[Senate Hearing 114-659]
[From the U.S. Government Publishing Office]
TRANSPORTATION AND HOUSING AND URBAN DEVELOPMENT, AND RELATED AGENCIES
APPROPRIATIONS FOR FISCAL YEAR 2017
----------
WEDNESDAY, NOVEMBER 16, 2016
U.S. Senate,
Subcommittee on the Committee on Appropriations,
Washington, DC.
The subcommittee met at 2:20 p.m., in room SD-192, Dirksen
Senate Office Building, Hon. Susan Collins, (chairman)
presiding.
Present: Senators Collins, Cassidy, Daines, Reed, and
Murphy.
The Automated and Self-Driving Vehicle Revolution:
What is the Role of Government?
opening statement of senator susan m. collins
Senator Collins. This hearing will come to order.
Good afternoon. I'm pleased to be joined today by our
ranking member, Senator Jack Reed, and by Senator Cassidy and
others who will be joining us, as we hold this hearing to
examine the role of government in enhancing roadway safety
through the careful deployment of automated and self-driving
vehicles, which could revolutionize our transportation system.
I also welcome our panel of witnesses. We are joined today
by Mark Rosekind, the Administrator of the National Highway
Traffic Safety Administration; Deborah Hersman, the President
and CEO of the National Safety Council; Paul Brubaker, the
President and CEO of the Alliance for Transportation
Innovation; and Dr. Nidhi Kalra, the Senior Information
Scientist at the RAND Corporation. I look forward to hearing
from each of you.
The potential of the new technology that we're examining
today is exciting, cars that drive themselves and avoid
accidents, seniors and disabled individuals able to retain or
gain their ability to get around town. At first, it sounds like
science fiction or an advance far, far off into the future. But
this technology is being tested and perfected right now, and it
could save thousands of lives.
According to the most recent statistics from the National
Highway Traffic Safety Administration, more than 35,000 lives
were lost in crashes on U.S. roadways in the year 2015.
Preliminary estimates indicate a 10 percent increase in the
first half of 2016. It is important to highlight that 94
percent of these crashes are the result of human factors or,
simply put, driver error such as distracted driving--sometimes
caused by hands-free technology--impaired driving, and
speeding. These crashes are preventable, and we should be doing
everything we can to eliminate them.
The testing and safe deployment of automated vehicle
technologies has the potential to reduce substantially the
number of driver-related crashes and fatalities. Let me
emphasize that point: if this technology were to be perfected,
automated vehicles could eliminate many of the crashes on our
Nation's roadways and save thousands of lives each year.
While fully self-driving autonomous vehicles are still
years away from being available to the general public, many new
vehicles already have driver-assisted features such as
automatic emergency breaking, rearview cameras, my personal
favorite, and lane-keep assist systems. These technologies are
already making a difference, saving lives and reducing
injuries.
In addition to improving safety, self-driving vehicles can
provide mobility options to our seniors and disabled people,
particularly those living in rural communities like my State of
Maine, where many of our older drivers currently do not have an
easy way to get to the doctor or to the grocery store. Public
transportation is nonexistent in much of Maine, and taxi
service is very limited in rural areas. Seniors who can no
longer drive often have very few options.
A self-driving car, or even one with limited automated
features, could help seniors feel more secure driving at night
and could help those who currently must rely on others to get
around, to maintain their independence. A recent survey of
drivers over the age of 50 showed that almost 80 percent of
those who plan to buy a car within the next 2 years will be
seeking automated features such as blind spot warning, crash
mitigation, and lane departure warning systems. Yet many
consumers remain wary of purchasing a fully autonomous vehicle
anytime soon. I know that the RAND Corporation has done a lot
of work in the area of the potential benefits of automated
technologies for our seniors.
To help accelerate the safe testing and development of
autonomous vehicles, the Department of Transportation released
its ``Federal Automated Vehicles Policy'' on September 20th,
which includes vehicle performance guidance, model state
policy, an assessment of current regulatory tools, and a
discussion of future regulatory tools that Congress may want to
consider.
While this was an important first step, it is clear that
the DOT's guidance document needs further refinement to help
ensure that automakers are able to bring the safety and
mobility benefits of autonomous vehicles into the marketplace
without unnecessary government regulations. I am particularly
interested in hearing from our witnesses today on what
improvements can be made to these guidelines.
Automobile manufacturers face a number of roadblocks
integrating autonomous vehicles onto our Nation's highways.
Some have called on the Administration to put the brakes on
autonomous vehicle deployment by imposing onerous requirements
through a rulemaking process which could take several years to
finalize, and my fear is that that would stifle innovation in
the meantime.
We must recognize that automated vehicle technology is
advancing faster than government agencies can act, and in this
instance impeding the advancement of technology may prevent us
from saving lives. On the other hand, we also know that there
have been some problems with the automated technology that has
increased the incidence of distracted driving and perhaps led
to additional accidents. The Department's guidance provides a
more effective voluntary 15-point safety assessment that
incentivizes automakers to certify that they have addressed all
relevant issues ranging from cybersecurity to human-machine
interface to ethical considerations.
Another challenge facing autonomous vehicles is the vast
number of State and local governments that are seeking to
regulate this technology. Unlike vehicle safety, which is
governed by Federal law, the safety and licensing of drivers
are under the jurisdiction of State governments. While the
Department's guidance deters States from interfering with
Federal vehicle safety standards, one topic we will explore is
whether or not DOT's approach also has the effect of
encouraging State governments to set prohibitive requirements
related to testing, licensing, and registration for automated
vehicle testing.
Given the public's keen interest in self-driving and
automated vehicles, it is important for State governments to
take a balanced approach in allowing the research, development,
safe testing, and deployment of such vehicles. However, I would
note that one serious accident, such as occurred in Florida,
can cause great mistrust of this new technology, and that could
lead to greater State restrictions on testing and deployment.
It is particularly important that DOT work with automotive
manufacturers, suppliers, and technology companies, as well as
with the commercial end users of automated vehicles, to ensure
that they are operating with an abundance of caution to avoid
excessive government intervention that could interfere with the
timely development of technologies that truly could save
thousands of lives.
[The statement follows:]
Prepared Statement of Senator Susan M. Collins
I am pleased to be joined today by our Ranking Member, Senator Jack
Reed, as we hold this hearing to examine the role of government in
enhancing roadway safety through the careful deployment of automated
and self-driving vehicles, which could revolutionize our transportation
system.
I also welcome our panel of witnesses. We are joined today by Mark
Rosekind, Administrator of the National Highway Traffic Safety
Administration; Deborah Hersman, President and C.E.O. of the National
Safety Council; Paul Brubaker, President and C.E.O. of the Alliance for
Transportation Innovation; and Dr. Nidhi Kalra, Senior Information
Scientist at the RAND Corporation. I look forward to hearing from each
of you.
The potential of this new technology is exciting. Cars that drive
themselves and avoid accidents. Seniors and disabled individuals able
to retain or gain their ability to get around town. At first, it sounds
like science fiction or an advance far off in the future. But this
technology is being tested and perfected right now, and it could save
thousands of lives.
According to the most recent statistics from the National Highway
Traffic Safety Administration, more than 35,000 lives were lost in
crashes on U.S. roadways in 2015, and preliminary estimates indicate a
10 percent increase in the first half of 2016. It is important to
highlight that 94 percent of roadway crashes are the result of human
factors, or simply put, driver error, such as distracted driving--
sometimes caused by hands-free technology, impaired driving, and
speeding. These crashes are preventable, and we should do everything we
can to eliminate them.
The testing and safe deployment of automated vehicle technologies
have the potential to reduce substantially the number of driver-related
crashes and fatalities. Let me emphasize that point: if the technology
were perfected, automated vehicles could eliminate many of the crashes
on our Nation's roadways and save thousands of lives every year.
While fully self-driving autonomous vehicles are still years away
from being available to the general public, many new vehicles already
have driver-assist features such as automatic emergency breaking,
rearview cameras, and lane-keep assist systems. These technologies are
already saving lives and reducing injuries on our roadways.
In addition to improving safety, self-driving vehicles can provide
mobility options to our seniors and disabled populations, particularly
those living in rural communities like my State of Maine, where many of
our older drivers currently do not have an easy way to get to the
doctor or to the grocery store. Public transportation is nonexistent in
much of Maine, and taxi service is very limited in rural areas. Seniors
who can no longer drive often have very few options.
A self-driving car, or even one with limited automated features,
could help seniors feel more comfortable driving at night and could
help those who currently must rely on others to get around to maintain
their independence. A recent survey of drivers over the age of 50
showed that almost 80 percent of those who plan to buy a car within the
next 2 years will be seeking automated features such as blind spot
warning, crash mitigation, and lane departure warning systems. Yet many
consumers remain wary of purchasing a fully autonomous vehicle anytime
soon. I know Dr. Kalra will provide us with more insight on the
potential benefits of automated technologies for our seniors.
To help accelerate the safe testing and deployment of autonomous
vehicles, the Department of Transportation released its ``Federal
Automated Vehicles Policy'' on September 20th, which includes vehicle
performance guidance, model State policy, an assessment of current
regulatory tools, and a discussion of future regulatory tools for
Congressional consideration.
While this was an important first step, it is clear that D.O.T.'s
guidance document needs further refinement to help ensure that
automakers are able to bring the safety and mobility benefits of
autonomous vehicles into the marketplace without unnecessary government
regulations. I am particularly interested in hearing from all of our
witnesses today on what improvements can be made to these guidelines.
Automobile manufacturers face a number of roadblocks integrating
autonomous vehicles onto our Nation's roadways. Some have called on the
Administration to put the brakes on autonomous vehicle deployment by
imposing onerous requirements through a rulemaking process, which could
take several years to finalize, stifling innovation in the meantime.
We must recognize that automated vehicle technology is advancing
faster than government agencies can act, and in this instance, impeding
the advancement of technology may prevent us from saving lives. The
Department's guidance provides a more effective voluntary 15-point
``safety assessment'' that incentivizes automakers to certify they have
addressed all relevant issues ranging from cybersecurity, to human-
machine interface, to ethical considerations.
Another challenge facing autonomous vehicles is the vast number of
State and local governments that are seeking to regulate this
technology. Unlike vehicle safety, which is governed by Federal law,
the safety and licensing of drivers are under the jurisdiction of State
governments. While the Department's guidance deters States from
interfering with Federal vehicle safety standards, one topic we will
explore is whether or not D.O.T.'s approach also has the effect of
encouraging State and local governments to set prohibitive requirements
related to testing, licensing, and registration for automated vehicle
testing.
Given the public's keen interest in automated and self-driving
vehicles, it is important for State and local governments to take a
balanced approach in allowing the research, development, safe testing,
and deployment of automated vehicles. However, one serious accident,
such as occurred in Florida, can cause mistrust of this new technology,
and that could lead to greater local restrictions on testing and
deployment. It is particularly important that D.O.T. work with
automotive manufacturers, suppliers, and technology companies, as well
as the commercial end users of automated vehicles, to ensure they are
operating with an abundance of caution to avoid excessive government
intervention.
With that, let me call upon my colleague and friend Senator Reed,
the ranking member, for his opening statement.
Senator Collins. With that, let me call upon my colleague
and friend from Rhode Island, Senator Reed, the ranking member,
for his opening statement.
STATEMENT OF SENATOR JACK REED
Senator Reed. Thank you very much, Chairman Collins, for
your leadership on so many issues, but particularly for holding
this very important hearing on the future of self-driving
vehicles. Let me join you in welcoming our witnesses. Thank you
for your service and your expertise.
Automated and self-driving vehicles are not yet common on
our roads, but autonomous driving and safety features such as
automatic emergency brakes and parking assist have been
introduced gradually over the last several years. Manufacturers
and innovators are now poised to take transformative leaps in
the development, integration, and adoption of these
technologies, and what was once novel is at the brink of
becoming commonplace.
As with any change, it brings opportunity as well as risk
and anxiety. Particularly for the millions of Americans who
earn a living as American commercial drivers, that is a
significant issue that we should address. What remains
unanswered is how this transportation revolution will evolve
and what steps regulators and industry should take to foster
and harness the positive aspects of this new technology.
Today auto manufacturers, ride sharing companies, and other
investors are funding research and development on driverless
vehicles. They are also launching pilot programs to gather data
and introduce consumers to different forms of the technology.
Uber is allowing customers in Pittsburgh to share rides in
self-driving cars. Otto, which is a self-driving bus and truck
company owned by Uber, autonomously delivered commercial goods
just last month. Tesla is collecting millions of miles of data
from its semi-autonomous vehicles and announced that it will
potentially make a full autonomous vehicle starting next year.
Google has been designing and testing cars with no human driver
for several years in California. And GM and Lyft have partnered
to build an autonomous fleet that will be available for ride
sharing.
These companies are all using different strategies to
achieve seemingly different goals. Some seek to provide
efficient, accessible, and cost-effective transportation
similar to Transit. Others want to improve freight
transportation through fleets of autonomous trucks that can
save gas and operator on the clock. What is clear is that
technology is fundamentally changing vehicles as we know them.
This innovation has the potential to dramatically improve
highway safety, as well as expand mobility access for thousands
of people, as the Chairman spoke about, seniors particularly.
And it also has safety features which, as the Chairman
commented on, are significant. In 2015, more than 35,000 people
died in crashes on the Nation's highways. The number represents
an 8 percent increase from the previous year and marked the
deadliest year on record since 2008. These vehicles might help
immensely in that regard. NHTSA (National Highway Traffic
Safety Administration) found that 94 percent of those deaths
were the result of human error, so autonomous vehicles could be
a real game changer in this regard.
We have a responsibility to fulfill this technology's
promise and foster American innovation. But we also must be
cognizant of the consequences of these technologies for the
shifts in the American economy and effects on the American
worker. Self-driving cars and trucks will certainly demand new
kinds of jobs and skills, but these jobs may be in different
sectors of the economy. For millions of Americans, particularly
those without a college degree or advanced training, driving a
bus, a cab, or a truck can provide a decent income. In fact,
for many, it is the ticket to the middle class.
In the latest year of data available from the Bureau of
Labor Statistics, there are more than 4 million workers in the
United States employed as drivers of trucks, taxis, chauffeurs
or delivery vehicles. All of these are potentially in a space
where they could be replaced by an autonomous vehicle. So we
have to make sure that this technology not only enables better
productivity but that it doesn't disqualify millions,
literally, of Americans from good, solid jobs.
We need to think through this policy very thoughtfully.
This is a broader conversation not involving just the
technologies but the Department of Labor, policymakers from a
huge swath of the government, so that we get it right, and we
get it right for the American workers as well as for the
technology interest.
Our regulatory agencies--and the Chairman has made the
point that it's not just Federal but its State and local--have
to be harmonized. We have to do this thoughtfully, and that is
why, again, I commend the Chairman for calling this hearing
today.
NHTSA--thank you--recently released a guidance document for
comment with the intent of promoting technological advancement
and preventing a patchwork of State regulations. So thank you,
Administrator Rosekind, for that work.
There are many questions. When will NHTSA initiate a formal
rulemaking on self-driving vehicles? When and how will the
Federal motor vehicle safety standards be updated? How much
data does NHTSA want to collect from industry and consumers,
and how will it protect that data from cyber threats?
We know industry will continue to innovate, deploy, and
develop these technologies at a faster pace than government can
respond, and this will present a challenge. We have to
recognize this up front.
But again, let me thank you, all of our witnesses for
participating in our hearing today. I apologize for my Rhode
Island accent. Other than that, I look forward to the
testimony. Thank you.
Senator Collins. Thank you very much, Senator Reed.
I am going to put all of the witnesses' impressive bios
into the record so that we can proceed with the testimony since
I mentioned all of you in my opening statement.
We will start with the Administrator of the National
Highway Traffic Safety Administration, known as NHTSA, the
Honorable Mark Rosekind.
STATEMENT OF THE HON. MARK ROSEKIND, ADMINISTRATOR,
NATIONAL HIGHWAY TRAFFIC SAFETY
ADMINISTRATION
Mr. Rosekind. Thank you, Chairman Collins, Ranking Member
Reed, other members of the committee. I really appreciate you
holding this hearing today and for the opportunity to testify.
At NHTSA, our mission is to save lives on America's
roadways. For 50 years, we have carried out that mission by
writing and enforcing strong regulations to make vehicles
safer, fighting against drunk-driving, building a national
consensus about seatbelt use, and so many other efforts that
have saved hundreds of thousands of lives on America's roads.
But we have far more work to do, and that work can be
measured by some alarming numbers that all of you cited
already. In 2015, we lost 35,092 people on our public roads. At
NHTSA, we know those are not just numbers. Those are mothers
and fathers, brothers and sisters, colleagues, co-workers,
friends. And the problem is that it's all getting worse.
Last month we announced that roadway fatalities in the
first half of this year are up over 10 percent, and it is
against this backdrop that the Department of Transportation,
under the leadership of Secretary Anthony Foxx, has been
working so hard on our efforts to accelerate the safe
deployment of automated vehicle technologies, because while
automated vehicles carry enormous potential to transform
mobility and reshape our transportation system, it is truly
their awesome potential to revolutionize roadway safety that
has all of us so motivated.
And there is one more number that helps us explain why, and
that number is 94. That's the percentage of crashes that can be
tied back to human choice or error. That's a choice to speed or
drive drunk, send a text message from behind the wheel, or
misjudge the stopping distance. That 94 percent represents the
untold potential of automated vehicle technologies.
We envision a future where advanced technologies not only
help reduce crashes but a world with fully self-driving cars
that hold the potential to eliminate traffic fatalities
altogether.
The Federal Automated Vehicles Policy, which the Department
issued on September 20th, is the world's first comprehensive
government action to guide the safe and efficient development
and deployment of these technologies.
The policy covers four areas: one, vehicle performance
guidance for automakers, tech companies, researchers, and other
developers, testers, and deployers of automated vehicle
technologies; two, a model state policy to build a consistent
national framework for the testing and operation of automated
vehicles; three, an exploration of the use of our current
regulatory tools that can be used to advance these
technologies; and four, a discussion of possible new tools that
the Federal Government may need to promote the safe deployment
of advanced technologies as the industry continues to develop.
I'd like to share just for a few moments a little bit about
our approach.
For 50 years, our traditional approach at NHTSA has largely
been reactive. NHTSA prescribes safety standards and then
responds to problems as they arise. A traditional method of
regulating these new technologies would be to engage solely in
the rulemaking process, writing new regulations that prescribe
specific standards and typically take years to take effect. Our
view is that approach would be slow, it would stymie innovation
and would stall the introduction of these technologies.
Our policy takes a very different path, built on proactive
safety which will better serve both safety and innovation. This
policy allows us to work with automakers and developers on the
front end, to ensure there are sound approaches to safety
throughout the entire development process. This is a new
approach, and it's going to take some adjustment for everyone
involved.
But we are confident that it will help us accomplish two
goals: first, to make sure that new technologies are deployed
safely; and second, to make sure that we don't get in the way
of innovation. Our approach is non-prescriptive. It does not
tell developers how they must provide safety but instead it
builds a transparent and proactive approach to ensure that they
are properly addressing the critical safety areas.
But that future is not without threats. As President Obama
wrote when announcing the policy, quote, ``The quickest way to
slam the brakes on innovation is for the public to lose
confidence in the safety of new technologies. Both government
and industry have a responsibility to make sure that doesn't
happen.''
It is our view that the best way that we can build that
public confidence is by working together, showing the public
that the government is on the side of innovation, and that the
industry is on the side of safety.
I submit the balance of my statement for the record, and I
look forward to taking your questions. Thank you.
[The statement follows:]
Prepared Statement of Hon. Mark Rosekind
Chairman Collins, Ranking Member Reed, and Members of the
Committee: Thank you for holding this hearing and inviting me to
testify. My name is Mark Rosekind, and I am the Administrator of the
National Highway Traffic Safety Administration, or NHTSA.
At NHTSA, our mission is to save lives on America's roadways. For
50 years, we have carried out that mission by writing and enforcing
strong regulations to make vehicles safer, fighting against drunk
driving, building a national consensus about seatbelt use, and so many
other efforts that have saved hundreds of thousands of Americans.
But we have far more work to do. And that work can be measured by
some alarming numbers.
In 2015, we lost 35,092 people on our public roads. At NHTSA, we
know that is not just a number. Every one of those is a mother or
father, a son or daughter, a coworker, a friend. In the United States,
we lose the equivalent of a fully-loaded 747 on our roadways every
single week.
And the problem is getting worse. Last month we announced that
roadway fatalities in the first half of this year are up over 10
percent.
It is against this backdrop that the Department of Transportation,
under the leadership of Secretary Anthony Foxx, has been working so
hard on our efforts to accelerate the safe deployment of automated
vehicle technologies.
Because while automated vehicles carry enormous potential to
transform mobility, reshape our transportation system and transform our
economy, it is their awesome potential to revolutionize roadway safety
that has us so motivated.
And there is one more number that helps explain why. That number is
94. That is the percentage of crashes that can be tied back to a human
choice or error. That's a choice to speed or drive drunk, to send a
text message from behind the wheel or misjudge the stopping distance.
And that 94 percent figure represents the untold potential of
automated vehicle technologies. We envision a future where advanced
technologies not only help reduce crashes, but also make possible a
world in which fully self-driving cars hold the potential to eliminate
traffic fatalities altogether.
The Department of Transportation views this moment as the cusp of a
new technological revolution that may transform roadway safety forever.
The Federal Automated Vehicles Policy, which the Department and
NHTSA issued in mid-September, is the world's first comprehensive
government action to guide the safe and efficient development and
deployment of these technologies. Today, I will discuss that Policy,
how we developed it, and where we are going next.
In January of this year, Secretary Foxx made two important
announcements.
First, he announced that President Obama was making a $3.9 billion
budget request for automated vehicles research. This is a major
commitment from the Administration to advance this technology, and DOT
continues to strongly support this request.
Second, he directed NHTSA to write a new policy covering four
areas: One, vehicle performance guidance for automakers, tech
companies, researchers and other developers and testers of automated
vehicle technologies. Two, a model State policy to build a consistent
national framework for the testing and operation of automated vehicles.
Three, an exploration of the use of our current regulatory tools that
can be used to advance these technologies. And four, a discussion of
possible new tools that the Federal Government may need to promote the
safe deployment of advanced technologies as the industry continues to
develop.
