With tourist and job centers quiet, restaurants serving in street-side parklets, and residents walking and biking on ‘slow streets,’ the pandemic’s impact on San Francisco streets can’t be overstated. But one thing looks remarkably constant. Autonomous vehicles (AVs) still drive our streets every day. We see the future coming, and now is the time to set federal AV priorities.
The most urgent priority is developing safety standards. Watching testing on San Francisco streets makes me an AV safety optimist. Developers appear to be taking a careful approach. Speeding, running red lights, and failing to yield to pedestrians kill people on American roads every day. If AVs are programmed to comply with traffic laws and accurately predict the actions of other road users, there is reason for hope.
But appearances and hope are different from public accountability. Measuring AV safety is not simple. Washington should lead by ensuring automated driving is actually safer than human driving before letting AVs flood public roads. Developers are testing on public roads in at least 10 states, but there is not one federal safety standard governing automated driving. Nothing requires developers to demonstrate proficiency in recognizing and classifying ALL human beings on our roads — including those who walk, ride, roll, or use a cane. Nothing requires AV developers to report collisions to the National Highway Traffic Safety Administration (NHTSA) or give NHTSA sensor data to assess whether AV driving caused a crash.
NHTSA doesn’t need legislation to develop AV safety standards, but the Agency needs new funding and expertise and should consider a new approach. When Congress adopted the Motor Vehicle Safety Act in 1966, there were already more than 75 million cars on U.S. roads. Safety standards were built from backward looking analysis of crashes and complaints. The data feeding those analyses are human lives.
There has only been one fatal AV crash on U.S. roads. The NTSB crash analysis shows that human error isn’t just the domain of drivers. AV designers and executives also make human errors – as well as consequential economic and other judgments about AV driving decisions. When AVs number in the hundreds and not millions, NOW is the time for NHTSA to build a forward-looking safety accountability system to catch errors before they cause injuries and deaths. Developers want Congress to authorize deployment of hundreds of thousands of AVs in just a few years. Congress should give NHTSA resources to get ahead of this curve.
Some developer proposals are ill-timed. An industry group recently asked the Federal Highway Administration to change the Manual on Uniform Traffic Control Devices (MUTCD) to make driving easier for automated drivers. A stop sign is a red octagon from California to Connecticut thanks to the MUTCD. We need better traffic engineering standards, but the draft regulation would require every city and state to adapt streets – starting now — to the weaknesses of today’s machine vision technology. Some of the requested changes may also help humans drive more safely, but others are designed for AVs alone and are premature. This would come at huge taxpayer expense while gas tax revenues dwindle and state and local budgets are in crisis. Before asking taxpayers for help, developers need to show that AVs can operate safely on the streets we actually have — complete with worn paint, unexpected behavior by other road users, low-tech signs, people biking, rolling and pushing strollers, and other forms of predictable unpredictability found on real streets. In the meantime, federal policy makers should expedite other initiatives.
First, the Washington agenda for infrastructure investment should free city streets from standards that were developed for highways and suburban roads. Whether drivers are humans or machines, urban roads with high volumes of pedestrians and cyclists should be built using urban multi-modal design standards that put people first.
Second, as developers build AV fleets from the ground up to provide ride services to the public, Washington should insist that AV services be born accessible to people of every age and ability — including people who use wheelchairs. Third, while it is exciting to see the auto industry throwing its weight behind EVs, electrifying cars is not enough to achieve energy efficient transportation. AVs will be deployed first in “robo-taxi” fleets in cities. Washington should review the recent history of ride-hailing and consider whether and how AVs can be a climate solution. Even if AVs are EVs, we need a cautious approach to business models that put small passenger cars on city streets while drawing people from more efficient and equitable modes.
If AVs are stuck in city traffic, they will not help anyone — even those who hope to profit from their deployment. In San Francisco, we hope to work closely with developers to demonstrate how AV passenger services can be deployed to support the overall efficiency of our urban transportation network by, for example, giving people rides when transit service is limited and helping people get from our steep hills to high frequency transit lines. Washington should expand support for collaborative research and development focused on the overall efficiency of our transportation system.
The bottom line is that transit remains the most energy and space efficient way to move large numbers of people long distances in and around cities. To keep up with Europe and Asia, make room for AVs on city streets, and prevent climate catastrophe, expanding Washington investment in world class public transit is essential.
Jeffrey Tumlin is the Director of Transportation at the San Francisco Municipal Transportation Agency. He is the former director of strategy at NelsonNygaard Consulting Associates, a planning and engineering firm that focuses on sustainable mobility. His major development projects have accommodated millions of square feet of growth with no net increase in motor vehicle traffic, winning awards from the U.S. General Services Administration, Institute of Transportation Engineers, American Planning Association, American Society of Landscape Architects, Congress for the New Urbanism, and Urban Land Institute. He is the author of “Sustainable Transportation: Tools for Creating Healthy, Vibrant and Resilient Communities” (Wiley, 2012).