Dispatches from the Automated Vehicle Symposium 2018 – July 11

For a recap of the July 10 proceedings, see my previous post here.

At the second (or third?) day of the Automated Vehicle Symposium 2018 in San Francisco, we had a diverse set of sessions, covering everything from labor force impacts to truck platooning, and from modeled impacts of AV adoption in several cities to AV interactions with other roadway users.

I’ll note this conference does an outstanding job broaching every topic imaginable related to AVs, well beyond the technology itself and the operational hurdles it needs to overcome. 

In terms of labor force impacts, Cornell University’s Erica Groshen spoke about Preparing U.S. Workers and Employers for an Autonomous Vehicle Future. She noted that previous technological disruptions have seen economic benefits produced broadly for the general population, though some of those gains are unevenly allocated.

Conversely, the costs of those transformations were heavily concentrated on those workers negatively impacted by them, and those costs are both high and precede the widely distributed benefits. Her report estimates that 1.3 million to 2.3 million workers will be displaced between 2018-2051, a consequential impact roughly half of the “China shock” of the past (roughly) several decades, but a manageable impact nonetheless.

While job losses are eventually recovered, this cost/benefit lag and disparity can produce significant social frictions, and therefore she recommends proactive policies to ease our transition to driverless systems. Rather than follow a laissez-faire or reactionary approach, she argued for embracing an “investment” approach whereby some economic gains from automation are captured and redirected to mitigate impacts on short-term “losers.”

John Moavenzedah of the World Economic Forum shared findings from a study performed with Boston Consulting Group that modeled the impacts of AVs on several city transportation systems, primarily Boston’s. This study did not provide a clear timeline for its conclusions, but rather made date-agnostic adoption assumptions based on consumer surveys and 2030 cost-of-service projections.

Key findings include:

• In Boston, transit ridership would drop significantly in urban areas, from 47% of all vehicular (non-bicycle/pedestrian) trips today to 33% in the future. Transit’s mode share in suburban areas actually improved from 30% to 32%, presumably because mobility-on-demand improves last-mile connectivity for faster, higher-capacity commuter rail.
• AV adoption varied considerably by city, and was correlated to both income and age. Higher income areas would adopt AVs at a greater rate, due to their preference for and ability to pay for mobility-on-demand services. Adoption was also highest in the 28-45 age bracket at 38%, versus a low of 21% for the 65+ demographic.
• Quantity of vehicles on the road would decrease, and parking demand would drop by nearly half, but vehicle miles traveled would increase.
• Travel time would improve marginally, but cities can exert significant influence on traffic flow through effective policy levers. Occupancy-based pricing schemes could reduce citywide travel time by 15.5%, converting street parking by 10%, and installing dedicated AV lanes by 8.3%.

Finally, he shared an overview of the AV pilot programs accompanying the study, and recommendations for other cities looking to launch their own pilot programs. The most compelling of these included building public awareness early, possibly through gimmicky events like “AV petting zoos” and “robot block parties” that can expose thousands of people to the technology, as well as publicizing regular updates on testing progress.

Find the full report here. It had actually been on my docket for a couple weeks now, and look forward to finally reading the entire piece.

Other sessions of note from the morning program:

• Peleton CEO Josh Switkes and then Embark CEO Alex Rodrigues discussed various dimensions of automated trucking. Switkes explained the technical progression towards platooning, and safety benefits derived from that implementation. Rodrigues described the current (and future) long-haul trucker shortage, and the way in which Embark is trying to solve it through highway automation.
• A panel of regulatory experts addressed various…regulatory issues related to AVs. One common theme was the need for any regulations implemented to be flexible and highly iterative to keep up with rapidly evolving technology.
• The University of Leeds’ Natasha Merat discussed he research on interactions between AVs and other road users. She noted that contrary to popular belief, very little external communication actually occurs between vehicles and other road users, and that most potential interaction-demanding situations are resolved before they actually arise, mostly through “adjusting kinematic motion.”

The breakout session I attended was titled OEM/DOT Dialog on Dedicated Lanes, Work Zones, and Shared Data. Law enforcement was a serious concern for any dedicated AV lanes, both in terms of ensuring vehicle authorization and accident/incident response. If the lane has to shut down, then what happens? How does law enforcement remove an AV from a dedicated facility? Is it even law enforcement’s responsibility to do so?

Utilization was also a major concern. Several panelists raised equity concerns and potential political backlash from road users who don’t have access to dedicated facilities, especially if AV adoption is low or deadheading (zero-occupancy) AVs are allowed access.

I asked a question about the maintenance obligations from dedicated AV truck lanes, positing the induced demand from such facilities could impose unsustainable financial burdens on DOTs. One panelist noted freight alliances have already pushed for these types of lanes, and acknowledged serious risks to assess before committing the state to construction and maintenance. Another said load management would become increasingly critical, but fortunately electronic logging tools are making that easier to track and enforce.

Another insight is that we may require some sort of clearinghouse for work zone and construction data that is both standardized and easily accessible for any AV operator. Technology improvements could help produce more accurate and higher fidelity data of benefit to all parties involved. And finally, DOTs are increasingly willing to enter into data sharing agreements with private operators, though data security and business model development are still open questions.