Over the subsequent 9 months, NHTSA hit the road, traveling to
discuss automated vehicles with industry, academics, State governments,
safety and mobility advocates, and the public. This Policy is the
product of that significant input.
Before discussing the individual components, I would like to share
a few thoughts about our approach.
First it is important to understand our traditional approach to
regulating motor vehicles. For 50 years, our approach has largely been
reactive. NHTSA has prescribed safety standards, and then responded to
problems as they arise.
A traditional approach to regulating these new technologies would
be to engage solely in rulemaking process, writing new regulations that
prescribe specific standards. Our view is that approach would stymie
innovation and stall the introduction of these technologies.
It would also be a long process. Rulemakings, and the research
necessary to support them, take years, meaning that any rule we might
offer today would likely be woefully out-of-date by the time it took
effect, given the pace of technological development in this space. Let
me be clear that using the notice-and-comment rulemaking process to
establish new standards will absolutely play an important role as this
technology matures and is adopted. But it is not the only tool in our
bag, and we have created an innovative approach that will better serve
both safety and innovation in the immediate term.
Our Policy represents a continuation of the new proactive safety
approach that we have built at NHTSA under the leadership of Secretary
Foxx. This Policy allows NHTSA to work with automakers and developers
on the front end, to ensure that sound approaches to safety are
followed throughout the entire design and development process. This is
a new approach, and it's going to take some adjustment for everyone
involved. But we are confident that it will help us accomplish two
goals: first, to make sure that new technologies are deployed safely;
and second, to make sure we don't get in the way of innovation.
As the Federal regulator with the responsibility of ensuring
vehicles are as safe as they can possibly be, we play an important role
on behalf of the American public to ensure that vehicle technologies do
not present safety threats.
At the same time, we recognize the great lifesaving potential of
these new technologies, and want to do everything we can to make sure
that potential is fully realized and that they are deployed as quickly
as possible to save as many lives as we can.
Some people have talked about safety and vehicle automation as on
the opposite ends of a spectrum, as if there were a trade-off between
safety and innovation. But at the Department of Transportation, we view
our role as promoting safety innovation. Our Policy is designed to
promote the safe and expeditious deployment of new technologies that
have the potential to reduce crashes and save lives.
Our approach is not prescriptive. It does not tell developers how
they must provide safety, but instead it builds a transparent and
proactive approach to ensure that they are properly addressing the
critical safety areas.
Finally, I want to be clear that while this Policy establishes an
important framework for the development and deployment of automated
vehicles, it is not the final word. In our view, this Policy is the
right tool at the right time. It answers a call from industry, State
and local governments, safety and mobility advocates and many others to
lay a clear path forward for the safe deployment of automated vehicles
and technologies.
But we intend this Policy to evolve over time. That evolution will
be based on comments we receive from the public, our own experience in
implementing it over the coming months and years, and, perhaps most
importantly, based on the rapid evolution of the technology itself. We
have designed this Policy to be nimble and flexible, to allow us to
stay at the leading edge of this revolution.
Before I discuss each component of the Policy, allow me to say a
few words on definitions.
First, it is important to note that with this Policy, we are
officially adopting the SAE International levels of automation, ranging
from zero to five. The primary focus of the Policy overall is on what
we refer to as ``highly automated vehicles'', or HAVs. Those are
vehicles at levels three through five on the SAE level scale, or
vehicles that--at least in some circumstances--take over full control
of the driving task. A portion of the first section of the Policy also
applies to Level 2 vehicle systems, which include advanced driver-
assistance systems already on the road today.
The Policy covers all automated vehicles that are designed to
operate on public roads. That includes personal light vehicles, as well
as heavy trucks. It even includes vehicles that might be designed to
not carry passengers at all.
Finally, I note that most of the Policy is effective immediately.
We expect that developers and manufacturers of AV technologies will use
the Policy to guide their safety approach. Some portions of the
Policy--notably the Safety Assessment Letter in the Vehicle Performance
Guidance--will become effective following a Paperwork Reduction Act
process that we expect to be completed within the next few months.`
Vehicle Performance Guidance for Automated Vehicles
The first section is the Vehicle Performance Guidance for Automated
Vehicles. This is guidance for manufacturers, developers and other
organizations involved in the development of automated vehicles. The
heart of the Guidance is a 15 point ``Safety Assessment'' that spells
out the critical safety areas that developers should address for the
safe design, development, testing and deployment of highly automated
vehicles prior to the sale or operation of such vehicles on public
roads.
The Safety Assessment covers areas such as the operational design
domain--essentially the where and when an AV is designed to operate
automatically--fallback conditions, cybersecurity, privacy, and the
human-machine interface.
We identified these areas through our extensive consultations with
industry, academia and advocates as the critical safety issues that
must be addressed to ensure that automated technologies are safe.
Critically, the Guidance does not specify how AV developers are
intended to address the areas. Instead, the Guidance asks developers to
document their own processes and then provide NHTSA with a Safety
Assessment letter in which they explain their approach. This process is
expected to yield a variety of different approaches for every one of
the areas. That is intentional, and is one of the ways that we are
preserving and promoting the innovation process. Government does not
have all the answers, and our view is that the more approaches that
innovators take to solving these problems, the more likely we are to
find the best way.
Model State Policy
The second section is the Model State Policy. For the last 50
years, there has been a fairly clear division of responsibility between
the Federal Government and the States for the oversight and regulation
of motor vehicles. Generally speaking, it has been the Federal
Government's responsibility to regulate motor vehicles and equipment
safety, while the States have regulated drivers and traffic laws.
That division of responsibility may be less clear in a highly
automated vehicle world where increasingly the vehicle's automated
systems become the driver.
The Model State Policy delineates the Federal and State roles for
the regulation of these vehicles, and it outlines the approach we
recommend to States as they consider the regulation of testing and
operation of automated vehicles on their public roads. Our goal is to
build a consistent national framework for the development and
deployment of automated vehicles, so that users can take their vehicles
across State lines as they can today, and so that developers are
building toward a single set of standards, rather than 50.
The Model State Policy confirms that States retain their
traditional responsibilities for vehicle licensing and registration,
traffic laws and enforcement, and motor vehicle insurance and liability
regimes. At the same time, the Policy reaffirms that the Federal
Government will continue to be responsible for the oversight of vehicle
safety and design, including automated features.
The Policy was developed in close coordination with the American
Association of Motor Vehicle Administrators (AAMVA), individual States
and other stakeholders. It suggests recommended areas for States to
consider in the development of their own regulations, including testing
regimes and registration. It also identifies a number of areas that
need to be further discussed and developed, including how law
enforcement will interact with highly automated vehicles, and the
development of a consistent approach to insurance and liability
challenges. We also note in the Policy that States do not have to take
any action at all.
NHTSA's Current Regulatory Tools
The third section addresses NHTSA's Current Regulatory Tools. This
section discusses how NHTSA will use the tools currently at its
disposal to promote and expedite the safe development and deployment of
highly automated vehicles.
The first of those tools discussed is our interpretation authority.
The current Federal Motor Vehicle Safety Standards generally do not
contemplate automated vehicle technologies. Therefore, it can sometimes
be unclear how those standards apply to advanced technologies. In this
section, we lay out the process by which developers of AV technologies
can submit interpretation requests to the agency to determine whether
and how their technologies conform with the standards. The agency also
commits to a greatly expedited process for reviewing these
interpretation requests. On simple safety-related interpretation
requests, we commit to providing answers within 60 days. Compared to
historical norms, that is lightning speed.
The second tool discussed is our exemption authority. Congress has
granted NHTSA the authority to provide exemptions to manufacturers to
deploy vehicles that do not conform to the Federal Motor Vehicle Safety
Standards. While these exemptions are admittedly limited--to 2,500
vehicles for each of 2 years--the Agency views this tool as an
important way of enabling a manufacturer to put a test fleet on the
road to gather critical safety data and improve its technologies. The
Policy similarly commits to an expedited process on simple safety-
related exemptions, providing an answer within 6 months from the
application.
The Agency's broadest power is its ability to write new safety
standards. While this tool tends to take the longest amount of time--
usually a period of years--it is the method that will ultimately allow
for the large-scale deployment of nontraditional vehicle designs and
equipment under consistent, broadly applicable standards. In addition,
to the extent that performance-based standards are adopted, this tool
has the potential to allow for technological innovation while
maintaining safety.
In this section, we also highlight that the Agency retains its
broad defects and enforcement authority. We use that authority to
investigate any unreasonable risks to safety, and to recall unsafe
vehicles from the road. The same day NHTSA issued the Policy, we also
issued an Enforcement Guidance Bulletin that makes clear that the
Agency's traditional enforcement authorities extend to advanced vehicle
technologies.
Modern Regulatory Tools
The fourth and final section of the Policy discusses Modern
Regulatory Tools, identifying 12 potential new tools, authorities and
resources that could aid the safe deployment of new lifesaving
technologies and enable the Agency to be more nimble and flexible.
Today's governing statutes and regulations were developed before
highly automated vehicles were even a remote notion. For that reason,
current authorities and tools alone may not be sufficient to ensure
that highly automated vehicles are introduced safely, and to realize
their full safety promise. This challenge requires NHTSA to examine
whether the ways in which the Agency has addressed safety for the last
several decades should be expanded and supplemented.
The new tools identified in this section include premarket
approval, expanded exemption authority, imminent hazard authority, new
research and hiring tools, and others that may better equip the Agency
in the future as more technologies move from the lab to the road. These
tools are offered for consideration by policymakers, industry,
advocates and the public as we move forward.
One thing we know for certain is that the agency will need
additional resources as this technology develops and is adopted. I have
great confidence in the NHTSA team's expertise and ability. But it is
undeniable that as more automakers move technology from the lab to the
test track to the road, we will need to make sure our Agency is
properly resourced to maintain pace.
We continue to support the President's budget request for more
research dollars, and are committed to working with you in the coming
months and years to identify what resources--both in personnel and
research funding--will be necessary to achieve our mission.
Next Steps
Finally, with respect to the Policy, I would like to highlight once
again that we fully intend this Policy to be the first iteration of
many to come. The Policy is effective now, and will continue to evolve
based on feedback and our experience implementing it, and, most
importantly, to keep pace with innovation. To that end, each section of
the Policy highlights a series of next steps that we will take to
implement and improve the Policy over time.
The first is our solicitation of public input. We are doing that
through an open comment period that is open now through November 22nd.
NHTSA is also hosting a series of public workshops that began earlier
this month on different sections of the Policy. I will note here that
the full Policy, additional materials, and the portal for public
comments can be found at www.nhtsa.gov/AV.
Over the coming months we will be engaging experts to review the
Policy, issuing further guidance on the Safety Assessment letter, and
engaging stakeholders across the spectrum to help flesh out other areas
of the Policy. For example, we will work with law enforcement
organizations to further the conversation about how AVs will interact
with the police, and work with industry to build the framework for the
data sharing discussed in the Vehicle Performance Guidance. We are also
engaged with other operating modes throughout the Department of
Transportation, recognizing the roles and responsibilities they play
with respect to public transit, commercial freight operations, and the
highway system on which automated vehicles will operate.
We do not pretend to have answered every question in this Policy,
and we will continue the conversation with the public about the best
ways to develop and improve our Policy as we learn more. To that end,
the Department of Transportation has committed to reviewing and
updating the Policy annually.
As I conclude, I want to say a few words about the importance of
the present moment in history. We have an industry that is rapidly
developing innovative new technologies. And we have a government that
is inspired and excited about the future of this technology.
But that future is not without threats. Bad actors or bad incidents
could threaten to derail our collective efforts.
I want to close with the words President Obama used when he
announced our new Policy in an op-ed in the Pittsburgh Post-Gazette. He
wrote, ``There are always those who argue that government should stay
out of free enterprise entirely, but I think most Americans would agree
we still need rules to keep our air and water clean, and our food and
medicine safe. That's the general principle here. What's more, the
quickest way to slam the brakes on innovation is for the public to lose
confidence in the safety of new technologies. Both government and
industry have a responsibility to make sure that doesn't happen.''
It is our strong view that the best way we can build that public
confidence is by working together, showing the public that the
government is on the side of innovation and the industry is on the side
of safety. We encourage you to join with us as we continue to develop
this Policy and show the American public that their safety is the
highest priority for all of us.
Thank you.
Senator Collins. Thank you very much.
Ms. Hersman.
STATEMENT OF HON. DEBORAH HERSMAN, PRESIDENT AND CEO,
NATIONAL SAFETY COUNCIL
Ms. Hersman. Chairman Collins, Senator Reed, thank you for
having me today. I'd like to acknowledge my board chair, Mr.
John Surma, who is here with me in the audience today.
Today we have millions of drivers behind the wheel, we
spend millions of dollars on education and enforcement
campaigns, but we still lose tens of thousands of people on our
roadways and experience billions of dollars in economic losses
as a result of highway crashes. In spite of safer vehicle
designs and record-setting seatbelt use rates across the
Nation, operating a motor vehicle remains one of the deadliest
things that we do every day.
Compared to other high-income countries, the United States'
death rate is more than double our counterparts, and it's not
because we don't have automated vehicles. It's because we
aren't willing to do the hard things that we know will save
lives.
Three to five thousand lives per year would be saved if
everyone buckled up. Ten thousand lives would be saved if
nobody drove drunk. Just three technologies that are available
on cars today--automatic emergency braking, lane departure
warning, and blind spot detection--if these were all standard
equipment on cars, 10,000 lives could be saved.
The AV policy begins an important discussion, and we
applaud NHTSA and DOT for issuing it. Federal leadership on
motor vehicle safety is necessary because we can only have one
level of safety when it comes to our stakeholders and our
constituents. Your constituents need to be confident that
vehicles are safe regardless of where they reside.
Manufacturers need certainty in order to invest in design and
production, and States do not possess the expertise or the
resources to replicate design, testing, and reporting programs.
Further, a patchwork of requirements will result in
confusion for consumers and increased costs for manufacturers
and operators attempting to comply with myriad requirements.
Finally, the absence of a safe, workable standard will
drive development and testing and deployment overseas,
resulting in the flight of innovation from the United States
and the jobs that accompany it to locations outside of our
borders.
NSC fully believes that automated vehicles have the
potential to save lives and prevent injuries, but here are
several key issues that we recommend that policymakers
consider.
Delayed integration into the fleet. It will be a long time
before highly automated vehicles replace our current fleet. How
do we ensure rapid acceptance and adoption of lifesaving
technologies that are available today? Electronic stability
control paints the picture of the headwinds that we're facing.
DSRC--also known as V2V, V2I, and V2X--complementary
technology must be taken into account. We need a belt-and-
suspenders approach when it comes to highway safety.
Data sharing in a post-crash event. My experience at the
NTSB taught me that we must rely on the data for truths. If you
can't access it, you won't learn the lessons.
De-identified data. The automotive industry must figure out
how to balance privacy and proprietary concerns against the
predictive value of big data. Too many lives are at stake.
Cooperation will help us unravel unintended consequences and
identify new failure modes.
Consumer education. Today, 40 percent of consumers are
startled by the way that their car has behaved. We will not
realize the full benefit of technology if we don't keep human
beings in the loop.
Lack of common taxonomy and standards. Today, manufacturers
have multiple names and different performance expectations for
similar systems. Shouldn't all AEB systems prevent collisions?
NHTSA resources. We cannot ask NHTSA to do more with less.
Automated vehicles will require more evaluation, more testing,
and more expertise than NHTSA possesses today.
Although we can imagine a future with automated vehicles,
the transition from Level 1 to Level 5 will be messy as we deal
with predictable human-machine issues. We cannot afford to
ignore the carnage on our roadways that is a national epidemic.
Efforts like Road to Zero will decrease fatalities today,
tomorrow, and in the future if we embrace proven counter-
measures and accelerate deployment of effective ADAS
technologies and highly automated vehicles.
The National Safety Council appreciates the committee's
leadership. If safety for the traveling public is the ultimate
goal, advanced technology provides the most promising
opportunity to achieve that outcome and will go a long way
toward eliminating preventable deaths in our lifetime. Thank
you.
[The statement follows:]
Prepared Statement of Hon. Deborah A.P. Hersman
Chairman Collins, Ranking Member Reed and other members of the
subcommittee, thank you for inviting me to testify before you today.
The National Safety Council (NSC) is a 100-year-old nonprofit with a
vision to end preventable deaths in our lifetime at work, in homes and
communities and on the road through leadership, research, education and
advocacy. Our more than 13,500 member companies represent employees at
more than 50,000 U.S. worksites. For decades we have advocated for
safer cars, safer drivers and a more forgiving environment in and
around vehicles. We have led large scale public education campaigns on
the importance of seatbelts and airbags, eliminating distracted
driving, and helping consumers understand the technologies in their
vehicles to reduce deaths and injuries on our roadways. We also educate
close to 1.3 million drivers a year in Defensive Driving courses.
national highway traffic safety administration automated vehicles
policy
The rapid pace of technological advancement means that regulators
are hard-pressed to keep ahead of industry as manufacturers offer
systems unheard of just a decade ago. However, NSC believes there are
appropriate and necessary roles for both innovation by manufacturers
and regulation by officials charged with ensuring public safety. The
NHTSA (National Highway Traffic Safety Administration) Federal
Automated Vehicles Policy (AV policy) is a step in the right direction.
It provides a framework in which manufacturers can work and States can
establish appropriate enforcement and oversight, while underscoring
NHTSA safety authority and recognizing the need for additional tools to
keep pace in this fast-moving environment. It also provides guidance
for more uniform Federal oversight rather than a potential patchwork of
regulations by the States.
Federal leadership on motor vehicle safety is necessary because
there should only be ONE LEVEL OF SAFETY. Consumers need confidence in
vehicles regardless of where they reside; manufacturers need certainty
in order to invest in design and production, and States do not possess
the expertise and the resources to replicate design, testing and
reporting programs. Further, a patchwork of requirements will result in
confusion for consumers and increased cost for manufacturers and
operators attempting to comply with a myriad of requirements. Finally,
the absence of a safe, workable standard will drive development,
testing and deployment overseas, resulting in the flight of innovation
and the jobs that accompany it to locations outside of the U.S.
the lifesaving potential of advanced technology
NSC believes advanced vehicle technology, up to and including fully
automated vehicles, can provide many benefits to society. The most
important contribution will be the potential to greatly reduce the
number of fatal crashes on our roadways, which are increasing. Every
day we lose approximately 100 people in motor vehicles crashes, and
every year more than 4 million people are injured. Beyond the human
toll, these deaths and injuries cost society over $379 billion,
including productivity losses, medical expenses, motor vehicle property
damages and employer costs.\1\
---------------------------------------------------------------------------
\1\ Injury Facts 2016.
---------------------------------------------------------------------------
NSC data reveal that the 19,100 roadway fatalities during the first
6 months of 2016 are 9 percent higher than the same period last year
and 18 percent higher than the same period 2 years ago. If we are to
make a meaningful change in this trend, there must be a sense of
urgency coupled with large, near term gains to save lives on our
roadways.
Motor Vehicle Deaths on the Rise
Source: NSC analysis of National Center for Health Statistics (NCHS)
mortality data and NSC estimate for 2016.
While the absolute numbers of fatalities change from year to year,
many of the same behavioral problems remain persistent and have been
represented in the data for decades. For example, in 2014:
9,967 people were killed in alcohol-impaired driving crashes \2\
---------------------------------------------------------------------------
\2\ Https://crashstats.nhtsa.dot.gov/Api/Public/Publication/812231.
---------------------------------------------------------------------------
3,179 people were killed in distraction related crashes \3\
---------------------------------------------------------------------------
\3\ Https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/
812260.
---------------------------------------------------------------------------
9,385 people were killed while unrestrained.\4\
---------------------------------------------------------------------------
\4\ Https://crashstats.nhtsa.dot.gov/Api/Public/Publication/812246.
---------------------------------------------------------------------------
NHTSA estimates that 94 percent of all fatal crashes have an
element of human error. Therefore, if we are to eliminate or reduce the
number of fatalities on our roadways, advances in vehicle technology
must be part of the solution. However, it will likely be decades before
we have meaningful fleet penetration of fully automated vehicles.
Last month, the NSC and the National Transportation Safety Board
(NTSB) hosted a full day event with dozens of expert panelists focused
on Reaching Zero Crashes: A dialogue on the Role of Advanced Driver
Assistance Systems (ADAS).\5\ While there is a great deal of excitement
about highly automated vehicles (HAVs), automated vehicles and their
potential to save lives, it is important to recognize that many legacy
technologies represent the building blocks for fully automated
vehicles. Greater consumer acceptance of the dozens of safety
technologies that are available today would lead to more rapid adoption
of them, saving lives and preventing injuries.
---------------------------------------------------------------------------
\5\ Http://www.ntsb.gov/news/events/Pages/2016_dte_RT_agenda.aspx.
---------------------------------------------------------------------------
As an example, Electronic Stability Control (ESC) is a technology
that uses automatic computer controlled braking of individual wheels to
help the driver maintain control in risky driving scenarios. ESC
primarily mitigates single vehicle, loss of control crashes in which
drivers would run off the road. For passenger cars as well as light
trucks and vans, it is estimated that ESC systems have saved more than
4,100 lives during the 5-year period from 2010 to 2014, but
incorporation into vehicles on the road remains slow.\6\ The following
charts from the Highway Data Loss Institute (HDLI) reveal how slowly
ADAS technologies are achieving penetration in the U.S. fleet due to
normal turnover of inventory--with the average age of cars in the U.S.
fleet being 11.5 years old.\7\ Electronic stability control has been
available for decades and was mandated on all new passenger cars by the
2012 model year, but in 2015 only 40 percent of registered vehicles
were equipped with ESC. Despite a clear life-saving benefit, full fleet
penetration of this technology is not predicted until the 2040s.\8\
---------------------------------------------------------------------------
\6\ Https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/
812277.
\7\ Http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/
publications/national_transportation_
statistics/html/table_01_26.html_mfd.
\8\ Http://www.ntsb.gov/news/events/Documents/
2016_dte_RT_p1_p3_moore.pdf.
---------------------------------------------------------------------------
New Vehicle Series With Electronic Stability Control by Model Year
Source: HLDI.
Registered Vehicles With Electronic Stability Control by Calendar Year
Source: HLDI.
