How Self-Driving Cars Can (and Should) Improve Transit

The recent 2014 Automated Vehicles Symposium brought the opportunities and perils of vehicle automation clearly into view. The week-long event—largely, though not completely, open to the media—featured keynote remarks by self-driving technology providers, car makers, federal and state officials, and academic researchers. Speakers highlighted the current and near-term trajectory of the technology's fundamental components, detailed the ongoing regulatory efforts to bring self-driving vehicles into the system safely, and debated the likely consequences of highly-automated cars, trucks, and buses. 

Many of the event's speakers provided a one-way flow of information, with much of the interactive action occurring in breakout sessions (which were closed to the media). I had the pleasure of participating in the transit and shared mobility session, which emphasized intelligent rollouts that take advantage of existing mass transit systems to avoid a nightmare scenario of drastically increased vehicle miles traveled (VMT). As Rod Diridon, Sr. of the Mineta Transportation Institute noted, substantial resources have been invested in existing transit systems, and it would be foolhardy to write them off. That is, the goal shouldn’t be replacement of existing systems; rather, the goal should be adaptation and cooperation.

During his plenary session presentation, Mike Gucwa of Stanford University succinctly described how automated vehicles (AVs) could lead to more VMT, based on his simulations of San Francisco Bay Area travel. Assuming current residential patterns, a 4-8 percent VMT increase is plausible, simply due to more efficient traffic flows (because connected self-driving vehicles could travel closer together, improving speeds) and driving-hassle reductions (as perceived values of travel time fall, because AV occupants are free to pursue other activities while in the car). At the same time, Gucwa projected average travel times to improve overall, even with more VMT. Ken Laberteaux of Toyota raised additional land use issues (e.g., easier commute = more sprawl?) that could lead to further increases in car travel. Because average time spent per day on personal travel seems to be about the same across a wide range of times and places (known as the travel time budget), at about 1.2 hours per day, faster travel may lead to new market pressures for exurban living. Moreover, the impacts discussed by Gucwa and Laberteaux are conservative in that they do not account for the possibility of AVs traveling with no human inside, in search of inexpensive, remote parking or to pick up other travelers.

The good news that merges from these questions is that recent technological innovations may be melded with new or existing shared mobility platforms to offer a pathway for avoiding these unfortunate scenarios. Susan Shaheen, from the University of California, Berkeley, detailed synergies with shared AV (SAV) car-sharing frameworks, Neil Hoose spoke about how SAVs could help solve the first-mile and last-mile problem of connecting mass transit systems to travelers’ ultimate destinations, and Adriano Allesandrini of the University of Rome La Sapienza shared his experiences in deploying a low-speed 12-passenger SAV pilot program across multiple European cities. Related developments discussed by Joseph Kopser showed how new platforms like his company Ridescout could help provide more seamless linkages across modes—whether public transit options, bikeshare programs, or, someday, SAVs. Sweden’s Christer Lindstrom continued the theme in his discussion of Stockholm’s challenges as a city constrained by islands and bridges. The city’s recently developed transport concept envisions a series of main-haul mass transit trunk lines connected by smaller local feeders to neighborhoods and businesses—a vision building on the earlier speakers' thoughts.

So where do we go from here? As many speakers noted, ultimately there will likely be a cocktail of privately owned and shared AVs, in addition to the human-driven vehicles that exist today. A big part of the value proposition of SAV systems will be pooling multiple passengers in vehicles (so that each pays only a fraction of the cost), and providing "last-mile" connectivity with higher-capacity mass transit. Land use and road pricing policies that encourage density and discourage unoccupied travel may also be required if we are uncomfortable with the private market being the sole deciding factor in the impact of AVs on city form. Though, as Lindstrom made clear, local land-use planning must be tailored and context-specific, rather than painted with a broad brush.

None of the experts assembled expect the private car to wither away, even though shared fleets of AVs offer exciting possibilities. Speaking personally, I place great value on owning a vehicle with fixed car seats for my children and locked mobile storage for a stroller and other items in the back. But having access to a reliable and affordable SAV system that links us to Salt Lake City’s excellent TRAX mass transit rail lines would allow my family, and I suspect many others as well, to save what we currently spend every month on our household’s second car. "Wheels within wheels" is a fair way to describe the complexity of introducing vehicle automation to the general public, and there are good reasons to closely monitor developments in the process. But we must not lose sight of the fact that there are tangible bottom-line benefits to the prospect of automated vehicles, and those benefits will provide excellent motivation for most people to adopt AVs when they become available.

Dan Fagnant is completing his Ph.D. in civil engineering at the University of Texas, and starts as an assistant professor at the University of Utah in Fall 2014. He is the lead author of the Oct. 2013 study Preparing a Nation for Autonomous Vehicles: Opportunities, Barriers and Policy Recommendations (published by the Eno Center for Transportation). A full listing of Dan's academic and general-readership publications can be found here.