Registered Vehicles With Available ESC, Actual and Predicted
Source: HLDI.
ADAS already operate on the roadways today, but more could be done
to encourage greater fleet penetration. Features like lane departure
warning systems, blind spot monitoring, adaptive cruise control and
others help to prevent or mitigate crashes. The cost of these
technologies is declining and their impact is measurable. According to
the Insurance Institute for Highway Safety (IIHS), if four current
technologies--forward collision warning/mitigation, lane departure
warning/prevention, side view assist/blind spot monitoring, and
adaptive headlights--were deployed in all passenger vehicles, they
could prevent or mitigate as many as 1.86 million crashes and save more
than 10,000 lives per year.\9\ However, front crash prevention,
commonly referred to as automatic emergency braking, which was an
option in about half new 2015 model year cars, was in only 8 percent of
registered cars in 2015.\10\
---------------------------------------------------------------------------
\9\ Http://dx.doi.org/10.1016/j.aap.2010.10.020.
\10\ Http://www.ntsb.gov/news/events/Documents/
2016_dte_RT_p1_p3_moore.pdf.
---------------------------------------------------------------------------
Crashes Relevant to 4 Crash Avoidance Systems FARS and GES, 2004-2008
Source: Insurance Institute for Highway Safety.
Similar conclusions were reached in a July 2016, Carnegie Mellon
study which stated that just three technologies--forward collision
warning, lane departure warning and blind spot monitoring--could have
prevented or reduced as many as 1.3 million crashes annually and over
10,000 fatal crashes.\11\ This study further found that almost one
quarter of all crashes could be affected by these crash avoidance
systems, but only 2 percent of 2013 model year cars included these
systems as standard.
---------------------------------------------------------------------------
\11\ Http://dx.doi.org/10.1016/j.aap.2016.06.017.
---------------------------------------------------------------------------
While many of these technologies are available on higher value cars
or as part of an upgraded technology package today, they are not
standard equipment on all makes and models. Safety should not be just
for those who can afford it, especially for technologies that will
result in thousands of lives saved every year. The Carnegie Mellon
study estimated that if all light-duty vehicles were equipped with the
three technologies, they would provide a lower bound annual benefit of
about $18 billion. With 2015 pricing, it would cost about $13 billion
to equip all light-duty vehicles with the three technologies, resulting
in an annual net benefit of about $4 billion or a $20 per vehicle net
benefit. By assuming all relevant crashes are avoided, the total upper
bound annual net benefit from all three technologies combined is about
$202 billion or an $861 per vehicle net benefit, at current technology
costs.
NSC recognizes and applauds the voluntary commitment made earlier
this year by 20 automakers to include automatic emergency braking (AEB)
on all vehicles sold in the U.S. by 2022. Toyota has already committed
to beat this date by several years. Given the slow turnover of the
fleet, we encourage other manufacturers to view the 2022 date as a
finish line rather than a starting point and accelerate the roll out of
AEB and other lifesaving technologies.
Whether mandated or optional, in many cases these systems can
perform driving tasks more predictably, more conservatively and more
safely than a human driver, and may act without driver input if a
driver is distracted, impaired or incapacitated. However, because there
are no minimum standards for many of these technologies, legitimate
questions about their effectiveness remain. The line between ADAS and
highly automated vehicles may be blurred, as we have progressed far
down the path of advanced features with few questions and widespread
acceptance of semi-automated features.
dedicated short range communication
Another component of ADAS and automated vehicle systems is
dedicated short range communication (DSRC), which would allow vehicles
to communicate over dedicated spectrum bands with each other,
pedestrians, and infrastructure to prevent collisions. This technology,
often referred to as V2V (vehicle-to-vehicle), V2I (vehicle-to-
infrastructure), V2P (vehicle-to-pedestrian), or V2X (vehicle-to-
everything), is pending a rulemaking decision by NHTSA to establish
performance standards. NSC encourages NHTSA to release this standard
soon so that implementation of V2X can be more widespread.
DSRC can create redundant safety systems in motor vehicles. In
other modes of transportation, fail-safe designs can support operator
error, but in highway vehicles that task has fallen solely on drivers.
DSRC would allow a vehicle to communicate with a red light to
compensate for a fatigued driver, stop a car to prevent a collision
with a pedestrian if a driver fails to detect him or her, and prevent
or mitigate collisions between vehicles equipped with DSRC. DSRC has
been deployed by some manufacturers, but NSC believes it is an
important option in a safe systems approach to the design of HAVs and
anticipates it will be more widely deployed if there is more regulatory
certainty.
education and training
One component in the AV policy that should be a requirement moving
forward is the incorporation of driver education and training about new
safety technologies. With nearly 17.4 million new passenger cars and
trucks sold in 2015,\12\ understanding the technology on these vehicles
is necessary, yet a University of Iowa survey found that 40 percent of
respondents reported they had experienced a situation in which their
vehicle acted in an unexpected way.\13\ When this occurs in a real-life
driving situation, among multiple drivers, it can lead to disastrous
outcomes.
---------------------------------------------------------------------------
\12\ Http://www.autoalliance.org/auto-marketplace/sales-data.
\13\ University of Iowa. National Consumer Survey of Driving Safety
Technologies. July 30, 2015. Accessible at http://ppc.uiowa.edu/sites/
default/files/national_consumer_survey_
technical_report_final_8.7.15.pdf.
---------------------------------------------------------------------------
The National Safety Council and our research partners at the
University of Iowa are focused on educating consumers about in-vehicle
safety technology through our MyCarDoesWhat campaign.\14\ This brand
agnostic education campaign informs drivers about how safety
technologies work, how to best interact with them, and how to identify
situations when the technology may not perform optimally and should not
be relied upon. Because of the need for continued human involvement in
the operation of many of these features, the campaign tagline is You
are your car's best safety feature.
---------------------------------------------------------------------------
\14\ www.mycardoeswhat.org.
---------------------------------------------------------------------------
Visitors to MyCarDoesWhat.org realize improvement in general
knowledge and accurate comprehension of vehicle safety features.
Drivers cannot effectively use these life-saving technologies if they
do not understand both their functions and limitations. The AV policy
proposes that this education be delivered in multiple ways, including
computer based, hands-on and virtual reality training, and other
innovative approaches. The MyCarDoesWhat education campaign follows
that approach, and is developing virtual reality modules for release
early next year. Further, we recommend ongoing evaluation to determine
the effectiveness of the various messages, methods of delivery and
media so they can be improved over time.
standardized nomenclature and performance outcomes
Another way to reduce consumer confusion is to standardize the
nomenclature or taxonomy for advanced technologies. NSC, the State of
California, and Consumer Reports have recommended that, at the very
least, systems that are not completely automated or Level 5 should not
be described as such. ADAS, with emphasis on driver assist, represents
the vehicles being sold today and requires drivers to remain fully
engaged in the driving task. That fact is often lost in marketing,
media reports and consumer expectations. Labeling a motor vehicle as
``autonomous'' today, or even using terms such as ``autopilot'', only
confuses consumers and can contribute to losses of situational
awareness around the driving task.
By establishing standard nomenclature and establishing clear
performance outcomes, consumers will better understand what they should
expect from these technologies. For example, vehicles marketed as
having AEB will not necessarily come to a complete stop before a
collision.\15\ Some AEB systems only operate at higher speeds, and some
are designed to slow rather than stop prior to a collision. These
nuances may not be easily understood by consumers. IIHS reports that
systems with a warning only, but no automatic corrective action, reduce
frontal crash rates by about 25 percent, but vehicles with automatic
braking reduce crashes by more than 40 percent. Vehicles with a warning
and automatic braking reduce crash rates by about 50 percent.
Establishing a standardized, results-based, understandable definition
of AEB and other ADAS technologies would benefit consumers,
manufacturers, and dealers, as well as organizations that evaluate
vehicles for their safety benefits.
---------------------------------------------------------------------------
\15\ Http://www.nsc.org/learn/safety-knowledge/Pages/Driver-Assist-
Technologies.aspx.
---------------------------------------------------------------------------
Finally, the New Car Assessment Program (NCAP) program has operated
for nearly 40 years with a goal of testing vehicle safety systems and
educating consumers about them. Practically, it has created a mechanism
to allow consumers to evaluate vehicles on safety systems. NSC supports
NCAP and believes it is an important program to improve the safety of
the motor vehicle fleet. Standardized nomenclature and performance
outcomes will ensure NCAP can more effectively compare vehicle safety
systems between manufacturers, and even between a manufacturer's own
models.
data protection and sharing
The National Safety Council is very bullish on ADAS, and eventually
fully automated vehicles, because we know when implemented safely and
properly, they will help us realize huge gains in reducing roadway
fatalities. But a minimum requirement, if we are to realize the life-
saving benefits, will be ensuring that we have reliable event data
recorders that produce data in a standardized format that is
downloadable for investigators, law enforcement, State highway safety
offices, insurers and other relevant stakeholders. Following a crash,
we must be able to answer simple questions like whether the vehicle
systems or the human driver had control of the car, if and how the
vehicle was communicating with the driver, and if all systems were
working as designed.
The AV policy details the importance of data collection as these
technologies begin to be tested in real-world scenarios. Understanding
the circumstances and causes surrounding malfunctions, including at
lower levels of automation, will help make this technology stronger and
safer, and ensure failures are less likely to occur as technology
evolves. This will be especially important in assuring consumers of the
reliability of ADAS and automated technology. As the former Chairman of
the NTSB, I believe that minimum parameters should be set for data
preservation, standardization of formats, ease of access for post-crash
evaluation, and establishment of privacy protections early in the
process. Data-sharing programs require greater maturity and a strong
safety culture committed to continuous improvement.
Event data recorders (EDRs) are widely used throughout the
automotive industry in vehicles of all shapes and sizes, yielding
valuable data in crash reconstruction efforts. Similar devices are used
in other modes of transportation as well. Amtrak utilizes event data
recorders that automatically transmit data from locomotive recorders to
a control center when an event occurs. In the aviation industry, crash-
hardened ``black box'' data recorders store thousands of parameters of
data. Much of this information is collected after normal flights and
analyzed by the operator to learn about and improve operations, and in
the case of an incident, the data is invaluable to investigators to
determine what occurred. The same could be true for motor vehicles.
Missing from the policy, however, is clarification on whether
lower-level systems (below Level 3) should be subject to the same data
collection guidelines. The current event data recorder standards do not
require the majority of Levels 1 and 2 safety systems' sensors be
tracked.\16\ This lack of information limits real world data
availability that could guide the future development of these
technologies to make them safer. Currently, there is no easy way for
manufacturers, law enforcement, investigators or vehicle owners to
understand whether deployed systems were active during a crash, whether
they malfunctioned, or whether they helped mitigate damage or injury or
returned the car to a safe state in event of a malfunction.
---------------------------------------------------------------------------
\16\ 49 CFR 563.
---------------------------------------------------------------------------
Information sharing is included in the AV policy. However, the
policy fails to mention the public health argument for collecting data
from electronic devices in the car in the event of a crash. Acquiring
an understanding of what happens when systems perform as intended, fail
as expected, or fail in unexpected ways will yield valuable information
for manufacturers--some of whom have common suppliers. Further, in-
service data, as well as near miss and post-crash information sharing,
can help civil engineers and planners design better and safer roadways,
as well as help safety and health professionals design better
interventions to discourage risky driving or affect the behaviors of
other roadway users.
De-identified data sharing has been in existence in the aviation
industry for many years and proven highly successful. The Aviation
Safety Information Analysis and Sharing (ASIAS) system allows for
sharing of de-identified data across the aviation industry, making it
possible for the industry to identify trends and act on them. Analysis
of de-identified data will provide windows in to leading indicators in
the motor vehicle industry. Leading indicators are ``proactive,
preventative and predictive measures that monitor and provide current
information about the effective performance, activities and processes
of a . . . system that drive the identification and eliminate or
control of risks.'' \17\ The NSC Campbell Institute, a leader in
workplace safety, health and sustainability, states that tracking
leading indicators allows world-class safety organizations to make
further improvements to their safety records.\18\
---------------------------------------------------------------------------
\17\ Http://www.thecampbellinstitute.org/file/
download.php?id=20130925358263a8956de938e7c00
a2bbbb8413d.
\18\ Http://www.thecampbellinstitute.org/file/
download.php?id=2015092336b107f72d10a379134
af9249d3457ab.
---------------------------------------------------------------------------
The AV policy also suggests that auto manufacturers use EDRs to
gain a better understanding of how human operators engage with advanced
technology. This knowledge will allow manufacturers to be nimbler and
make adjustments in near real time for some systems based on what is
actually occurring in the driver's seat, rather than making changes
based on assumptions and estimations that must be accommodated in a
later model year. Collecting and sharing de-identified data about near
misses and other relevant problems could also help by aggregating
useful information for the automotive industry, allowing them to take
proactive steps based on leading indicators rather than waiting for a
crash or a series of crashes to occur. Finally, this data would be
useful to researchers and the safety community in analyzing the safety
benefits--and potential drawbacks--of these technologies as they
continue to mature.
While there are competing priorities regarding protecting personal
privacy and proprietary systems or designs, NSC believes that safety
should be the ultimate priority, and that other concerns need to be
accommodated to prioritize safety. NHTSA should facilitate data sharing
as widely as possible and require that manufacturers provide
accessible, standardized data to law enforcement, State highway safety
offices, investigators, insurers, and/or other relevant stakeholders.
american national standards institute standard
As important as it is for the average consumer to know and
understand the ADAS and automated technology, there is also work to be
done on this issue as it relates to the technology and its rollout to
commercial fleets. As such, NSC is taking a leading role working with
the American Society of Safety Engineers (ASSE) and a wide array of
experts in the automotive industry, technology sector, academia and
fleet management, to develop an ANSI (American National Standards
Institute) standard to address policies, procedures and management
processes that will assist in the control of risks and exposures
associated with the operation of autonomous fleet vehicles on public
thoroughfares.
national highway traffic safety administration resources and oversight
in the midst of changing technology
In response to reports of repeated incidents of unintended
acceleration in Toyota vehicles in 2009-2010, the National Research
Council Transportation Research Board conducted an investigation into
whether NHTSA (National Highway Traffic Safety Administration) had
reached the correct conclusion in its own investigation, as well as to
produce advice on how to best handle future issues involving the safe
performance of vehicle electronics. This research resulted in a report,
released on January 18, 2012.\19\
---------------------------------------------------------------------------
\19\ Http://www8.nationalacademies.org/onpinews/
newsitem.aspx?RecordID=13342.
---------------------------------------------------------------------------
The Research Council found that while NHTSA's decision to close its
investigation was justified, it was ``troubling that NHTSA could not
convincingly address public concerns about the safety of automotive
electronics.'' More specifically, the Research Council stated that ``to
respond effectively and confidently to claims of defects in the more
complex electronic systems . . . NHTSA will require additional
specialized technical expertise.'' While they acknowledged it was
unrealistic to expect NHTSA to hire and maintain these specialized
personnel in a constantly evolving field, they made a specific
recommendation that NHTSA establish a standing technical advisory panel
with members representing a wide array of technical expertise central
to the design, development, and safety assurance of automotive
electronics systems.
With technology advancing as quickly as it is, it is difficult to
keep up with advances without appropriate resources. Last year, NSC
supported the administration's NHTSA funding request of $1.1813
billion, which included $200 million for the Autonomous Vehicle
Development program. We encourage this Committee and Congress to fully
fund NHTSA, including requested investments in programs that will
support the development of ADAS and HAVs. We also recommend that NHTSA
consider how to best take advantage of existing knowledge by
establishing an advisory committee or similar mechanism to engage
experts in the field of advanced technology and automotive electronic
systems.
road to zero
On October 5, NSC, NHSTA, the Federal Highway Administration
(FHWA), and the Federal Motor Carrier Safety Administration (FMCSA)
announced the Road to Zero (RTZ) Coalition. RTZ is a partnership
initiative focused on dramatic reductions in roadway fatalities. Over
80 public and private organizations attended the announcement to learn
more about committing to a shared vision of zero fatalities on our
roadways. The first meeting of the coalition will be on December 15.
The purpose of the Road to Zero Coalition is to (1) encourage and
facilitate widespread implementation of countermeasures to reduce motor
vehicle crash deaths in the near term; (2) develop a scenario-based
vision for zero U.S. traffic deaths in the future; and (3) provide a
roadmap for policymakers and stakeholders to eliminate traffic deaths.
NSC is joined on the Steering Group for the Road to Zero Coalition
by the following organizations: Advocates for Highway and Auto Safety,
American Association of Motor Vehicle Administrators (AAMVA), American
Association of State Highway and Transportation Officials (AASHTO),
Commercial Vehicle Safety Alliance (CVSA), Governors Highway Safety
Association (GHSA), Institute of Transportation Engineers (ITE),
Insurance Institute for Highway Safety (IIHS), Intelligent Car
Coalition, International Association of Chiefs of Police (IACP),
Mothers Against Drunk Driving (MADD), National Association of State
Emergency Medical Services Officials (NASEMSO), National Association of
City Transportation Officials (NACTO), National Association of County
Engineers (NACE), and the Vision Zero Network.
On behalf of the Coalition, the NSC will administer a grant program
to support national non-profit organizations committed to roadway
safety programs that address the overlaps and gaps between roadway
users, vehicles and infrastructure. In addition, the Coalition will
look at engaging others in near term solutions and countermeasures to
reduce the death toll on our roadways. Finally, we will also provide
critical input for the development of a future community scenario with
zero traffic fatalities--an effort to look at the measures, programs
and technologies will be necessary to reach zero highway fatalities in
30 years and work back from there. NHTSA, FHWA, FMCSA, and NSC are
sponsoring the development of the scenario-based vision for zero
traffic deaths in the U.S. in a 30-year timeframe, and the RAND
Corporation has been retained to produce the scenario over the next 12-
18 months. I look forward to briefing this Committee and others in
Congress on the results of these activities and the efforts of the
Coalition to reach zero deaths on our roadways.
conclusion
Today, we have millions of drivers behind the wheel, spend millions
of dollars on education and enforcement campaigns, and still recognize
billions in economic loses as a result of crashes. In spite of safer
vehicle designs and record-setting seat belt use rates across the
Nation, operating a motor vehicle remains one of the deadliest things
we do on a daily basis.
NSC believes fully automated vehicles have the potential to save
lives and prevent injuries, but--as outlined above--there are several
key issues that policymakers must address. The AV policy begins this
discussion, and we applaud NHTSA and DOT for issuing it.
It will be a long time before HAVs replace our current fleet. The
transition will likely be messy as we deal with a complex and ever-
changing Human-Machine interface. There will be an evolution of the
existing technologies and perhaps a revolution when it comes to new and
different technologies. We need to be prepared for unanticipated
consequences and new failure modes.
Although we can imagine a future with automated vehicles, it will
be a long and winding road to get to the destination of zero fatalities
as a result of HAVs. We cannot afford to ignore the carnage on our
highways that is a national epidemic today. The U.S. trails other
industrialized countries in addressing highway deaths. Efforts like
Road to Zero will decrease fatalities today, tomorrow, and in the
future if we embrace proven countermeasures and accelerate deployment
of effective ADAS technologies.
NSC appreciates this Committee's leadership on vehicle technology
and safe roadway transportation. If safety for the traveling public is
the ultimate goal, advanced technology provides the most promising
opportunity to achieve that outcome, and will go a long way toward
reaching the goal of eliminating preventable deaths in our lifetime.
Senator Collins. Thank you very much, Ms. Hersman.
Mr. Brubaker.
STATEMENT OF HON. PAUL BRUBAKER, PRESIDENT AND CEO, THE
ALLIANCE FOR TRANSPORTATION INNOVATION
Mr. Brubaker. Chairman Collins, Ranking Member Reed, and
members of the subcommittee, on behalf of the Alliance for
Transportation Innovation, I would like to thank you for the
opportunity to share our views on the role of government in
integrating autonomous vehicles onto our Nation's roads.
The safety and social benefits that can be achieved by
replacing human drivers with modern sensors and computers are
too profound not to be encouraged through government policy.
Yet regulation has failed to keep pace with innovation, and we
see evidence of a significant gap in understanding of these
technologies within all levels of government. It's a gap that
must be bridged so regulators can better understand when it's
best to lead, follow, or move out of the way.
We believe the Administration's recently released AV policy
is a good first step in opening communication channels that can
help to bridge this gap. We will formally be commenting on this
policy by the end of the week.
But despite this disconnect industry and innovators are
still working hard to integrate the development and deployment
of autonomous systems and are investing billions of dollars
with safety top of mind. These technologies are already being
tested on our roads. In fact, just 3 weeks ago we saw the first
commercial delivery of a truckload of beer delivered by a self-
driving tractor trailer. Shortly, some in the auto industry
will be prepared to offer full self-driving capability to the
public, and government leaders must ensure that our current
regulatory approaches do not constrain the development of these
technologies. Lives depend on it.
To put a fine point on the subject, human drivers are
simply killers. Errors in human judgment while at the controls
of motor vehicles account for more than 90 percent of the car
crashes that last year killed over 34,000 people in the United
States and claimed 1.3 million lives around the globe.
Computers don't get tired. They don't text. They don't
drink and drive, and they don't get road rage. Networks,
sensors, maps, computing, artificial intelligence, machine
learning, and vehicle controls work in concert to detect
objects, analyze probabilities, consider options, make
decisions, and take actions, all at the speed of light, and all
with significantly greater reliability and efficiency than
humans.
This capability is improving rapidly, and government needs
to keep pace or risk undermining the promise of tomorrow by
applying the regulations of yesteryear. Government cannot
simply allow itself to be the obstacle to saving lives.
We hear a lot about these 34,000 fatal crashes and the fact
that 94 percent of them are caused by human error. But the
problem with statistics is they numb us with the fact that
we're talking about people, individual parents, siblings, sons
and daughters who are suddenly and cruelly taken away from us.
Sons like Leo Vagias--and his father Teddy is right here behind
me--and his best friend Sam Cali were killed in a very
preventable car crash just this past June in New Jersey in a
scenario that gets repeated too often and claims 96 lives a
day. It needs to stop, and we have a collective obligation--
industry, government, and citizens--to accelerate the
development, testing and deployment of life-saving self-driving
technologies.
But this journey to self-driving is complex, and
integrating these technologies onto our Nation's highways will
require unprecedented coordination and cooperation among
industry, government, and citizens. We must win over a
skeptical public by demonstrating that the underlying
technologies are effective, safe and secure, and because these
vehicles are going to operate on our public roadways, we've got
to ensure that government encourages their safe and responsible
deployment. Legal liability insurance issues must be addressed,
and economic impacts and potential social disruption must be
understood.
Given these complexities, ATI21.org believes that only
executive leadership with Congress on this issue will provide
the necessary national vision, goals, and direction required
across all sectors of our economy. That is why we recently
published our National Strategic Framework to Accelerate Life-
Saving Self-Driving Vehicles. We designed this document to
develop a pathway to successful integration of autonomous
vehicles into America's transportation system.
[The document link follows:]
Http://www.ati21.org/wp-content/uploads/2016/10/Final-ATI-Strategy-
Document.pdf.
One area of emphasis in that framework is data. We believe
that data is the key to speeding the safe deployment of these
technologies. That's why we're recommending the creation of a
National Self-Driving Data Repository. We envision that this is
a highly secured, trusted, opt-in data repository with a number
of positive incentives that can be leveraged in a way to
encourage participation, incentives like indemnification, and
rapid regulatory approvals. Such a repository will help us
better understand the level of safety and performance, as well
as enable us to identify real and emerging potential issues,
and inform regulators, industry and the public based on near
real-time data.
We have shared this national framework with the
subcommittee and with the Trump transition team, and we're
available to discuss this in detail.
[The article link follows:]
Https://www.washingtonpost.com/local/trafficandcommuting/the-next-
president-should-make-driverless-cars-a-white-house-priority-group-
says/2016/10/20/6c548212-9636-11e6-bc79-
af1cd3d2984b_story.html?utm_term=.3299c046b7aa.
The convergence of technology and transportation has the
potential to dramatically improve the safe and efficient
movement of people and goods. Over the next few years the Trump
Administration and Congress will have an unprecedented
opportunity to lead a cooperative effort between industry,
government, and the public that holds the promise of saving
lives. We look forward to working with you to advance this
agenda. Thank you.
[The statement follows:]
Prepared Statement of Hon. Paul R. Brubaker
Chairman Collins, Ranking Member Reed, and Members of the
Subcommittee, on behalf of the Alliance for Transportation Innovation
(ATI21.org), I would like to thank you for the opportunity to share our
views on integrating autonomous vehicles onto our Nation's roads.
ATI21.org is a not-for-profit created to accelerate the deployment
of technologies and innovation that can dramatically improve the safe
and efficient movement of people and goods. Our members are
associations, companies, and government and academic entities, as well
as individuals who are dedicated to our mission.
The need to accelerate the path to self-driving is not only
critical to our country's economic future, but also our technological
standing in the world. The potential safety and social benefits are
enormous, and we are on the cusp of reliably and safely replacing human
drivers with technology enabled mobility platforms--self-driving
vehicles--that will transform how we move people and goods.
Accelerating development and deployment of self-driving vehicles
holds the promise of saving tens of thousands of lives every year,
reducing crash-related injuries, and fundamentally transforming
personal mobility. As soon as we achieve full self-driving, distracted
and drunk driving will no longer claim lives; adult children will not
have to have the conversation with aging parents about taking away
their car keys; the poor will have access to much more convenient and
affordable transportation; and the physically and developmentally
disabled, including my soon to be adult son on the autism spectrum,
will have access to transportation options. There is no doubt that
safe, affordable, accessible, and convenient transportation is critical
for everyone's quality of life.
While the technologies to enable self-driving are being rapidly
developed, the desire to get these capabilities to market are
highlighting profound shortcomings in our national approach to
regulation. Specifically, the regulatory process is simply not keeping
up with the pace of innovation. While this is not unique to self-
driving, the rapid development of autonomous vehicles presents us with
an opportunity to revisit our regulatory approach and offer reforms
that are more suitable to the digital age. It is imperative. In this
case the cost of delay is measured in lost members of our families, of
pain and suffering, and the denial of economic, health and social
opportunity for the elderly, disabled and the impoverished.
That said, in the Obama Administration, Transportation Secretary
Anthony Foxx and Administrator Mark Rosekind of the National Highway
Traffic Safety Administration (NHTSA) get it. It is clear that the
administration has a profound understanding of the critical
technologies that have rapidly developed during the President's tenure
and have laid the groundwork for the next administration to propel us
into the self-driving future. The path will not be easy. There are
considerable and complex cultural and regulatory barriers that must be
overcome, and industry, researchers, the public, and government at all
levels will need to engage and collaborate.
In late September the Department of Transportation issued its much
anticipated autonomous vehicle (AV) policy guidance in what we believe
represented a well-intentioned and thoughtful first-step toward
ensuring that the Department, and NHTSA specifically, is heading toward
a more responsive regulatory approach that is more suitable for
accelerating the safe deployment of highly autonomous vehicles on our
Nation's roads.
While these initial guidelines are not perfect, ATI21.org believes
the Department of Transportation should be commended for recognizing a
number of critical issues that must become priorities for industry and
government to safely deploy autonomous vehicles. We also commend the
Department for its clear willingness to engage with both the
traditional auto industry and technology sectors. We believe such an
open and productive dialogue with all interested parties is the first
step to evolving toward a much more responsive and effective regulatory
framework than the existing model.
Specifically, there are several positive provisions included in
this initial version of the AV policy. First, the Department's 15-point
safety assessment covers all the major key areas that are critical to
the safe design of highly autonomous vehicles. Second, the Department
rightfully asserted its appropriate and exclusive role as the Nation's
vehicle safety regulator and offered a model policy that can help guide
States in formulating appropriate regulatory frameworks. Lastly, the
Department was quite clear that it intended this process to be
iterative and recognized the need to evolve and even adopt new
regulatory models that could keep pace with the innovation cycle. We
believe these are all excellent first steps in developing an
appropriate regulatory framework.
There are however two sections of the AV Policy document that gave
us pause. Specifically, the section on the use of existing regulatory
tools and the examples given as possible future regulatory approaches
is no better and potentially worse than the existing tools.
Although the policy includes language about revising and
streamlining processes related to interpretations and waivers, we
believe the underlying Federal Motor Vehicle Safety Standards (FMVSS),
while arguably effective for traditional motor vehicle design approval,
will prove to be an unworkable foundation for accelerating the design
and deployment of highly autonomous vehicles. As was highlighted in the
March 2016 report by the Volpe National Transportation Research Center,
there are considerable challenges in applying these standards to SAE
level 4 and 5 autonomous vehicles.\1,\\2\ For example, the FMVSS makes
more than 250 individual references to human drivers, and eliminating
traditional design features such as steering wheels and pedals, which
can be anticipated in level 4 and 5 vehicles, would violate roughly a
third of the standards and half of the Series 100 crash avoidance
requirements.
---------------------------------------------------------------------------
\1\ John A. Volpe National Transportation Systems Center, Review of
Federal Motor Vehicle Safety Standards (FMVSS) for Automated Vehicles,
Preliminary Report-March 2016, prepared for the Intelligent
Transportation Systems Joint Program Office, NHTSA.
\2\ Levels of driving automation are defined in the SAE
International Standard J3016. It identifies six levels of driving
automation from ``no automation'' (0) to ``full automation'' (5), with
level 4 being ``high'' and Level 5 being ``full.''
---------------------------------------------------------------------------
While we commend the Department for beginning what we believe is a
long overdue conversation on the need to create a much more responsive
and timely regulatory approach, we believe the examples offered such as
pre-approval of designs based on the FAA model or the FDA medical
device approach are neither particularly efficient nor applicable or
fitting.
Data is the key to reforming the regulatory process in a manner
that will allow the creation of a regulatory approach that is rapid,
efficient, and effective. Specifically, a central repository for
collecting, storing, and analyzing all operational and testing data
from across the industry is necessary to create a responsive regulatory
environment. Neither the NHTSA nor any existing industry group is
positioned to perform this task. Consequently, we are recommending the
establishment of an independent, federally chartered organization to
securely and confidentially collect and analyze all operational and
test data from across industry that will include simulation data. By
collecting real-time data and applying modern analytical methods, we
are confident that, if properly established, this organization can
rapidly inform industry and NHTSA of real and potential issues. We
recommend that this effort be funded from existing dollars. We would be
happy to consult with the Subcommittee on the specific details of this
recommendation.
The second area we would like to highlight in our testimony this
morning is our call for a National Strategic Framework to Advance Life-
Saving Self-Driving Vehicles. We created this document calling for
presidential level leadership on self-driving.
Accelerating the path to self-driving is an enormously complex
undertaking that will go beyond the responsibility of the U.S.
Department of Transportation. While there are major roles and policy
levers that both the Federal Government as a whole and U.S. Department
of Transportation can take to accelerate deployment and ensure that
only safe vehicles make it to market, it is critical that all levels of
government work together, that industry and the research community
continues to drive innovation and has a path to market, and perhaps
most importantly that citizens are informed, engaged and welcoming of
what will prove to be a major cultural shift.
Not unlike past transitions to transportation-related innovation,
success will depend on high levels of cooperation and engagement across
government, industry, and the public. This was true of steamships,
railroads, motor vehicles and aviation. The disruptive transition to
self-driving will require similar coordination, cooperation, and
support.
As is true with all disruptive innovation throughout history, we
can expect to experience both positive and negative effects as we begin
integrating autonomous vehicles onto the Nation's roads. It is critical
that we thoughtfully consider the obvious consequences, both good and
bad, that will undoubtedly arise from our transition from what we refer
to as the ``crash economy'' to a new mobility paradigm.
As I mentioned earlier, the major reason for making this transition
is to save lives and reduce injuries. Last year, over 35,000 Americans
died in car crashes. This year that number is expected to climb and may
exceed 40,000. Millions more are injured including many with life-
changing debilitating injuries. We know that crashes cost the U.S.
economy almost $1 trillion annually. Let us be clear, we would not
tolerate such cost and carnage in any other mode of transportation, and
now, the tools are arriving to make these tragedies preventable.
Autonomous features and ultimately self-driving vehicles hold
promise to dramatically reduce fatalities and injuries as NHTSA
estimates that more than 94 percent of these crashes are the result of
human error. We simply should not allow more than 90 of our fellow
citizens to die every day when we can create the conditions to prevent
it. The sooner we fully integrate autonomous vehicles onto our
highways, the sooner we will realize a dramatic reduction in the number
of preventable fatalities and injuries.
Getting this transition right will not be easy. Moving from the
crash economy to a dramatically safer, accessible, affordable, and
convenient mobility paradigm is a complex undertaking requiring
unprecedented cooperation and coordination among industry,
associations, and research and government institutions at all levels.
We believe such an effort will require presidential leadership and
that the incoming administration and Congress should work in concert to
create the conditions necessary to accelerate the development, testing,
and deployment of these capabilities as well as proactively address
some of the anticipated consequences of transition.
At ATI21.org, we announced the publication of our initial framework
for accelerating the deployment of self-driving vehicles last month.
The framework is not a prescriptive list of recommended actions but
rather an outline of key challenges areas that must be addressed in
order to accelerate the integration of autonomous vehicles on our
Nation's highways.
Between now and Inauguration Day, we will be soliciting reaction to
the challenge areas and hosting a series of expert panels to explore
and produce recommendations on each challenge area. We intend to update
the framework based on panel input and host a final review and input
session for the entire document on January 4th with the Consumer
Technology Association in advance of the Consumer Electronics Show.
In short, ATI21.org is calling for a National Strategy to Advance
Self-Driving (NSAS) and urging the next president to issue an executive
order that would describe the challenges and create a Program
Management Office (PMO) within the Office of Science and Technology
Policy (OSTP). We recommend funding the activity as a government-wide
technology initiative through existing Office of Management and Budget
(OMB) authority under a provision in the Clinger-Cohen Act. The PMO
would operate under a 4-year charter to collaborate with relevant
stakeholders, assemble experts, develop outcome-focused and actionable
strategies, and identify funding streams to address each of the
challenge areas.
We have identified the following challenge areas:
--Earning public trust
--Increasing confidence in self-driving technology
--Ensuring robust cybersecurity
--Developing standards and regulations that encourage self-driving
--Creating the legal, liability, and insurance framework for
accelerating the deployment of self-driving vehicles
--Adopting reasonable data and privacy policies
--Accelerating the transition to a fully self-driving fleet
--Understanding and planning for economic disruption and labor
transition
As we say in the framework, accelerating the path to self-driving
will not be easy. The incoming president by leading us though this
ambitious effort early in the new term can ensure dramatic progress
toward creating a new mobility paradigm that will create safe,
convenient, resilient, and accessible transportation options to all
Americans regardless of their ability, age, or economic condition.
ATI21.org believes this is a transportation legacy that we can be proud
to leave to future generations.
In the coming months, we look forward to working with the
Subcommittee and Congress to work on creating the conditions necessary
to speed integration of autonomous vehicles on the Nation's highways.
Thank you. I look forward to answering your questions.
Senator Collins. Thank you.
Dr. Kalra.
STATEMENT OF DR. NIDHI KALRA, CO-DIRECTOR, RAND CENTER
FOR DECISION MAKING UNDER UNCERTAINTY
Dr. Kalra. Chairman Collins, Ranking Member Reed, members
of the subcommittee, thank you for the opportunity to testify
today on the future of autonomous vehicles.
My name is Nidhi Kalra, and I'm an Information Scientist at
the RAND Corporation. RAND is a non-profit, non-partisan
research institution committed to improving public policy
through objective research and analysis. I have a Ph.D. in
robotics and have been studying autonomous vehicles for 10
years. And in the interest of full disclosure, my spouse is the
co-founder of an autonomous vehicle start-up in Silicon Valley,
though his work has no bearing on my testimony, or vice-versa.
Now, there are three issues I'd like to discuss today, the
safety of autonomous vehicles, their benefits for mobility, and
the urgency of addressing distortions in the transportation
market so we can harness their benefits and mitigate their
drawbacks.
So first let's talk about safety. Will autonomous vehicles
be safe before they're allowed on the road for consumer use? We
may not know. Real-world driving is currently the only method
of demonstrating their safety. But suppose a fleet of
autonomous vehicles has a 20 percent lower fatality rate than
human drivers. They would have to be driven 5 billion miles to
prove it, and a fleet of test vehicles, 100 test vehicles would
have to be driven 225 years to cover this distance. It's
impossible if we ever want them on the road for consumer use.
But now let's consider two things. First, Americans as a
whole drive more than this every day. We drive 8 billion miles
every day. So once autonomous vehicles are widely deployed for
consumer use, we may know very quickly how safe they are. And
consider that autonomous vehicles will improve over time. So
when a human driver makes a mistake, only that person, at best,
is able to learn from the experience, but that's not with
autonomous vehicles. When one vehicle makes a mistake or has an
experience on the road, that information can be used to improve
the entire fleet.
So experience may be one of the most important tools for
both demonstrating and improving autonomous vehicle safety. So
I think an important question is how do we enable autonomous
vehicles to get experience while mitigating the safety risks
they pose? And I suggest two things--strategic pilot studies
and data sharing--that can help.
Pilot studies could start with real-world testing in lower-
risk conditions like operating vehicles in well-maintained
areas and at low speeds, and then could be expanded as safety
is demonstrated.
As for data sharing, developers are already using the
experiences of one vehicle to improve their fleet, but data
sharing across the industry could mean that the experiences of
every vehicle can improve the entire industry.
Second, let's talk about mobility. Autonomous vehicles
could improve mobility for millions of Americans who are
elderly, young, have disabilities, or live in poverty. But
policies may be needed to make them affordable, available, and
accessible. Policymakers can incentivize developers to bring
these technologies to those markets sooner than they might
otherwise, or to integrate them with transit and para-transit.
Policymakers may need to work with developers to help make sure
they're accessible, meaning that they comply with ADA
guidelines and use different forms of payment, for example.
And while the cost of shared autonomous vehicles is
expected to be low, policymakers could extend transit and para-
transit reduced fares to these other technologies to make them
all the more affordable for these vulnerable populations.
Third, let's talk about market distortions. Autonomous
vehicles could also significantly affect traffic congestion,
energy pollution and land use, maybe for the better, but maybe
for the worse. So to harness their benefits and mitigate their
drawbacks, policymakers should address the existing distortions
in the transportation market that lead to undesirable outcomes
and subsidize desirable outcomes.
Now, there are many policy options to address distortions,
like strengthening the auto insurance market to encourage
safety, implementing congestion pricing to reduce congestion,
and offering rebates for fuel-efficient vehicles to reduce
pollution.
Now, I can't today give each policy the discussion it
warrants, but now is the time to address these distortions,
because autonomous vehicles could improve our transportation
system tremendously if the right market signals are in place.
And there is currently a window of opportunity to send those
new market signals because consumer expectations about
autonomous vehicle performance and cost aren't yet set.
So, let me conclude. In the 1990s, we couldn't forecast the
impact that the Internet has now had. Similarly, today we can't
forecast the impact that autonomous vehicles will have on our
future, but we can shape that future, and I would like to leave
you with two recommendations.
First, as I noted, widespread experience may be the key to
figuring out both how safe autonomous vehicles are and how they
get better. So it's important to get them that experience as
safely as possible.
And second, correcting market distortions will be critical
to encouraging better outcomes for our transportation system
and for our society.
Chairman Collins, Ranking Member Reed, and members of the
subcommittee, thank you for allowing me to appear before you
today, and I look forward to your questions.
[The statement follows:]
Prepared Statement of Dr. Nidhi Kalra
Chairman Collins, Ranking Member Reed, and distinguished members of
the subcommittee, my name is Nidhi Kalra of the RAND Corporation.\1\
Thank you for the opportunity to testify on important emerging
opportunities and risks related to autonomous vehicles. Autonomous
vehicles have the potential to change transportation profoundly, in the
United States and around the world. There is much opportunity for
improvement, but also potential for added risks and harms. How
autonomous vehicles ultimately shape our future is not foretold; it
depends on many technology and policy choices we make today.
---------------------------------------------------------------------------
\1\ The RAND Corporation is a research organization that develops
solutions to public policy challenges to help make communities
throughout the world safer and more secure, healthier and more
prosperous. RAND is nonprofit, nonpartisan, and committed to the public
interest.
---------------------------------------------------------------------------
Today, I would like to discuss three important questions about the
future of autonomous vehicles and how policies can shape it. First,
will autonomous vehicles be safe before they are allowed on the road
for consumer use? Second, how can autonomous vehicles improve mobility
for Americans who currently may have limited mobility? Third, what
mechanisms can help realize the benefits and mitigate the drawbacks of
autonomous vehicles? I will focus most of my remarks on fully
autonomous vehicles--those that can operate without a human driver some
or all of the time--rather than on vehicles that require continuous
human oversight.
will autonomous vehicles be safe before they are allowed on the road
for consumer use?
In the United States, roughly 32,000 people are killed and more
than 2 million are injured in motor vehicle crashes every year.\2\
Although safety has generally improved over the past several decades,
2015 saw 35,000 road fatalities, the largest increase in fatalities in
this country in more than 50 years. This occurred partly because
Americans drove more and partly because they drove worse.
---------------------------------------------------------------------------
\2\ Bureau of Transportation Statistics, Motor Vehicle Safety Data,
Table 2-17, Washington, D.C.: Research and Innovative Technology
Administration, U.S. Department of Transportation, 2015.
---------------------------------------------------------------------------
U.S. motor vehicle crashes as a whole can pose enormous economic
and social costs--more than $800 billion in a single year.\3\ And more
than 90 percent of crashes are caused by human errors,\4\ such as
driving too fast and misjudging other drivers' behaviors, as well as
alcohol impairment, distraction, and fatigue.
---------------------------------------------------------------------------
\3\ Lawrence Blincoe, Ted R. Miller, Eduard Zaloshnja, and Bruce A.
Lawrence, The Economic and Societal Impact of Motor Vehicle Crashes
2010 (Revised), Washington, D.C.: National Highway Traffic Safety
Administration, DOT HS 812 013, 2014, revised May 2015.
\4\ National Highway Traffic Safety Administration, Traffic Safety
Facts, A Brief Statistical Summary: Critical Reasons for Crashes
Investigated in the National Motor Vehicle Crash Causation Survey,
Washington, D.C.: National Center for Statistics and Analysis, U.S.
Department of Transportation, DOT HS 812 115, February 2015.
---------------------------------------------------------------------------
Autonomous Vehicles Present Benefits and Risks to Safety
Autonomous vehicles have the potential to significantly mitigate
this public safety crisis by eliminating many of the mistakes that
human drivers routinely make.\5\ To begin with, autonomous vehicles are
never drunk, distracted, or tired; these factors are involved in 41
percent, 10 percent, and 2.5 percent, respectively, of all fatal
crashes.\6\ Autonomous vehicles could perform better than human drivers
because of better perception (e.g., no blind spots), better
decisionmaking (e.g., more-accurate planning of complex driving
maneuvers), and better execution (e.g., faster and more-precise control
of steering, brakes, and acceleration).
---------------------------------------------------------------------------
\5\ James M. Anderson, Nidhi Kalra, Karlyn D. Stanley, Paul
Sorensen, Constantine Samaras, and Oluwatobi A. Oluwatola, Autonomous
Vehicle Technology: A Guide for Policymakers, Santa Monica, Calif.:
RAND Corporation, RR-433-2-RC, 2014; and Daniel J. Fagnant and Kara
Kockelman, ``Preparing a Nation for Autonomous Vehicles: Opportunities,
Barriers and Policy Recommendations,'' Transportation Research Part A:
Policy and Practice, Vol. 77, July 2015, pp. 167-181.
\6\ National Highway Traffic Safety Administration, Traffic Safety
Facts: Crash Stats, Washington, D.C.: National Center for Statistics
and Analysis, DOT HS 811 449, March 2011; Bureau of Transportation
Statistics, Occupant and Non-Motorist Fatalities in Crashes by Number
of Vehicles and Alcohol Involvement (Updated July 2014), Table 2-20,
Washington, D.C.: U.S. Department of Transportation, 2014; and U.S.
Department of Transportation, Fact Sheet: Enhanced Mobility of Seniors
and Individuals with Disabilities Section 5310, Washington D.C., 2015.
This does not mean that 53.5 percent of all fatal crashes are caused by
these factors because a crash may involve, but not be strictly caused
by, one of these factors, and because more than one of these factors
may be involved in a single crash.
---------------------------------------------------------------------------
However, autonomous vehicles might not eliminate all crashes. For
instance, inclement weather and complex driving environments pose
challenges for autonomous vehicles, as well as for human drivers, and
autonomous vehicles might perform worse than human drivers in some
cases.\7\ There is also the potential for autonomous vehicles to pose
new and serious crash risks--for example, crashes resulting from cyber
attacks.\8\ Clearly, autonomous vehicles present both enormous
potential benefits and potential risks to transportation safety.
---------------------------------------------------------------------------
\7\ Lee Gomes, Hidden Obstacles for Google's Self-Driving Cars:
Impressive Progress Hides Major Limitations of Google's Quest for
Automated Driving, Massachusetts Institute of Technology, August 28,
2014.
\8\ Anderson et al., 2014.
---------------------------------------------------------------------------
When the National Highway Traffic Safety Administration released
much-anticipated guidelines intended to outline best practices for
autonomous vehicle safety, many looked to that guidance to answer the
key question: Will autonomous vehicles be safe? I believe the answer is
``maybe.'' Answering the question requires considering two issues.
First, how should autonomous vehicle safety be measured, and second,
what threshold of safety should be required before autonomous vehicles
are made publicly available? In essence, what test do autonomous
vehicles have to take and what constitutes a passing grade? Both are
genuinely open questions, so it is understandable that Federal
guidelines have not yet answered them.
There Is No Proven, Feasible Way to Determine Autonomous Vehicle Safety
There are no road tests that could demonstrate how safe an
autonomous vehicle is--there are too many conditions and scenarios to
test them all. (A road test that a person takes at the Department of
Motor Vehicles also does not prove that he or she will be a good
driver; rather, the road test determines whether the person can perform
a specific set of driving skills under regular traffic situations.
While this type of evidence is viewed as adequate for licensing human
drivers, it is not generally viewed as adequate for robot drivers.)
A logical alternative is to test-drive autonomous vehicles
extensively in real traffic and observe their performance before making
them commercially available. Although this is a helpful first step, it
is not sufficient to prove safety. Even though the number of crashes,
injuries, and fatalities from human drivers is high, the rate of these
failures is low in comparison with the number of miles that people
drive. Americans drive nearly 3 trillion miles every year.\9\ The
35,092 fatalities and 2.44 million injuries in 2015 correspond to a
failure rate of 1.12 fatalities and 78 injuries per 100 million miles
driven. Given that current traffic fatalities and injuries are rare
events compared with vehicle miles traveled, fully autonomous vehicles
would have to be driven hundreds of millions of miles and sometimes
hundreds of billions of miles to demonstrate their reliability in terms
of fatalities and injuries. Under even aggressive testing assumptions,
existing fleets would take tens and sometimes hundreds of years to
drive these miles--an impossible proposition if the aim is to
demonstrate their performance prior to releasing them on the roads for
consumer use.\10\ And, in the meantime, human drivers would continue to
cause avoidable crashes and enormous harms to people and property.
---------------------------------------------------------------------------
\9\ Bureau of Transportation Statistics, 2015.
\10\ Nidhi Kalra and Susan M. Paddock, Driving to Safety: How Many
Miles of Driving Would It Take to Demonstrate Autonomous Vehicle
Reliability? Santa Monica, Calif.: RAND Corporation, RR-1478-RC, 2016.
---------------------------------------------------------------------------
Developers of this technology and third-party testers need to
develop innovative methods of demonstrating safety and reliability.
These methods may include but are not limited to accelerated
testing,\11\ virtual testing and simulations,\12\ mathematical modeling
and analysis,\13\ scenario and behavior testing,\14\ and pilot
studies,\15\ as well as extensive focused testing of hardware and
software systems. This is a rapidly growing area of research and
development. There are promising ideas but no demonstrated and accepted
methods of proving safety. In sum, no one yet knows how autonomous
vehicles should be tested. It is therefore reasonable that the current
Federal guidelines have not specified a test either.
---------------------------------------------------------------------------
\11\ Wayne B. Nelson, Accelerated Testing: Statistical Models, Test
Plans, and Data Analysis, Hoboken, N.J.: John Wiley & Sons, 2009.
\12\ Suren Chen and Feng Chen, ``Simulation-Based Assessment of
Vehicle Safety Behavior under Hazardous Driving Conditions,'' Journal
of Transportation Engineering, Vol. 136, No. 4, 2010, pp. 304-315;
Siddartha Khastgir, Stewart A. Birrell, Gunwant Dhadyalla, and Paul A.
Jennings, ``Development of a Drive-In Driver-in-the-Loop Fully
Immersive Driving Simulator for Virtual Validation of Automotive
Systems,'' paper presented at IEEE 81st Vehicular Technology
Conference, Glasgow, Scotland, May 11-14, 2015; and Stephanie Olivares,
Nikolaus Rebernik, Arno Eichberger, and Ernst Stadlober, ``Virtual
Stochastic Testing of Advanced Driver Assistance Systems,'' in Tim
Schulze, Beate Muller, and Gereon Meyer, eds., Advanced Microsystems
for Automotive Applications 2015: Smart Systems for Green and Automated
Driving, New York: Springer, 2015.
\13\ Khashayar Hojjati-Emami, Balbir Dhillon, and Kouroush Jenab,
``Reliability Prediction for the Vehicles Equipped with Advanced Driver
Assistance Systems (ADAS) and Passive Safety Systems (PSS),''
International Journal of Industrial Engineering Computations, Vol. 3,
No. 5, 2012, pp. 731-1742; and R. Kianfar, P. Falcone, and J.
Fredriksson, ``Safety Verification of Automated Driving Systems,'' IEEE
Intelligent Transportation Systems Magazine, Vol. 5, No. 4, Winter
2013, pp. 73-186.
\14\ California Department of Motor Vehicles, Express Terms Title
13, Division 1, Chapter 1 Article 3.7--Autonomous Vehicles, 2015; and
Michael Sivak, and Brandon Schoettle, Should We Require Licensing Tests
and Graduated Licensing for Self-Driving Vehicles? University of
Michigan: Transportation Research Institute, Technical Report UMTRI-
2015-33, 2015.
\15\ ANWB, Experiments on Autonomous and Automated Driving: An
Overview 2015, 2015.
---------------------------------------------------------------------------
There Is No Consensus on How Safe Autonomous Vehicles Should Be
The second issue of how safe autonomous vehicles should be is worth
considering, even if their degree of safety cannot yet be fully proven.
Some will insist that anything short of totally eliminating risk is a
safety compromise. They might feel that it is acceptable if humans make
mistakes, but not if machines do. But, again, waiting for autonomous
vehicles to operate perfectly misses opportunities to save lives
because it leaves far-from-perfect human drivers behind the wheel.
It seems sensible that autonomous vehicles should be allowed on
America's roads when they are judged safer than the average human
driver, allowing more lives to be saved and sooner while still ensuring
that autonomous vehicles do not create new risks. An argument can be
made that autonomous vehicles could be allowed even when they are not
as safe as average human drivers if developers can use early deployment
as a way to rapidly improve the vehicles. The vehicles could become at
least as good as the average human sooner than they would otherwise,
and thus save more lives overall.
The lack of consensus on this point is not a failure of sound
thinking. It is not a failure at all, but rather a genuine expression
of Americans' different values and beliefs when it comes to humans
versus machines. It is therefore reasonable that the Federal guidelines
also do not draw a line in the sand.
While these are difficult decisions, our differences in values and
beliefs can be informed by thinking not only about safety today but
also about the arc of safety in the coming decades. Our discourse on
the question of how safe the vehicles need to be focuses on the safety
of autonomous vehicles at the time that they are first introduced for
consumer use. But this thinking should be expanded to consider the
evolution of autonomous vehicle safety over time, not just at the start
of vehicle deployment. When a human driver makes a mistake on the road,
typically only that individual can learn from that experience to
improve his or her driving habits. The other drivers on the road are
largely unaffected. This is not the case with autonomous vehicles,
which can use experience and learning to improve performance, not just
of the individual vehicle but of the entire fleet. This is because,
when an algorithm or software is updated and improved for one vehicle,
it can be updated for all vehicles. For this reason, experience may be
one of the most important tools for improving autonomous vehicle safety
and, by extension, transportation safety.
Policymakers Can Promote Autonomous Vehicle Safety
This raises an important question: How do we enable autonomous
vehicles to improve as quickly as possible while lowering the risks
they pose? There are several tactics policymakers could consider to
accelerate autonomous vehicles' improvement.
A first step is to conduct real-world but lower-risk pilot studies
of autonomous vehicles. Risk can be lowered first by operating
autonomous vehicles in conditions in which crashes are less likely.
This can include limiting autonomous vehicle pilots to areas with less-
complex terrain, to routes that are well maintained and easier to
navigate, to nondangerous weather conditions, or to some combination of
these controls. It can also include educating communities about safe
behavior in and around autonomous vehicles. Furthermore, risk can be
lowered by designing and operating vehicles so that when crashes occur,
the consequences of the crash to passengers and bystanders are fewer.
This could be accomplished by limiting vehicle speed, ensuring that all
pilot-study passengers wear seatbelts, and so forth. These
strategically limited pilot studies can then be expanded as safe
operation of autonomous vehicles is demonstrated.
A second consideration is the role of sharing driving data across
the industry and with policymakers. Autonomous vehicle developers
already use the experiences of a single vehicle to improve the safety
of their individual fleets. This improvement could occur even faster if
the experiences of each vehicle in each fleet could be used across all
developers to improve the entire industry. There are certainly
nontrivial concerns about protecting trade secrets, but these concerns
could be addressed and must be balanced with the societal need for safe
autonomous vehicle technology.
In sum, it may not be possible to know what the safety risk of
autonomous vehicles is, and Americans may not agree on what it should
be. All the same, there are ways of lowering that risk that deserve
careful consideration.
how can autonomous vehicles improve mobility for americans who
currently may have limited mobility?
For almost all Americans, the ability to get around is essential
for living a rich, productive, and healthy life: being able to get to a
place of work, to visit friends and family, to access healthcare and
other services, to participate in civic activities, and to be connected
to the external world in all other ways. Even with the increasing
ability to interact and transact online, the importance of mobility in
today's world remains vital. Despite its importance, many Americans
have limited, and sometimes very limited, mobility as a result of
advanced age, disabilities, or lack of means. Whatever the reason,
limited mobility has significant negative consequences. Autonomous
vehicles could help.
Autonomous Vehicles Could Help Many Older Americans Who Face Limited or
Declining Mobility
The experiences of older Americans, especially those over 75, are
emblematic of the challenges of limited mobility. The number of
Americans 65 and older will increase from 48 million in 2015 (15
percent of today's population) to 74 million in 2030 (23 percent of the
population). The number of Americans 75 and older will increase from 20
million in 2014 (6 percent of today's population) to 35 million in 2030
(10 percent of the population.) \16\ Older Americans are living longer
and working longer than ever before. The labor force participation of
those over 65 is expected to be 21.7 percent in 2024, up from 12.4
percent in 1994.\17\
---------------------------------------------------------------------------
\16\ U.S. Census Bureau, Projections of the Population by Sex and
Age for the United States: 2015 to 2060 (NP2014-T9), spreadsheet,
December 2014. As of November 8, 2016: http://www.census.gov/
population/projections/files/summary/NP2014-T9.xls.
\17\ Bureau of Labor Statistics, ``Civilian Labor Force
Participation Rate by Age, Gender, Race, and Ethnicity,'' web page,
December 2015. As of November 8, 2016: http://www.bls.gov/emp/
ep_table_303.htm.
---------------------------------------------------------------------------
Driving is important to their quality of life. Of adults over 65,
80 percent live in car-dependent areas and 90 percent say they intend
to age in place.\18\ Eighty-five percent of adults aged 65 to 84 hold
licenses, and almost 60 percent of adults over 85 hold licenses.\19\
---------------------------------------------------------------------------
\18\ David Dudley, ``The Driverless Car Is (Almost) Here,'' AARP
The Magazine, December 2014/January 2015. As of November 8, 2016:
http://www.aarp.org/home-family/personal-technology/info-2014/google-
self-driving-car.html.
\19\ Policy and Governmental Affairs Office of Highway Police
Information, ``Distribution of Licensed Drivers--2014 By Sex and
Percentage in Each Age Group and Relation to Population,'' web page,
U.S. Department of Transportation, September 2014. As of November 8,
2016: https://www.fhwa.dot.gov/policyinformation/statistics/2014/
dl20.cfm.
---------------------------------------------------------------------------
Driving is risky for many older Americans. A recent study found
that, when compared with drivers aged 55 to 64, drivers over 75 were
more than 2.5 times as likely to die in a car crash, and drivers over
85 were almost four times as likely.\20\ This is due both to increased
likelihood of getting into crashes and greater vulnerability to
injuries.
---------------------------------------------------------------------------
\20\ AAA Foundation for Traffic Safety, ``Drivers Over 65 Almost
Twice as Likely as Middle-Aged Drivers to Die in Car Crashes, According
to AAA Foundation Study,'' February 18, 2004. As of November 8, 2016:
https://www.aaafoundation.org/sites/default/files/DriversOver65.pdf.
---------------------------------------------------------------------------
But giving up driving has risks as well. Driving cessation almost
doubles the risk of increased depressive symptoms and is correlated
with (though not strictly a cause of) cognitive, social, and physical
declines and higher rates of entry into long-term care.\21\
---------------------------------------------------------------------------
\21\ Stanford Chihuri, Thelma J. Mielenz, Charles J. DiMaggio,
Marian E. Betz, Carolyn DiGuiseppi, Vanya C. Jones, and Guohua Li,
``Driving Cessation and Health Outcomes in Older Adults,'' American
Geriatric Society, Vol. 64, 2016, pp. 332-341.
---------------------------------------------------------------------------
Geography can further affect mobility. Approximately 18 percent of
the rural population is 65 years or older, compared with 13.5 percent
in non-rural areas.\22\ Compared with their counterparts in urban
areas, older adults in rural areas must take longer trips for
healthcare and other services and have fewer alternatives to
driving.\23\
---------------------------------------------------------------------------
\22\ U.S. Census Bureau, ``Percent of the Total Population Who Are
65 Years and Over--United States--Urban/Rural and Inside/Outside
Metropolitan and Micropolitan Area,'' American FactFinder, 2014. As of
November 8, 2016: http://factfinder.census.gov/bkmk/table/1.0/en/ACS/
14_5YR/GCT0103.US26
\23\ J. E. Burkhardt, A. T. McGavock, C. A. Nelson, and C. G. B.
Mitchel, Improving Public Transit Options for Older Persons Transit
Cooperative Research Program, Washington D.C.: Transport Research
Board, 2002.
---------------------------------------------------------------------------
Autonomous vehicles offer a promising solution. Fully automated
vehicles that do not require human intervention would allow many older
adults to travel by car, without having to drive. It could increase
their mobility, with all of the associated social and economic
benefits, while mitigating much of the safety risk. This, in turn, may
allow more people to age in place, remaining in their homes for much
longer than they might otherwise be able to.
Autonomous Vehicles Could Improve Mobility for Many Others
Older adults are just one group of Americans that could benefit
from increased mobility from autonomous vehicles. Many people with
disabilities, young people, and people living in poverty face mobility
challenges that could be alleviated by autonomous vehicles.
In 2010, 56.7 million individuals (18.7 percent of the population)
identified as having a disability.\24\ Only 65 percent of individuals
with disabilities drive, compared with 88 percent of individuals
without disabilities.\25\ In spite of the Americans with Disabilities
Act, which mandates that transit authorities operating a fixed route
system provide paratransit or a comparable service to individuals with
a disability,\26\ individuals with disabilities often have limited
mobility because of a lack of availability or access to services. One
survey showed that 12 percent of persons with disabilities reported
having a harder time obtaining the transportation they need to be
independent, compared with 3 percent of others, the top two reasons
being no or limited public transportation (33 percent) and not having a
car (26 percent).\27\
---------------------------------------------------------------------------
\24\ Matthew W. Brault, Americans with Disabilities: 2010, U.S.
Census Bureau, July 2012.
\25\ U.S. Department of Transportation, Freedom to Travel,
Washington D.C.: Bureau of Transportation Statistics, 2003.
\26\ U.S. Department of Justice, Information and Technical
Assistance on the Americans with Disabilities Act, 2016.
\27\ U.S. Department of Transportation, 2003.
---------------------------------------------------------------------------
There are also 25 million young Americans between the ages of 12
and 17 who have mobility needs but are not yet old enough to drive or
are novice drivers.\28\ Getting to school and academic enrichment
opportunities, social and extracurricular activities, and even first
jobs can be a challenge. Many depend on buses (principally to school)
or their parents--or forgo travel. For many working parents, there is a
trade-off between supporting their own and their children's mobility
needs.
---------------------------------------------------------------------------
\28\ Federal Interagency Forum on Child and Family Statistics,
``POP1 Child Population: Number of Children (in millions) Ages 0-17 in
the United States by Age, 1950-2015 and Projected 2016-2050,'' 2016. As
of November 8, 2016: http://www.childstats.gov/americaschildren/tables/
pop1.asp.
---------------------------------------------------------------------------
Americans living in poverty also face mobility challenges. About
43.1 million people (13.3 percent of the population) live in
poverty.\29\ This includes older adults and many individuals with
disabilities. In 2014, 10 percent of older adults and 28.5 percent of
individuals with a disability had a yearly income below the poverty
line.\30\
---------------------------------------------------------------------------
\29\ Bernadette D. Proctor, Jessica L. Semega, and Melissa A.
Kollar, Income and Poverty in the United States: 2015, U.S. Census
Bureau, September 2016. As of November 8, 2016: http://www.census.gov/
library/publications/2016/demo/p60-256.html.
\30\ C. DeNavas-Walt and B. D. Proctor, Income and Poverty in the
United States: 2014, Washington D.C.: U.S. Census Bureau, 2015.
---------------------------------------------------------------------------
About 24 percent of households below the poverty line do not own a
vehicle, compared with just 2 percent of households with incomes over
$100,000. Individuals living in poverty are about three times as likely
to take transit and 1.5 times more likely to walk.\31\ While these are
desirable ways to get around for environmental and physical health
reasons, they can take much more time and limit travel to destinations
that are accessible by these modes. This is important because research
shows that access to efficient transportation is important for escaping
poverty (via access to education, training, and work) and achieving
upward economic mobility.\32\ In sum, there are millions of Americans
with limited mobility, and autonomous vehicles could help them.
---------------------------------------------------------------------------
\31\ Federal Highway Administration, ``Mobility Challenges for
Households in Poverty: 2009 National Household Travel Survey,'' FHWA
NHTS Brief, 2014.
\32\ Raj Chetty and Nathaniel Hendren, ``The Impacts of
Neighborhoods on Intergenerational Mobility: Childhood Exposure Effects
and County-Level Estimates,'' Harvard University, 2015.
---------------------------------------------------------------------------
Affordability, Availability, and Accessibility Are Keys to Realizing
These Benefits
Simply bringing autonomous vehicles to market might not fully solve
the mobility challenges Americans face. Autonomous vehicles, like other
transportation options, must also be affordable, available, and
accessible. Fortunately, autonomous vehicles may have advantages over
conventional transit, taxi, or vehicle-sharing services.
For many older adults, individuals with disabilities, and other
people living below the poverty line, the costs of a personally owned
vehicle are prohibitive. The costs of a privately owned autonomous
vehicle are expected to be much higher, particularly initially. Shared
autonomous vehicles will be the key to affordability. Shared vehicles
are vehicles that are not personally owned but instead are available
for many people to use, either on demand or through a reservation
system, and are typically pay-per-use. Some estimates suggest that the
per-mile cost of using a shared autonomous vehicle service could be 30
percent to 90 percent less than owning a conventional vehicle or using
conventional taxis, depending on the nature of the service.\33\ In
other words, the per-trip costs could be comparable to transit, but
with greater convenience and speed.
---------------------------------------------------------------------------
\33\ Lawrence D. Burnes, William C. Jordan, and Bonnie A.
Scarborough, Transforming Personal Mobility, The Earth Institute,
Columbia University, January 27, 2013; and Tasha Keeney, ``What If Uber
Were to Adopt Shared Autonomous Vehicles (SAVs)?'' ARK Invest, June 22,
2015.
---------------------------------------------------------------------------
Second, shared autonomous vehicles must be available where people
live. Car-sharing vehicles and taxis are not readily available in most
small towns and rural communities because there are too few people to
support the services. Furthermore, those who live in poor urban areas
are another underserved segment in today's mobility market. Transit may
not offer complete solutions, and taxis have historically been scarce
because of the low demand compared to wealthier urban areas.\34\ The
lower cost of shared autonomous vehicles may increase the availability
in underserved regions where other transportation solutions are
limited.
---------------------------------------------------------------------------
\34\ Mark W. Frankena and Paul A. Pautler, An Economic Analysis of
Taxicab Regulation, Bureau of Economics, No. 1103, May 1984; Nelson
Nygaard, Boston Taxi Consultant Report, 2013; and Hara Associates Inc.,
Best Practices Studies of Taxi Regulation: Taxi User Surveys, prepared
for San Francisco Municipal Transportation Agency, 2013. Today's ride
share services may be helping provide better service in these
underserved populations (see the Uber-funded study by Rosanna Smart,
Brad Rowe, Angela Hawken, Mark Kleiman, Nate Mladenovic, Peter Gehred,
and Clarissa Manning, Faster and Cheaper: How Ride-Sourcing Fills a Gap
in Low-Income Los Angeles Neighborhoods, BOTEC Analysis Corporation,
July 2015).
---------------------------------------------------------------------------
Third, shared autonomous vehicles need to be accessible. This
includes vehicle design, websites, and technology interfaces that are
consistent with Americans with Disabilities Act and other accessibility
standards and guidelines. It also includes implementing diverse payment
systems that do not require smart phones or credit cards. Meeting these
design goals can be expensive. For example, the National Highway
Traffic Safety Administration estimates that the cost of a new vehicle
with adaptive equipment (e.g., mechanical hand controls, power transfer
seats, and lifts and ramps) can be $20,000-$80,000.\35\ The cost for
accessible autonomous vehicles may be lower because the vehicle only
needs to be modified for passenger use; it does not need to be modified
to enable driving.
---------------------------------------------------------------------------
\35\ National Highway Traffic Safety Administration, Adapting Motor
Vehicles for People with Disabilities, June 2015.
---------------------------------------------------------------------------
Policymakers Can Promote Affordability, Availability, and Accessibility
All of this suggests that autonomous vehicles may increase mobility
for historically underserved populations in a way that is more
affordable, available, and accessible than existing transportation
options. However, there is a clear and essential role for sound policy
in realizing these benefits.
First, policymakers can create incentives for manufacturers to
prioritize these markets and reach them sooner than they might
otherwise. Incentives can include cost-sharing programs, subsidies, or
other financial levers. They can also include partnerships to integrate
both public and private shared autonomous vehicles into existing
transit and paratransit services so that they are complementary rather
than competing. This may involve making payment seamless across modes,
providing transfer benefits across modes, and integrating scheduling.
Private ride-share services are already working with transit agencies
to provide connections to existing transit services, but primarily in
urban areas.
Second, policymakers may need to incentivize technology developers
to ensure that accessibility for diverse populations is a priority when
designing these vehicles. This includes facilitating collaboration
between developers, healthcare providers, independent living centers
and other facilities, and, most importantly, the users themselves.
Participatory design will be key.
Third, while the cost of shared autonomous vehicles is expected to
be lower than many alternatives, public assistance may still be
warranted. In many regions, seniors and individuals with disabilities
ride transit at a discounted rate or even for free. Policies would be
needed to extend these discounts to shared autonomous vehicle services.
In sum, autonomous vehicles present an enormous opportunity to
improve mobility for millions of Americans who are currently
underserved by our existing transportation system. The social, health,
and economic benefits could be enormous. Policymakers can play an
important and distinct role in prioritizing and enabling the technology
so that autonomous vehicles can help democratize America's
transportation system.
what mechanisms can help realize the benefits and mitigate the
drawbacks of autonomous vehicles?
In addition to transforming safety and mobility, autonomous
vehicles may also shape other areas of transportation, including
congestion, energy and pollution, and land use. Some potential impacts
will be positive while others will be negative. All of the impacts are
complex and difficult to predict, but despite the uncertainty,
policymakers can help nudge the free market in the right direction.
The Impacts of Autonomous Vehicles on Congestion
Congestion has enormous societal costs. Travel delays resulting
from traffic congestion caused drivers to waste more than 3 billion
gallons of fuel and kept travelers stuck in their cars for nearly 7
billion extra hours--42 hours per rush-hour commuter.\36\ The total
cost to the United States was $960 per commuter, or $160 billion for
the Nation as a whole.
---------------------------------------------------------------------------
\36\ David L. Schrank, Bill Eisele, and Timothy J. Lomax, The 2015
Urban Mobility Scorecard, College Station, Tex.: Texas A&M
Transportation Institute, 2015.
---------------------------------------------------------------------------
Even if autonomous vehicles had no impact on the incidence of
congestion, they could reduce the cost of congestion. If individuals
can work in their cars, the cost of the time spent in traffic could be
reduced substantially, even if the time itself is not reduced.
Nevertheless, the potential impact of autonomous vehicles on
traffic congestion itself could be substantial but is uncertain.
Traffic congestion could be significantly reduced because more vehicles
can fit on a given stretch of roadway if they are autonomous. In the
near term, autonomous vehicle platooning (where cars drive close
together to reduce air resistance and increase fuel economy) can enable
greater throughput; in the longer term, if a large number of vehicles
are autonomous, lanes could be made narrower, creating more usable road
space. If autonomous vehicles are much safer, they could significantly
reduce crashes, which are a major source of congestion. Shared
autonomous vehicles could provide better connections to main transit
lines, leading to increases in use.\37\
---------------------------------------------------------------------------
\37\ Johanna Zmud, Jason Wagner, Richard T. Baker, Ginger Goodin,
Maarit Moran, Nidhi Kalra, and Dan Fagnant, Policy and Planning Actions
to Internalize Societal Impacts of CV and AV Systems in Market
Decisions, interim deliverable to the National Cooperative Highway
Research Program, Transportation Research Board of the National
Academies of Sciences, Engineering, and Medicine, May 2016.
---------------------------------------------------------------------------
However, there is a flip side. Because autonomous vehicles will
lower the costs of driving by car--by enabling productivity in the
vehicle, reducing fuel costs through greater fuel economy, avoiding
parking fees, and lowering insurance costs through greater safety--they
could also increase the amount of driving. Improvements in mobility for
underserved populations would also add to the amount of driving. If
people can do the same things from the comfort of their own cars, fewer
people might take public transit.\38\
---------------------------------------------------------------------------
\38\ Zmud et al., 2016; Anderson et al., 2014.
---------------------------------------------------------------------------
Accurately predicting the net effect on transportation demand is
impossible because of the disruptive nature of autonomous vehicles.
Just as we could not predict in 1990 how the Internet would change how
and how much we would communicate 20 years later, we cannot confidently
predict today how autonomous vehicles will change how and how much we
will travel 20 years from now.
The Impacts of Autonomous Vehicles on Energy
Autonomous vehicles could increase fuel efficiency, but the net
effect is unclear because they may increase travel demand, which could
negate those gains. To the extent that fossil fuels remain the primary
source of transportation energy, this would have knock-on effects in
foreign oil dependence, air pollution, and greenhouse gas emissions.
The way people operate and maintain vehicles is inefficient.
Aggressive driving alone can drop fuel economy by 25 percent, and not
using cruise control on highways can drop it another 7 percent.\39\
Autonomous vehicles can avoid these behaviors and thus reduce fuel
consumption. Adding to this, even relatively simple levels of
automation can enable platooning.
---------------------------------------------------------------------------
\39\ Michael Sivak and Brandon Schoettle, ``Eco-Driving: Strategic,
Tactical, and Operational Decisions of the Driver That Influence
Vehicle Fuel Economy,'' Transport Policy, Vol. 22, July 2012, pp. 96-
99.
---------------------------------------------------------------------------
In the longer term, if autonomous vehicles that crash less are
widely used, they could be built lighter, which will further reduce
fuel consumption and emissions. Less obviously, fully autonomous
vehicles might be able to jump-start alternative transportation fuels.
One of the key obstacles to both plug-in electric and hydrogen fuel
cells, which have zero tailpipe emissions and can use renewable energy,
is the lack of refueling or charging infrastructure. This becomes much
less of a problem if cars can drive themselves to refueling or
recharging stations because far fewer stations are needed.\40\ One
recent study showed that electric shared autonomous vehicles could
reduce greenhouse gas emissions in 2030 by 87-94 percent relative to
current conventional vehicles and 63-82 percent below projected model
year 2030 hybrid vehicles because of decreases in future carbon
intensity of electricity, ``right sizing'' of vehicles, and higher
miles traveled per vehicle.\41\
---------------------------------------------------------------------------
\40\ Anderson et al., 2014.
\41\ Jeffery B. Greenblatt and Samveg Saxena, ``Autonomous Taxis
Could Greatly Reduce Greenhouse-Gas Emissions of U.S. Light-Duty
Vehicles,'' Nature Climate Change, Vol. 5, No. 9, 2015, pp. 860-863.
---------------------------------------------------------------------------
On the other hand, passengers may prefer larger autonomous vehicles
to allow them to take better advantage of the opportunity to do things
other than driving, resulting in lower fuel economy and greater
emissions.\42\ And, of course, they may drive more.
---------------------------------------------------------------------------
\42\ Anderson et al., 2014.
---------------------------------------------------------------------------
The Impacts of Autonomous Vehicles on Land Use
Automobile use has influenced the form and extent of land
development in the United States, leading in large part to sprawl (that
is, low-density, inefficient land-use patterns).\43\ The land allocated
to automobile infrastructure poses a cost to society: It could
otherwise be used for farms, open space, homes, businesses, and other
facilities, with associated environmental, economic, and public health
effects.\44\
---------------------------------------------------------------------------
\43\ Robert Burchell, George Lowenstein, William R. Dolphin,
Catherine C. Galley, Anthony Downs, Samuel Seskin Katherine Gray Still,
and Terry Moore, Costs of Sprawl-2000 TRCP Report 74, Federal Transit
Administration, 2002.
\44\ M. A. Delucchi and J. J. Murphy, ``How Large Are Tax Subsidies
to Motor-Vehicle Users in the U.S.?'' Transport Policy, Vol. 15, 2008,
pp. 196-208.
---------------------------------------------------------------------------
Autonomous vehicles may affect land use in two opposite ways, and
both could take place. Commute time and distance are among the key
factors households consider in deciding where to live. While areas
father away from central business districts offer many benefits,
particularly in housing size and cost, a longer commute may be too
costly, both in terms of travel and time costs. However, given the
ability to engage in other activities while in an autonomous vehicle,
the opportunity cost of transportation declines. This could increase
the willingness of households to locate farther away from the urban
core, increasing urban sprawl.\45\
---------------------------------------------------------------------------
\45\ Anderson et al., 2014.
---------------------------------------------------------------------------
On the other hand, autonomous vehicles could also lead to greater
density in core urban areas. Driving remains the dominant mode of
passenger travel in the United States, even in large cities with good
transit options, but the typical automobile is parked for about 95
percent of its lifetime.\46\ As of a decade ago, the total area devoted
to parking spaces in major central business districts was, on average,
about 31 percent of the district area.\47\ The emergence of autonomous
vehicles could sharply reduce the amount of parking needed in core
urban areas in several ways. First, after dropping off its passenger or
passengers in a downtown location, an autonomous vehicle could pilot
itself to a remote lot in a peripheral area, reducing the amount of
parking needed in the densest urban areas where land values are
highest. Second, as described earlier, autonomous vehicle technology
might lead to a new model for urban mobility in the form of driverless
taxis. Under such a system, autonomous vehicles would not need to park
after every trip; rather, after dropping off one passenger, they would
simply travel to pick up the next passenger. Third, the convenience and
low cost of such a system might induce many urban dwellers to forgo car
ownership, or at least to reduce the number of cars owned. Thus,
driverless taxis could reduce the number of parking spaces needed in
residential buildings, as well as at commercial centers.\48\ These
effects, emphasizing the service character of transportation, could
free up substantial amounts of space in urban areas for other valuable
uses: homes, businesses, parks, hospitals, and so on.
---------------------------------------------------------------------------
\46\ Donald C. Shoup, The High Cost of Free Parking, Chicago:
Planner's Press, 2005.
\47\ Shoup, 2005.
\48\ Zmud et al., 2016.
---------------------------------------------------------------------------
Driving Externalities May Prevent the Benefits from Being Realized
While the effects of autonomous vehicles are complex, some outcomes
are clear wins. If safe autonomous vehicles are developed and used
widely and responsibly, the current public safety crisis in the U.S.
transportation system could be mitigated. If safe and usable autonomous
vehicles are developed, mobility could increase for millions of
Americans who currently have limited mobility. In addition, if the
potential increase in transportation demand created by autonomous
vehicles were mitigated or decoupled from fossil fuels, there could be
enormous energy security, public health, and environmental benefits.
Related to this, if shared autonomous vehicles are widely available and
widely used, this could reduce private vehicle ownership and the need
for road infrastructure, allowing repurposing of land to more
economically productive uses.\49\
---------------------------------------------------------------------------
\49\ Zmud et al., 2016.
---------------------------------------------------------------------------
Yet these outcomes may not actually be realized because many
benefits accrue to society rather than to either the producers or
consumers of autonomous vehicles. Consumers may be unwilling to pay for
expensive technology if much of the benefits go to others, and
consequently, producers may be less willing to develop them. Thus,
there is less incentive for producers and consumers to take actions
that would achieve beneficial outcomes.\50\
---------------------------------------------------------------------------
\50\ Zmud et al., 2016.
---------------------------------------------------------------------------
Safety is a good example of this phenomenon, with significant
consequences for autonomous vehicles. When an individual drives
unsafely or operates an unsafe vehicle, he not only puts his own well-
being at risk but also the well-being of all other road users around
him, including pedestrians and bicyclists. However, in our current
transportation and legal system, an individual is responsible for only
a fraction of the full cost of being unsafe. In many States, motorists
are required to carry only $30,000 (or less) worth of liability
insurance--far less than is necessary to compensate someone for a
serious injury, much less a loss of life. This leaves a huge gap
between the harms that are regularly inflicted by drivers and the
amount available for legal recovery. In essence, society subsidizes
dangerous vehicles and driving behavior, creating less incentive for
safer vehicles and behaviors.
Economists call this an externality. An externality is an effect
that one party imposes on another party without compensating them for
the effect if it is negative or charging them for it if it is
positive.\51\ The free market does not allocate resources well in the
presence of externalities because the true costs and benefits of
actions are distorted.
---------------------------------------------------------------------------
\51\ James M. Buchanan and Wm. Craig Stubblewine, ``Externality,''
Economica, Vol. 29, No. 116, 1962, pp. 371-84.
---------------------------------------------------------------------------
Consider how the safety externality dampens the market for safe
vehicles, including safe autonomous vehicles. First, buyers' incentive
to purchase safe autonomous vehicles (which we can expect to be
expensive, at least at first) is less than it would be if full social
benefits of safe vehicles were reflected in a lower price tag. Second,
and related, auto manufactures' incentives to create as-safe-as-
possible autonomous vehicles are less than they should be, because
safety is undervalued in the marketplace.\52\ The result is that very
safe autonomous vehicles could be technologically feasible, but fewer
firms will develop them and fewer individuals will buy them because
many of their benefits accrue to the public rather than the purchaser.
---------------------------------------------------------------------------
\52\ This externality explains, in part, why there is little market
for vehicles that are designed to better protect individuals outside of
the vehicle in the event of a crash. The technology exists, but the
societal benefit of protecting others does not reach the buyer in the
form of a relatively lower-priced vehicle.
---------------------------------------------------------------------------
Safety is just one externality. Many of the benefits and the costs
of autonomous vehicles (and vehicles in general) are external. If a
buyer's car is energy efficient, it helps the buyer somewhat, but most
of the benefits go to other people (e.g., the people who breathe the
air in the area where that vehicle is driven). Those costs (e.g., of
poor air quality) are real and are borne by society. If the benefits of
reducing pollution are not factored into the buyer's cost of the
vehicle, there is little incentive for them to buy it, particularly if
the vehicle is more expensive than less-efficient alternatives.
Policymakers Can Promote Beneficial Outcomes by Internalizing
Externalities
So, how can we solve the externality problem? The key is to use
policy tools to ``internalize'' externalities so that market prices
reflect the true costs and benefits of private-sector actions. This can
be done with subsidies, user fees, mandates, and privileges to ensure
that producers and consumers of autonomous vehicles receive the benefit
from (and are thus incentivized toward) making choices that benefit
society.
As just one example, when a driver uses a busy road, he adds to
congestion that other travelers experience, but he does not have to pay
for the cost of that extra congestion--the lost productivity of others
as they sit in traffic, the delay in goods movement, and the local
increase in pollution. But congestion is a problem that could be
solved. Nearly all passenger vehicles in this country have space for at
least four people, but on average, there are just 1.67 passengers.\53\
Those unused seats are extra, already-built transportation capacity.
But that capacity is not used because, in large part, the costs of
carpooling are internal (the driver bears the cost of the effort and
hassle) but the benefits of carpooling remain external (the driver does
not benefit from reducing society's congestion). High-occupancy-vehicle
(HOV) lanes are one partial remedy: They help internalize the positive
externalities of carpooling by enabling carpoolers to themselves bypass
congestion and get to their destinations faster.
---------------------------------------------------------------------------
\53\ Federal Highway Administration, Summary of Travel Trends: 2009
National Household Travel Survey, June 2011. As of November 8, 2016:
http://nhts.ornl.gov/2009/pub/stt.pdf.
---------------------------------------------------------------------------
There are many policy options to internalize not only the
congestion externality but also the other driving externalities related
to safety, pollution, oil dependence, and mobility. These include
creating insurance requirements that strengthen the market for road
safety, offering transit incentives that reduce congestion, and
offering rebates for using fuel-efficient vehicles, among others. Each
option has a long history of research and discussion, and these and
other options have been implemented to varying degrees.
It is not possible today to give each policy the discussion it
warrants. Nevertheless, now is the time to revisit the impact of
driving externalities and the policies to internalize them, because
autonomous vehicles could improve our transportation system
tremendously, provided the right market signals are in place. In other
words, because so many of the benefits and costs of autonomous vehicles
would accrue to people other than the buyer, internalizing
externalities is a key step to ensuring that society minimizes their
disadvantages and maximizes their benefits.
There is also a current window of opportunity to send those new
market signals because consumer expectations about autonomous vehicle
costs, performance, and other characteristics are not yet set. For
instance, congestion pricing requires drivers to pay a fee to travel
during peak rush hour, because driving during rush hour imposes higher
congestion costs on everyone than driving at other times does. Although
congestion pricing is widely recognized as an effective means of
internalizing the cost of congestion and thereby reducing congestion,
it has been difficult to implement, partly because drivers are
unaccustomed to paying different prices based on when they travel.
However, today, many private ride-sharing companies charge an extra fee
for their services during rush hour, analogous to congestion pricing.
Because these services are new and the reasons for the charge are
understandable to consumers (greater demand for a limited supply of
goods), these fees have been generally acceptable. Similarly, consumers
may be more amenable to new policies that internalize the externalities
of driving now, before autonomous vehicles are available, rather than
later, once expectations about autonomous vehicles are set. Of course,
these policies must apply to all auto travel, not just autonomous
vehicle travel, for the market signals to be clear.
In sum, it is not possible to fully predict what a future with
autonomous vehicles will look like. However, by using the current
window of opportunity to internalize the externalities of driving, it
is possible to send the right market signals, paving the way for a
future transportation system that maximizes the potential advantages
while minimizing the potential disadvantages. This is an exciting
future of increased mobility and economic growth and greater
transportation safety, efficiency, equity, and sustainability.
[Conflict of Interest Statement: Nidhi Kalra's spouse, David
Ferguson, is co-founder and president of Nuro, an autonomous vehicle
startup. He previously served as a principal engineer for Google's
driverless car project. This written testimony was carefully reviewed
by subject-matter experts within the RAND Corporation; the research
quality assurance team for the RAND Justice, Infrastructure, and
Environment division; and the RAND Office of Congressional Relations.
However, the opinions and conclusions expressed in this testimony are
the author's alone and should not be interpreted as representing those
of the RAND Corporation or any of the sponsors of its research.]
Senator Collins. Thank you very much for your testimony,
and I am going to start my questions with you.
MOBILITY FOR SENIORS
As a senator representing a State with the oldest median
age in the country, I can see tremendous potential for seniors
whose vision has diminished to the point where they no longer
can safely drive being able to use these autonomous vehicles so
that they can maintain a measure of independence and not be
dependent on others. So I see that, in addition to the safety
benefits, and really the two are linked, as being a tremendous
advantage.
Another advantage which I wondered whether you had analyzed
at all is whether there would be an impact on the insurance
market and the rates that individuals would pay for insurance
if they are driving cars with either limited safety features
that have been added to it or are fully autonomous.
Dr. Kalra. I appreciate the note about seniors. My
grandparents are in that same boat, so I understand that very
clearly.
In terms of insurance, yes, in the long run we might expect
insurance rates to go down because fewer crashes mean lower
costs for personal injury and damage, and eventually they may
be needed for declining auto insurance. But in the short run,
it's actually hard to say what the effect of insurance is going
to be, for a few reasons.
First, it may take many years or many decades for the fleet
to become largely autonomous, so the risks may not change as
quickly as we anticipate.
Second, because at least for some types of autonomous
vehicles there's going to be shared control between the human
and the machine, there may be over-reliance on the technology,
which is one of the things we've seen with Tesla. So it may not
necessarily translate into lower crashes or crash rates. That
remains to be seen.
And third, the car repair costs for a crash could actually
increase, because right now if I rear-end someone, it's a few
hundred dollars to bang out that bumper, but in the future it
could be much more expensive.
Now, there's a different issue with people who are insured
at the minimum insurance rate, but that's a little more
complicated. We can go there if you want to.
Senator Collins. Thank you.
EXISTING VEHICLE TECHNOLOGIES
Administrator, we heard Ms. Hersman say that three
technologies, if they were mandated, could save some 10,000
lives. Now, whenever the Federal Government mandates, there is
obviously cost involved. But I am curious whether NHTSA has
considered mandating proven safety technologies that are
available right now on new vehicles as they're manufactured. I
should make that clear.
Mr. Rosekind. In fact, I think as soon as anybody says to
any of you ``We need to regulate to get safety,'' the first
thing you need to ask is what are we going to regulate? Because
for regulation we need performance criteria, testing, we need
to know that there's enough penetration that we have sufficient
data, cost/benefit analyses, et cetera.
I say that because the last technologies that we have seen
come through rulemaking are things like electronic stability
control, that rear visibility camera that you like so much,
advanced air bags. Those took 6, 8, and 10 years to actually
get through the regulatory process. That's really important
because in these new technologies, by the time those rules
would come out it would be irrelevant for the new technology
that would have evolved.
And I say that because, quickly, in September of 2015 we
actually challenged the auto industry: how do you take
automatic emergency braking and make it standard on all
vehicles? This is called democratizing safety. So it's not just
on high-end or an option. Basically what they did was in March
come back, and 20 automakers are going to make AEB standard by
2022 on all vehicles in our country. That will beat regulation
by 3 or 4 years. We can count the lives saved.
So there are ways to do this besides regulation, and I
think we have to use the right tool. For the moment, this
policy is the right tool at the right time.
STATE REGULATIONS OF AUTOMATED VEHICLES
Senator Collins. Well, let me ask you the other side of
that coin. The guidance the Department has put out said a 15-
point safety assessment is voluntary, but as you're well aware,
States like California are already proposing to make it
mandatory for companies that want to test or deploy automated
vehicles. What can you do, what can NHTSA do to ensure that
States do not arbitrarily mandate various aspects of your
guidance document, thereby creating a patchwork of ever-
changing State laws that would stymie innovation and the
deployment of these vehicles?
Mr. Rosekind. Really, a huge part of including that piece,
the model state policy, was to differentiate the Federal versus
State roles. In fact, directly to your point, we make it
absolutely clear that States actually have to do nothing in
this area. They can actually support the advancement of these
safety technologies with no policy or regulations. If they
choose to, we identify some areas where they can move forward.
Everyone on the State level that we've interacted with is
absolutely interested in seeing a consistent framework and
trying to avoid the patchwork. This is an area I think we need
to watch and see it develop.
And California, as you mentioned, if you look at their
early, sort of recommended policies, they've changed
dramatically in their latest proposals because they waited for
this policy.
So I think everybody is seeing an effort and an interest
right now in trying to prevent that patchwork, and we're going
to have to wait and see how it actually develops.
Senator Collins. But should the States even be involved in
this area, or should it be left up to the Federal Government so
that there is a nationwide standard and approach?
Mr. Rosekind. In the policy we make that explicit, here's
what the Federal Government is concerned about, standards for
the vehicles, defects, et cetera, and the States, they still
stay responsible and they don't have to do anything. We make
that explicit.
Senator Collins. You do, but if California is moving
forward to implement your guidance and make it mandatory, isn't
that contrary to the goal of your guidance?
Mr. Rosekind. Two things. One is we are explicit in there
saying that this policy was not intended to codify, and at the
same time I think right now California has not actually put
their regulations out. They've put a proposal out. They've
talked about other aspects, that they want to evaluate it, and
they're even questioning the language that's being used, and
that's what I'm saying. I don't think we even know what their
final stance is going to be, because they have continually
emphasized the need for consistent framework.
Senator Collins. Senator Reed.
CYBERSECURITY
Senator Reed. Well, thank you, Chairman, very much.
I agree with Mr. Brubaker; computers don't get tired, they
don't get road rage, but they open a whole new dimension of
cyber security which we have to deal with. I know,
Administrator, the FAVP encourages information sharing on cyber
security, but we've seen incidents in which systems have been
hacked recently, vulnerabilities in automobiles, GPS systems,
insurance plugins have been used.
So can you give us an idea of what you're doing in terms of
ensuring that we can reassure the public that these vehicles
will not be subject to cyber attacks?
Mr. Rosekind. Thank you for emphasizing this issue because,
as you've heard from pretty much everybody, without public
confidence in these vehicles, it's just not going to happen.
So specifically to cyber security, NHTSA has actually been
on this for many years. In 2012 we formalized that with a
group, an office within the agency. Just recently we have had a
roundtable with 300 people, pulling them together. We had an
intra-government meeting after that, which has resulted in
everything from urging and supporting an agency, Auto ISAC,
Information Sharing and Analysis Center, within the industry.
The industry has come out with their own best practices. Just a
few weeks ago, NHTSA came out with its own cyber security best
practices. Just yesterday the Department of Homeland Security
actually highlighted our cyber security best practices, a model
for other industries to go after.
So the good news is all of this has been done before an
incident has actually occurred within the auto industry. One
that everybody knows about, about a year ago in July Wired
Magazine, that was planned. That was a researcher demonstrating
it could be done. I'd point out that even with the authorities
and tools we had, within 3 days the defect was called and 1.4
million vehicles basically were already under remedy at that
point.
So I would just say it's a constant vulnerability. It
clearly is a threat that needs to be addressed. But for the
moment there's a lot of action going on in that arena.
Senator Reed. Let me just quickly follow up with Ms.
Hersman, and thank you. How do you think outside expertise
could assist in combatting these cyber threats and regulating
advanced technologies like HAVs?
Ms. Hersman. I think there's a tremendous opportunity to
call on experts, just like this committee does. Things change
very quickly, and when it comes to technology, we certainly
can't expect NHTSA to always stay on the cutting edge.
Potentially having advisory groups, organizations that will
assist them in evaluating new cutting-edge technology is an
opportunity that they should consider and take advantage of
going forward. We're going to ask so much of this agency when
it comes to looking at the evolution of technology that we need
to support them and give them opportunities to get the best and
the brightest to weigh in on it.
Senator Reed. Thank you very much.
IMPACT ON LABOR
Dr. Kalra, I seem to be emphasizing the potential
challenges rather than the golden opportunities, but we should
do that. And one of them, as I mentioned in my statement, is
employment, roughly 3 to 4 million jobs that are good jobs.
You've done some work, I believe, on this. You made some
comments in your discussion.
Can you just comment upon the implications of the labor
market, both positive and negative, for these vehicles?
Dr. Kalra. Absolutely. So you're absolutely right that
there are millions of Americans who make their living behind
the wheel, and there's no question that autonomous vehicles do
threaten those jobs. I think we have to stare that back in the
face. The question is what to do about it.
This transition to full autonomy is going to take time, so
time can be an advantage in this respect in that we are ahead
of the curve. We can start thinking now about how to develop an
alternative job market, where that's possible. We also know
that freight may be one of the areas that is first hit because
of the private nature of the industry and because we're talking
about goods movement rather than people movement. So the
technology may make its way there faster. One aspect of that is
that many freight experts project a lack of supply of drivers
in the future, so there may be a little cushion in there.
In terms of on the flip side, autonomous vehicles will
create new kinds of jobs or increase the technical nature of
existing jobs. An auto body shop is not going to look the same
as it does today. So we need to prepare our young people
especially, people in community colleges, for those high-tech
jobs.
The broader issue is that autonomous vehicles could and, I
believe, will democratize transportation and give mobility to
people who currently don't have it. Important to that is
physical mobility. Being able to get around is one of the most
important things to help people out of poverty, getting them
access to training, getting them to jobs. The unfortunate part
is people who have the least access to transportation often are
the ones who struggle to get jobs, people who are in poverty
who can't afford their cars.
So there's two sides of this coin. I think we need to do
everything we can to bring mobility to people who are currently
underserved by our transportation system, while providing
cushion to people who will be negatively affected.
Senator Reed. Your comments seem to be there has to be a
conscious, deliberate planning process because this is coming.
Dr. Kalra. That is right.
Senator Reed. And your best guess at when we'll see, for
example, significant autonomy in freight delivery? Is it 5
years? 10 years? Too far to guess?
Dr. Kalra. I can only guess. I would be surprised if we
don't have it in 10 years. I would be surprised if we do have
it in the next two to three.
Senator Reed. Okay. Anybody else want to take a wild guess,
like the lottery?
Senator Reed. Okay. Thank you, Madam Chairman.
Senator Collins. Thank you, Senator.
Senator Daines.
DATA SECURITY
Senator Daines. Thank you, Madam Chair and Ranking Member
Reed. And thank you for testifying here today. This is a topic
that covers many issues I care deeply about. As a 5th
generation Montanan, we have a lot of open space, a lot of
roadways. In fact, we just wrapped up the election season, and
one of the candidates running for governor traveled 64,000
miles during the course of the campaign on the roads. So that
kind of puts in perspective that we're not as big as Texas or
Alaska or California, but we have to drive a lot more. So I
care very much about what you're talking about, saving lives,
to improve passenger and freight mobility, very important for
us in a State that requires moving our products, ag products
particularly, around the world.
I'm very excited about these new technologies. I spent 12
years in the cloud computing business, executive capacity,
before I came to politics.
So I wanted to follow up on Senator Reed's question on the
cyber piece here, Mr. Rosekind. You said in your testimony that
the quickest way to slam the brakes on innovation is for the
public to lose confidence. I think that's well said. I was a
private employee for 28 years, and the best was when I didn't
have my information compromised, until I became a Federal
employee. Then I got the letter from OPM. Thank you, Federal
Government here. I never had that issue in the private sector.
Mr. Rosekind. I got that letter, too.
Senator Daines. Many of us did. I'm sure if we polled the
audience here, there would be a lot of hands going up. Clearly,
we want to make sure we're protecting privacy and hardening our
systems.
Many consumers no longer have confidence in the government.
I appreciate the fact that you're bringing some advisory groups
in to bring that perspective. Going fast, at the speed of
business out there, sometimes I think government needs to.
How is NHTSA gaining the public's confidence with this
latest guidance that you're looking at here as it relates to
cybersecurity?
Mr. Rosekind. I think in the policy, one of the ways
actually that we're trying to do that is through the innovation
approach. I think, just like when people talk about regulation,
when you talk about cyber security, if someone were to say
regulate that, as you know, the speed that this stuff is
changing, by the time you get through a 6- to 10-year
regulation it's not going to be relevant anymore.
So, one of the 15 safety assessment items is cyber
security. Everybody has to tell us how they're addressing that
particular issue, and our intent here is to see as many
different forms of innovation coming to us as we can, and the
data will drive the safest and best way to protect these
systems. In fact, at some point, if there are best practices,
it should be based on that data. In the future if there's
rulemaking, it should probably be based as a foundation on
whatever those best practices were.
Senator Daines. I know industry has been working
collaboratively to address cyber and published some best
practices in July. How did NHTSA's guidance incorporate their
experiences and expertise?
Mr. Rosekind. We had a lot of interaction with them. We
knew what was coming. In fact, I would highlight not just this
policy but our own best practices came out just a few weeks
ago, and they complement exactly what we know the industry was
doing.
Senator Daines. So I guess my understanding is they didn't
go through a notice and comment process before being issued.
Afterwards DOT solicited comments, and I think they're due next
week. So my question, I guess, is why wasn't there more
consultation with industry before the guidance was issued?
Mr. Rosekind. Well, let's keep them separated. The best
practices came out a few weeks ago, and the policy came out
September 20th. We actually have an extensive amount of open
time. So the policy is under a 60-day comment period right now.
That closes November 22nd. And what was already identified in
here is 23 next steps. So we actually have a whole new set of
public meetings that are coming up to talk about all the
different elements. We just did that last week. We had a
meeting on the letter, which includes the cyber security part.
So there's all kinds of other opportunity for people to add
input.
Senator Daines. And that's helpful, and I know Senator
Collins was talking with us a bit about some of this guidance,
which becomes quasi-regulation. You touched on the time
required to go through the formal rulemaking process, and the
guidance allows you to be more nimble, but it may not always be
as transparent. I think that's one of the political concerns we
hear.
RULEMAKING PROCESS
Should voluntary guidance policies be used to expedite the
rulemaking process or Federal enforcement action?
Mr. Rosekind. I think what we're trying to do is continue
our enforcement and regulatory tools. We have those
authorities. We're not giving any of them up. All we've done is
try and complement those with the way to support innovation at
this time. It is absolutely possible, if you come in with the
right data, that we will use that data to create best practices
and rulemaking. We actually identify potential rulemaking in
here as well. So that is just one more tool that's being
included to deal with this fast-paced technology.
Senator Daines. Thank you. I'm running out of time.
DRIVING IN RURAL AREAS
Dr. Kalra, a question. You discussed lowering the risk to
pilot programs and raised questions about the ability of this
technology to function better than humans in complex
conditions. I come from a State that has a lot of rural roads.
Fifty-four percent of automobile fatalities occur on rural
roads despite the fact that just 19 percent of Americans live
in rural areas.
How do you incorporate issues like we have in Montana of
unmapped roads, gravel, snow, wildlife, where 95 percent of the
roads are rural? As we know, accidents often happen when things
we don't anticipate occur. How do we lower the risk of these
technologies in rural settings?
Dr. Kalra. By testing them in those environments. And I'll
point to the University of Michigan's testing center for
autonomous vehicles, where they are committed to testing on
different kinds of road surfaces, in different kinds of weather
conditions. We need to get autonomous vehicles not only in
sunny Southern California but also in Montana, and I think that
has to be a priority.
Senator Daines. The sun shines a lot in Montana, but we
also have ice and big elk out there at times.
Dr. Kalra. I'm from North Dakota. I know that ice.
Senator Daines. Oh, you do understand that.
Dr. Kalra. I do.
Senator Daines. You might understand ice better than I
understand it, I think.
Anyway, thank you. I'm out of time.
CYBERSECURITY RISKS
Senator Collins. Thank you for raising that very important
issue, which is one that concerns me as well.
Mr. Brubaker, both of my colleagues have brought up the
cyber security issue, which is also of great concern to me as
we've seen we have rogue States and terrorist groups that seem
to be able to hack into virtually any database. What is the
industry doing to address the cyber security risks of
autonomous vehicles, and do you think that there's more that
NHTSA could do to proactively address such a risk?
Mr. Brubaker. Yes, I do. The one thing I would caution is
you want to be very careful in regulating cyber security and
establishing cyber rules because you tend to lock in a solution
that's relevant for that time period, but it's not timeless.
But to answer your specific question about what NHTSA can
do, I would actually start by going across the river to the
five-sided building and start talking to the people who have
been doing embedded system cyber security for years who are
really good at it. One of the key elements of the third offset
strategy the Department is pursuing right now--and I know both
you and the ranking member have a strong defense background, so
you would get this--is really focusing in on automation and
robotics, advanced robotics. So they're doing the kind of cyber
security strategy and employing the kind of tactics that you
need to really lock down these embedded systems, and they've
been doing it pretty successfully for years.
So I think the one thing that we could do is look across
the river, look to the IC, the intelligence community, for
lessons learned on how they do encryption, how they do
authentication. I know that the industry has issued its best
practices but, frankly, when I look at it, I think it's a lot
of reinventing the wheel, a lot of things that have already
been done.
I would also say, again, the caution on the regulatory
piece is we've had the Federal Information Security Management
Act, FISMA, for a number of years, but it didn't prevent the
OPM hack. So you can issue the regulations, and you can comply
with those regulations, and you can go down a cyber security
checklist, but unless it's substantive, and unless it's moving,
because it's a constant game with the adversaries, and you need
to understand their motives and what they're about and do the
risk assessment accordingly and structure your systems from the
ground up.
The one area that I think industry has really got to pay
very close attention to is supply chain integrity, where
they're securing the components, they know where those
components come from, they have visibility end to end, and
there are a handful of companies out there who are focused on
it, automakers that are focused on it. But by and large, the
industry is having a very difficult time adjusting its culture
to harden its systems.
Senator Collins. Thank you.
DISTRACTED DRIVING
Ms. Hersman, yesterday's New York Times quoted you in the
area of distracted driving, and you pointed out that new
technologies that are intended to allow the driver to keep his
or her hands on the wheel may actually be preventing the driver
from paying attention to his or her driving environment, and
the article quotes you as saying it's the cognitive workload on
your brain that's the problem.
Are you concerned that automated and self-driving
vehicles--well, self-driving vehicles are not a problem if we
really get to that stage. But the semi-automated, if you will--
you have your scale for the degree of automation--are going to
produce even more distracted drivers and thus more crashes?
Ms. Hersman. I would say there are always unintended
consequences when we introduce new things. We do know that
people are very distracted today. People are addicted to their
devices that they bring into the car with them. The challenge
with moving to things like hands-free is that people think that
the distraction is in their hands, but the distraction is
really in your brain. We would have outlawed stick-shift cars a
long time ago if it was our hands that were the problem.
So when we look at automating vehicles further, what that
does is potentially has the human being stepping back a little
bit, and human beings typically are not good monitors. When
everything works according to plan 99 times out of 100, it's
hard for the human being to pay attention 100 percent of the
time. In aviation we call this over-reliance on automation. We
saw pilots' skills deteriorate as airplanes performed more
functions for the pilot.
One of the challenges is how do you compel the human being
to pay attention so that when they need to take over or when
they need to intervene, they're ready and they're prepared to
do so? If you could text, if you could read the newspaper, if
you could do other things, 99 percent of the time it might work
well. It's that one time that you need to intervene that you're
not prepared to, and we haven't even touched on impaired
drivers and drivers who might be sleeping, who might be
fatigued or not paying attention.
So there definitely are some risks and some tradeoffs. I
think a number of the panelists have touched on these. That's
why we talk about the messy transition between Level 1 and
Level 5 and having to keep that human in the loop, how can you
compel them when they need to be engaged to do it.
Senator Collins. Thank you.
LESSONS FROM VOLVO DEMONSTRATION
Administrator, last month the world's first shipment by a
self-driving truck was made by an auto Volvo truck. I watched
the video of that. There was a professional driver on board
but, in fact, the truck was driving itself 120 miles along
Interstate 25, and it was operating fully autonomously without
that driver's assistance. It was very exciting to watch, but
I'll have to admit to you it also made me a bit nervous as I
watched it.
Are there any lessons from that test that NHTSA has learned
so far, and would it have any effect on where you're going with
your future guidance?
Mr. Rosekind. Absolutely. Let me just add a little bit to
that confidence. That was actually at 2:00 a.m. in the morning
with a patrol car behind them, so there was a lot more
protection there than just what you see in the video, which was
very promotional.
Having said that, there is a lot of lessons learned, and
part of the reason to have the policy out is so that people are
basically handling all these safety issues before they're ever
out testing or deploying these kinds of things. That's part of
what we're trying to do, is be proactive instead of reactive,
wait for something bad to happen and then react to it. We'd
rather deal with all of these issues, not just with that truck
example but what's going on in Pittsburgh. You cited all of the
exciting things that are happening, but we really need to make
sure that safety is being addressed, at least for these 15
items, before these things get tested or deployed on the road.
Senator Collins. Senator Reed. Thank you.
IMPACT ON LABOR
Senator Reed. Thank you very much, Ms. Chairman.
Mr. Brubaker, you in your testimony note also this
potential labor shift and that we have to be prepared to
respond to it. Could you just give us your advice or your
thoughts on the type of coordination at the Federal level we
have to take to anticipate this and deal with it?
Mr. Brubaker. Sure, absolutely. I want to commend the
subcommittee for the language that they put in the report
language last year asking the Secretary to devote funds to TRB
(Transportation Research Board) to study the economic impacts
of these self-driving technologies because, as we all know,
there will be displacement, and we really need to understand it
and get proactive.
So it's going to involve a number of different Federal
agencies. I mean, the obvious one is the Department of Labor.
They would be best poised to kind of understand where the
shifts are. But I think even to the point where you're
identifying the most vulnerable professions would be very, very
helpful. And then even to the point where if people wanted to
volunteer proactively for retraining and job placement in
needed skill areas, we could coordinate that, and it's one of
the areas that we lay out in our framework. It's the eighth
area where we really need to be mindful of it and very
sensitive.
I mean, I grew up in Youngstown, Ohio, and I saw what
happens when people don't--and how they vote later--when people
aren't treated with respect.
Senator Reed. That's exactly right. I was in Niles a few
months ago, right outside of Youngstown.
Mr. Brubaker. Near my hometown.
Senator Reed. So I get it. We have to be very, very
conscious of this.
NATIONAL HIGHWAY TRAFFIC SAFETY ADMINISTRATION GUIDANCE
Doctor, you and RAND published a guide on this technology
in 2014. As you look at NHTSA's guidance, do you think they're
addressing policy areas that need to be addressed at the
Federal level? I have harped on employment and cyber security.
Are there other issues?
Dr. Kalra. Can I ask, do you mean for NHTSA in particular,
or the Federal Government in general?
Senator Reed. NHTSA in particular, and then go big.
Dr. Kalra. Okay. For NHTSA, I think they need to continue
the course that they're going on. Their guidance is admirable.
It threads the needle well on how do you provide guidance and a
framework for safety while allowing innovation to occur. But
one of those items is, for example, ethics, and it asks
developers to say how will the car handle ethical issues. Even
if the reporting is done, it's not clear that reporting will
lead to safety. So there's a gap there because even though one
can follow the guidelines, that doesn't mean autonomous
vehicles will be safe, and that's because no one really knows
how to prove safety in advance of getting these vehicles on the
road.
So I think continuing in that direction is important.
For the Federal Government more broadly, in addition to the
economic issues, I just want to reiterate that there are market
distortions. I know this is wonky, this is what RAND does, but
there are market distortions in our transportation system that
mean that we don't take full advantage of the opportunity for
cost savings.
I'll be very specific. When we have pollution, whether it's
factory pollution or automotive pollution, that's a real cost
to our society. It is paid in healthcare. It is paid. It's not
an invisible cost; it's a real cost. The question is who pays
for it? So as long as the market does not take into account the
costs that we all bear for some of these activities, there
isn't an incentive, a market incentive to reduce them.
So I think that is something that often gets overlooked as
we talk about the safety of autonomous vehicles, which is
understandable. But I would encourage us, if we want this
future of sustainability and efficiency and mobility for
everyone, to look at the other things in our transportation
market that are preventing us already from achieving those
goals.
Senator Reed. Thank you.
Ms. Hersman, in terms of NHTSA's guidance, do you think it
effectively addresses safety concerns from your perspective?
Ms. Hersman. I would say it's a good first step, and I
think we can see the challenge that exists of not having
anything and having a vacuum, and then the States are stepping
in, there's a patchwork system. But I think we're also so in
the early days here. There's so much that we don't know. If
they were to put out regulations right now, they probably
wouldn't be the right ones.
So I think when we look at the balance between the Wild
West of doing nothing and saying nothing on this and letting
everyone do whatever they want versus locking everything down
with prescriptive regulations going into the future, I think
they've really tried to achieve a fine balance there. I think
there's more work to be done, absolutely. But I think this
gives people something to react to. It gives us a framework to
start from.
I do want to go back to the issue of workplace. We're a
safety organization, so we track deaths and injuries. The
transportation industry is one of the most dangerous when it
comes to workers' health and their injuries and deaths. They're
just behind the construction industry for absolute numbers of
fatalities, and just behind agriculture and mining for the
rates. So it's a very dangerous job. Forty percent of workplace
fatalities are motor vehicle crash related.
So those folks, we want to make them safer too. So if we
can have that path to figure out how to take care of them over
the long run, I think this will be good for them as far as
their safety.
ROLE OF GOVERNMENT
Senator Reed. There are so many opportunities for safety,
for improved productivity, for better health effects. But then
there's the other side of the equation of what about the
drivers that have done a good job, the defined pension, et
cetera, and suddenly there's an autonomous vehicles and thank
you very much for your service. We have to deal with that.
And that goes to the question of there are some people that
might prefer the Wild West, let the chips fall where they may
and this will work itself out. But that goes to, Mr. Rosekind,
the agency has to have a role here.
So can you just give us a sense of what that role is going
forward as you deal with a very sophisticated industry that is
conscious of these issues, the safety advocates of the cyber
dimensions? Can you just give us assurance that it won't be the
Wild West but it will be innovative and based on good
experience?
Mr. Rosekind. You've just heard from pretty much everybody
on the panel what a challenging arena to address. How do you
make sure you don't clamp it down, all this great American
entrepreneurial spirit that could save lives and make life so
much better for everybody? Henry Claypool is over there. We
talk about the elderly. He's representing the disabled
community. Never had driver's licenses; don't need them.
Autonomous vehicles can do it. We're just talking about great
entrepreneurship and innovation being tremendous and not
wanting the Wild, Wild West. That's why it's the right tool at
the right time.
The policy is intended, we hope, to be literally as
innovative for the government as it is for the technology
companies. This is a first step. In fact, in there we make a
commitment that in a year, with all of the different 23 next
steps that are going on, on an annual basis this policy will be
reviewed and updated. We can't do that with regulation in any
kind of effective way.
So we're trying to strike for a very first effort that
right balance, to support innovation and safety. People think
they're at the opposite ends. We're trying to support that
middle ground, safety and innovation, because we're talking
about 32,092 lives, and we could save all of them. Autonomous
vehicles should play a critical role in us getting there.
Senator Reed. Thank you.
Thank you, Madam Chairman.
ENERGY AND ENVIRONMENT
Senator Collins. Thank you and I appreciate very much you
bringing up the employment impact, which is certainly something
for us all to think about as well.
I just have a couple more questions. One, Dr. Kalra, do you
see any energy consumption and environmental benefits from the
deployment of autonomous vehicles?
Dr. Kalra. Absolutely. There are tremendous benefits and
risks. But the benefits include, for example, if we have a
large number of autonomous vehicles, we can increase the
throughput of vehicles on our roads, reduce congestion, because
a lot of congestion is actually caused by crashes. So there's a
reduction in congestion, which immediately translates into a
reduction in fuel consumption.
But if we think big, one of the biggest challenges we have
in getting real alternative fuels, like hydrogen, into our
transportation system is the infrastructure problem, meaning
the distribution problem of how many hydrogen stations do you
need. And if an autonomous vehicle, fully autonomous, can drive
itself to a hydrogen station in the future, gas up at 2:00 in
the morning, or maybe there are long lines but no one cares
because there's no one in there, autonomous vehicles could
actually make possible transportation fuels that right now are
struggling to get off the ground because of these logistical
issues.
So the opportunities, dare I say it, are limitless for
improvement. But again, the market signals should also be in
play.
Senator Collins. Thank you.
ETHICAL CONSIDERATIONS
I have one final question that I wanted to ask each of you,
and that is as I have gotten more involved in this fascinating
issue, there is a question that keeps occurring to me, and that
is the ethical considerations that a human driver makes when
faced with two unpalatable choices. I wonder how do you teach a
fully autonomous vehicle judgment, the kind of judgment that
that professional truck driver has because of his or her many
years of experience.
I was thinking--and we've all been in these kinds of
situations where you have to make a split-second judgment on
whether you're going to swerve around a vehicle that suddenly
stopped or has spun on an icy road and risk going off the road
and hitting a tree, perhaps killing yourself and your
passengers, or a pedestrian or a bicyclist has darted in front
of you in the road, and yet if you slam on the brakes you're
going to be back-ended.
I mean, driving is not a simple task. It requires complex
judgments all the time. So how does a self-driving vehicle make
those kinds of judgments?
We'll start with the Administrator, and I'd like to go
right down the line.
Mr. Rosekind. I am going to use that to address two things.
You've just raised a great example of all the unknowns that are
out there still related to autonomous vehicles, and the
questions you've just raised--and there are many, many more
related to ethics--there are no answers right now. But we are
taking hands off the wheel, human hands, and putting them into
the hands of a coder, because that car will be programmed to
make decisions, basically.
So I think that's why, for example, people have questioned
us--it's part of our 15-item assessment. We put ethics in
there, and people have questions. When you think about when
people are really going to accept these and have confidence,
and we know the ethics are going to be an issue people are
going to ask about, there are no answers right now. That's why
it's in there for us to really look at the innovation of what
people bring to this.
But let me just add one element to that that nobody is
really talking about. I've been fortunate over about the last
decade to get to know Captain Gene Cernan, the Apollo 17
astronaut, the last man who walked on the moon. One of my
favorite sayings at safety meetings with him is, you know, I'm
never going to live long enough to make all the mistakes that
could kill me, right? So why does he go to safety meetings? For
just that reason.
So think about how all of us learned to drive a vehicle.
Our experience is those ethical decisions we're going to make
are based on our personal experience. Now, with autonomous
vehicles, if we collect all of this information and share those
cases where somebody died in some unique edge or corner case,
if that information were shared, no other person should ever
lose their life in that same situation because that information
gets shared with the entire fleet, with every vehicle that's
out there. That's the future that we could look toward,
including in these ethical issues. We get to make decisions at
some point that we could share and make sure that everybody
basically is allowed the same opportunity to be safe.
Senator Collins. Thank you.
Ms. Hersman.
Ms. Hersman. I would probably fall back on the 10 years
that I spent at the NTSB and say that we've got to take care of
the data. That means we've got to learn from the data. It's got
to be accessible. We have to have standardized, accessible
formats.
If you think about black boxes on airplanes, that's how we
learn. We know what happened. When it comes to autonomous
vehicles, or even something that's in-between, in order to
understand those decisions or those outcomes that you talked
about, who was in charge? The car, the human, or some
combination of the two?
We've got to share that data. When we talk about mistakes
that occurred or close calls or things that happened, if we
don't have all manufacturers willing to share those lessons
learned, we're going to have to have each provider learn the
lesson anew every time. So we've got to look at this
holistically and say there will be failures. Will there be
deaths? Absolutely, as this technology rolls out. Will there be
things that happen that we don't expect? I can't tell you how
many investigations we went into and people said this was never
supposed to happen. This scenario was 10 to the minus 9th. It
was never going to occur, and it did.
So things will happen, but that data to me is incredibly
important to how this rolls out and to maintain that confidence
going forward. We've got to learn.
Senator Collins. Thank you.
Mr. Brubaker.
Mr. Brubaker. Yes. So these ethical issues when they get
raised cause me to lose the rest of my hair here. Part of the
issue is we're all thinking about this in a human context.
We're all sort of applying that judgment that we have to make,
and what we're not cognizant of or what we're really not
thinking through is that the level of situational awareness
that these vehicles are going to have far exceeds that of a
human. You're talking about redundant sensors. You're talking
about the ability to see far beyond what a human eye can see
and have this 360-degree, 24/7 situational awareness.
The data is important. I agree with the two prior speakers.
But the thing that we have to remember is that this is all
about artificial intelligence. This is about instantly
assessing the situation, constantly assessing it, calculating
probabilities and reducing the risk of being in an emergent
situation where you've got to make two really unpalatable
choices.
So my whole view on this, and this is from having lots and
lots of conversations with people in the AI world and people
who are looking at this and building the sensor suites and
mashing up this technology, that those ethical considerations
are--I don't want to call them a red herring, but in many cases
the whole notion is to create the kind of technology that will
avoid those emerging situations in the first place.
Senator Collins. Thank you.
Dr. Kalra.
Dr. Kalra. I agree with the other speakers in almost every
regard. I do want to say that when we raise these vivid
examples of swerve or do this, people don't actually make
ethical judgments when they decide what to do. They make snap
judgments. They don't have time. They barely react. It's just a
knee-jerk thing that you do without thinking about the
consequences. The ethical things that we do when we drive, or
the unethical things, are driving when we're intoxicated, or
texting while we drive, and autonomous vehicles won't make
those kinds of ethical mistakes.
But the ethical judgments that autonomous vehicles will
have to make I don't believe are the pedestrian versus the
driver but really about how the autonomous vehicle distributes
risk on the road at every moment of driving. The example I'll
give is when I'm driving and there's a bicyclist next to me,
I'll drive in the middle of the lane because I want to
distribute the risk differently between the bicycle, myself,
and the other people who might be on the road. Autonomous
vehicles will have to make those judgments, and I think
attention needs to be placed on how do autonomous vehicles
ethically distribute regular risk on the road rather than in
the sort of imminent crash cases, and there's no answers to
this because even ethicists don't have the thing that's the
right thing to do. But I would urge us to put our attention on
those everyday kinds of ethical problems.
Senator Collins. Thank you very much.
Senator Reed, do you have anything else?
Senator Reed. No.
Senator Collins. I want to thank you all for testifying
today. This was a superb panel.
This technology is so fascinating and moving so quickly,
and the fact that it could substantially reduce or perhaps one
day even eliminate the 94 percent of crashes that are
attributable to human error is truly an astonishing fact, and
that's one fact that prompted me to call this hearing.
I appreciate your advancing our understanding of the
issues, the technology, and the policy considerations that we
will have to face.
SUBCOMMITTEE RECESS
The hearing record will remain open until next Wednesday,
November 23rd, 2016. There may be additional questions
submitted by us or by our colleagues for the record. We very
much appreciate your cooperation, and we look forward to
continue to working with you.
This hearing is now adjourned.
[Whereupon, at 3:53 p.m., Wednesday, November 16, the
subcommittee was recessed, to reconvene subject to the call of
the Chair.]
MATERIAL SUBMITTED SUBSEQUENT TO THE HEARING
[Clerk's Note.--The following outside witness testimonies
were received subsequent to the hearing for inclusion in the
record.]
Prepared Statement of John Bozzella, President and CEO, Association of
Global Automakers
On behalf of the Association of Global Automakers (``Global
Automakers''), I am pleased to provide the following statement for the
record of the Senate Committee on Appropriations Subcommittee on
Transportation, Housing and Urban Development, and Related Agencies
hearing entitled ``The Automated & Self-Driving Vehicle Revolution:
What Is the Role of Government?'' Global Automakers represents
international automobile manufacturers that design, build, and sell
cars and light trucks in the United States. These companies have
invested $52 billion in U.S.-based facilities, directly employ more
than 100,000 Americans, and sell 47 percent of all new vehicles
purchased annually in the country. Combined, our members operate more
than 300 production, design, R&D, sales, finance and other facilities
across the United States.
The automotive industry is in the midst of an unprecedented wave of
technological innovation that is redefining how we think about
transportation. Advancements in connected and automated vehicle
technology promise to enhance mobility, help save lives, improve
transportation efficiency, and reduce fuel consumption and associated
emissions. Over the past several decades, our members have made
tremendous strides in safety by improving vehicle crashworthiness;
today, automakers are deploying crash avoidance technologies to help
prevent crashes from occurring altogether. Our members are at the
forefront of this innovation, as they have made, and continue to make,
substantial investments in the research and development of automated
vehicle systems and other advanced automotive technologies.
With the introduction of advanced sensors such as cameras and
radar, a number of vehicles on the road today already provide automated
functionality through advanced crash-avoidance and convenience features
like automatic emergency braking, lane keeping assist, and adaptive
cruise control. These systems, which are foundational to the
development of more highly automated systems, are designed to provide
support to the driver only in certain situations. As these systems
become more advanced, a vehicle's capability to operate without active
control by the driver will increase.
The next breakthrough in vehicle safety, and a critical technology
for realizing the benefits of automated driving, is Dedicated Short
Range Communications (DSRC) connected car technology. This technology
supports vehicle to vehicle communications (V2V) allowing cars to have
greater 360-degree situational awareness. Through DSRC, vehicles can
speak to each other and to surrounding infrastructure at the rate of
ten times per second to avoid crashes and improve mobility. This
technology is on the road today; pilot projects and deployments around
the country are using DSRC supported applications to demonstrate the
value of connected mobility to the traveling public. Soon, and
increasingly into the future, we will share our roads with automated
vehicles; V2V has the ability to connect all vehicles, regardless of
mode or level of automation.
While we are indeed at the cusp of a transportation revolution
through connected automation, these transformations are not inevitable
nor accidental. Public policy can either spur investment and
innovation, or hinder them, depending on which policy choices are made.
Effective public policy on connected and automated vehicles should have
two components. First, it should be flexible and provide room for
innovators to develop, test and sell new technologies. Overly
prescriptive and rigid regulation would slow and limit innovation.
Second, manufacturers should be able to build vehicles and systems that
can be sold in all fifty States. A patchwork of inconsistent laws and
regulation would be unworkable.
The National Highway Traffic Safety Administration (NHTSA) Federal
Automated Vehicle Policy, released in September 2016, provides a policy
framework that is more flexible and nimble than the formal rulemaking
process, and recognizes that technology can advance more rapidly than
regulation. Last month, NHTSA issued its Cybersecurity Best Practices
for Modern Vehicles to complement the important efforts already
underway within the Automotive Information Sharing and Analysis Center
(Auto-ISAC) to develop industry-led best practices to enhance vehicle
cybersecurity as systems become more electronic and connected. Issues
of consumer privacy have also been addressed through the automakers'
consumer privacy protection principles. These actions, by Federal
regulators and industry, help spur the development of live-saving
technologies and ensure that the public has confidence in them.
The NHTSA's Federal Automated Vehicle Policy is intended to address
a number of key policy questions and is a positive first step to
demonstrate Federal leadership. The Policy is divided into four main
sections. First, the Vehicle Performance Guidance for Automated
Vehicles outlines recommended practices for the safe pre-deployment
design, development and testing of highly automated vehicle systems
prior to the sale or operation on public roads. The Guidance was
designed to be flexible and dynamic; it is intended by NHTSA to
highlight important areas that manufacturers should consider and
address as they design and test their systems. The Guidance provides
for a ``Safety Assessment Letter,'' a voluntary tool by which
developers would communicate to the agency how it addresses fifteen key
safety areas in designing their vehicles and systems. NHTSA is in the
midst of developing a template for the Letter, and we believe NHTSA
should establish a clearly defined and practicable approach that does
not create an undue administrative burden that could slow innovation.
It is also our expectation that NHTSA will not use the Guidance and the
Safety Assessment Letter as a mechanism for ``premarket approval'' (or
``premarket disapproval'') of automated vehicle technology, as this
would extend beyond the agency's current authority.
Second, the agency has developed a Model State Policy which seeks
to provide guidance to the States in order to help support a more
uniform nationwide approach to automated vehicle policy. While the
Policy cannot in itself preempt State action, it does set a clear
marker in defining the roles of State government in addressing issues
related to vehicle automation. We support the strong statements in the
Policy that affirm that ``[t]he shared objective is to ensure the
establishment of a consistent national framework rather than a
patchwork of incompatible laws,'' and that ``[the] Guidance is not
intended for States to codify as legal requirements for the
development, design, manufacture, testing, and operation of automated
vehicles.''
However, despite the guidance in the Model State Policy, several
States are in the process of establishing their own regulatory programs
for automated vehicles. In some instances, State departments of motor
vehicles would assume the responsibility of determining whether a
particular automated vehicle or system is safe and thus may be sold or
operated in the State. Such State-by-State regulations would present a
significant obstacle to the future testing and deployment of automated
vehicles. While the Model State Policy clearly delineates the Federal
roles and States' roles, it does not clearly limit or prevent State
regulation of automated vehicle design and performance.
Additionally, we have some concerns with certain recommendations in
the Model State Policy that encourage States to regulate automated
vehicle test programs. Already, we have seen State proposals to require
manufacturers to obtain an ordinance authorizing testing from each
local jurisdiction in which testing will be conducted. However, Federal
law authorizes original manufacturers to conduct on-road test programs
and authorizes NHTSA to regulate test programs. Allowing a patchwork of
State and local test requirements for automated vehicle testing would
significantly obstruct the development of these vehicles. We are open
to working with NHTSA and Congress to ensure there is a path forward
for automated vehicle deployment without unnecessary obstacles at the
State level.
Third, the Federal Policy provides a useful description of the
agency's current regulatory tools, which includes issuance of safety
standards, interpretations of the meaning and application of standards,
and exemptions from standards, as well as the agency's ability to take
enforcement action regarding safety related defects. Each of these
tools could have a valuable application in facilitating and regulating
the entry of automated vehicles into U.S. commerce. At the same time,
we must consider the long-term efficacy of these tools in determining
whether other regulatory and non-regulatory policies may be appropriate
and necessary in the future. It is important that any action be data
driven and technology neutral.
Finally, the agency discusses the potential new tools and
authorities that may be necessary in addressing the challenges and
opportunities involved in facilitating the deployment of automated
vehicles. We agree with NHTSA's assessment that new authorities could
assist the agency in facilitating the development and introduction of
automated technology. However, imprudent legislation in this area could
have the opposite effect and delay technology development. For example,
we see no basis at all for any change to the self-certification system
for vehicles. The Federal Policy's discussion of the Federal Aviation
Administration (FAA) process of ``premarket approval'' is not practical
given the structural differences between the automotive industry and
aviation sector, and implementation of such an approach could
significantly slow innovation. Similarly, the Safety Assessment Letter
should not be used as a means to prohibit testing or deployment of
technology without adequate data to support an unreasonable safety
risk.
We believe that NHTSA's Federal Automated Vehicle Policy is an
important first step in the development of a flexible and nimble
approach that can adapt to the pace of technology. However, the
document requires further clarification and refinement to achieve these
goals. Global Automakers is currently preparing comments on the NHTSA
guidance and will provide a copy to the Committee upon submission to
NHTSA. Additionally, we agree with NHTSA that the agency should update
its Federal Automated Vehicle Policy and regularly review the Policy,
as it is designed to never be frozen or final. Global Automakers and
its members remain committed to working with Federal, State, and local
governments to ensure there is a flexible, consistent framework for
automated vehicle technologies so consumers can fully realize the
benefits as quickly as possible.
In addition, the Federal Government must move expeditiously to
establish a framework for the deployment V2V communications through
DSRC connectivity. NHTSA is developing a new vehicle safety standard
that would require vehicles to be equipped with DSRC technology. Global
Automakers looks forward to the release of the proposed rule, and will
continue to work with the Federal Communications Commission to ensure
that the 5.9 GHz Safety Spectrum remains free from harmful interference
to support DSRC technology.
The automobile industry continues to provide innovative
technologies with demonstrable safety, mobility, and environmental
benefits. To achieve these benefits, there must be close collaboration
and coordination among and between government, industry, academia, and
other stakeholders. Global Automakers and our member companies believe
that connected and automated vehicles represent the next giant leap
towards our shared long-term goal of safer and cleaner, and more
efficient vehicle transportation.
______
Prepared Statement of Property Casualty Insurers Association of America
There are public perceptions that auto accidents and insurance
costs are decreasing. In fact, our roads are becoming increasingly
dangerous and auto repair and medical costs are increasing. According
to the National Highway Transportation Safety Administration (NHTSA),
17,775 people died on our Nation's roads in the first half of 2016.
Traffic deaths are increasing at the fastest rate in 50 years, with a
10.4 percent increase in the first 6 months of this year. Even adjusted
for the increase in vehicle miles traveled (VMT), the fatality rate
increased 6.6 percent to 1.12 per 100 million VMT. Non-fatal injuries
are on the rise as well, increasing 28 percent since 2009, according to
the National Safety Council. Someday, self-driving cars may reduce the
number of accidents and deaths. However, the potential of automated
vehicle technology stands in sharp contrast to what is happening on our
roads today.
The Property Casualty Insurers Association of America (PCI) is
composed of nearly 1,000 member companies, representing the broadest
cross section of insurers of any national trade association. PCI
members write $202 billion in annual premium, 35 percent of the
Nation's property casualty insurance. That figure includes over $97
billion, or 42 percent, of the auto insurance premium written in the
United States. PCI's analysis has found that since 2013, auto claims
frequency has increased nearly 5 percent, increasing the overall cost
of claims by more than 18 percent. PCI has analyzed the recent increase
in auto insurance claim frequency and found strong correlations with
traffic congestion and distracted driving, weaker correlations from
increasing populations of novice and older drivers, and some
correlation with liberalized marijuana laws.
While it is important to prepare for the automated vehicle of the
future, we urge policymakers to continue to focus on the auto safety
challenges that face us today, such as distracted and impaired driving.
H.R. 22, the FAST Act, provides for increased public awareness,
improved enforcement, and establishing an enforceable impairment
standard for drivers under the influence of marijuana. These provisions
are critical to reducing accidents, injuries and deaths on our Nation's
roads. The importance of addressing these issues was also the subject
of a bipartisan letter from 23 members of Congress to Transportation
Secretary Foxx urging prompt implementation of these FAST Act
provisions.
[The bipartisan letter is attached.]
NHTSA recently unveiled its ``Federal Automated Vehicle Policy,''
which is intended to provide guidance for States on the testing and
deployment of highly automated vehicles (HAVs). While mentions of
insurance are few, the new policy raises issues that are important to
the automobile insurance market as it seeks to adapt and develop new
products to meet consumer needs.
recognition of state regulation of insurance and liability issues
NHTSA's policy identifies the following as Federal
responsibilities: setting and enforcing safety standards for motor
vehicles, recalls, promoting public awareness and providing guidance
for the States. NHTSA's policy also recognizes that it is the State's
role to license drivers and vehicles, enforce traffic laws, regulate
motor vehicle insurance, and legislate regarding tort and criminal
liability issues pertaining to automated vehicles. PCI shares the view
that the States should continue to have primacy on motor vehicle
insurance and liability issues as they do today, and we support NHTSA's
recognition of that role.
NHTSA's policy also repeats the recommendation from its 2013
guidance that entities testing automated technology should provide
proof of financial responsibility coverage of at least $5 million. PCI
has not taken a position on this coverage requirement. However, as HAVs
are deployed for public use, States will need to consider what, if any,
changes need to be made to existing State motor vehicle financial
responsibility laws.
data collection and access
As policymakers consider what data should be collected and retained
by automated vehicles, it is essential that insurers have reasonable
access to the data for providing customer service for claims handling
and underwriting purposes. In many auto accidents, apportionment of
liability is likely to hinge upon whether or not a human driver or the
vehicle itself was in control and what actions either the driver or the
vehicle took or did not take immediately prior to the loss event.
Access to data for insurers will speed claims handling and potentially
avoid disputes that could delay compensation to accident victims.
Access to historical anonymized data on the different automated vehicle
systems will also be important to help insurers innovate and develop
new insurance products as the nature of the risk changes.
conclusion
Automated driving technology holds great promise for the future.
Implementing clear policies on the Federal and State roles in
regulating automated vehicle technology and ensuring that insurers have
access to vehicle data on reasonable terms to efficiently handle
claims, develop products and underwrite are essential to that future.
However, policymakers must not lose sight of the auto safety issues
that face us today. We look forward to working with policymakers at the
Federal and State level to reduce accidents on our roads today and in
the future.
Attachment
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]