At this point, it seems clear that autonomous vehicles are on the verge of technical feasibility. Just last week, Waymo announced that it is testing self-driving minivans without a human back-up in the front seat. Its employees will be riding in the back with an emergency stop button – but no steering wheel. But do these technical advances mean that we’re ready for AVs? How should we manage the non-technical aspects of AV deployment to ensure they achieve promised improvements in safety and accessibility?
I decided to write this article to address these issues after participating in the Intelligent Transportation System World Congress earlier this month. There were tons of panels focused on autonomous vehicles, and I was lucky enough to be speak on one that dove into the critical questions for civic leaders and transportation professionals. We went beyond technical readiness to ask ourselves if should we deploy AVs, and, if so, how should we deploy them?
The name of the panel was "Automation as a Solution: Addressing 21st Century Mobility Challenges through AV Deployment.” I’ve collected a few of the key themes and interesting remarks in this post. Many of these ideas were generated by my co-panelists: Avery Ash from INRIX, Karla Taylor from City of Austin, and Dave Vermer from HMI Tech. To avoid any misquotation or misinterpretation, I’m not going to attribute specific ideas to specific people.
Personally, I don’t think that AVs are inherently good (or bad!). If they can help mitigate some of the serious challenges facing our environment and our economy, and if they bring greater prosperity to all segments of society, then I’m all for them. But what if they’re just new places for rich people to binge watch Netflix in? What if they only make it easier for people to live further out and isolated in the suburbs? If that’s the case, then I’m deeply concerned about them.
Coming out of the conference, there seemed to be a general consensus that AVs should be shared and electric as a key part of achieving the positive version of the AV future. And I agree. How can the transportation industry work to ensure that AVs will be deployed in a manner that is good for accessibility, affordability, and jobs equity?
Require AV Manufacturers to Share Their Data
Data about AVs should be easy to collect, at least in a technical sense. But historically it has been really hard to collect data about everyone else on the road, and that is critical for us to rigorously understand how AVs are impacting our system very early on, as soon as they are adopted by consumers. To achieve that rigorous understanding, we absolutely need good data from AVs – data that is relatively complete and demarcated as “AV only-data." But we also need data on everyone else.
We didn’t do a great job with this concept when transportation network companies (TNCs) such as Uber and Lyft hit the roads. First, we still don’t have much useful data about TNCs from the providers themselves. Second, this issue is compounded by the fact that most planners still use very clunky techniques like surveys to measure “everyone else.” Thus, we have very little idea of how TNCs are being used, and the evidence we do have about their impact is usually anecdotal.
As a result of this lack of data, we cannot and have not done a good job of managing the arrival of TNCs, nor have we effectively pushed for these new services to benefit all parts of our cities. If we take the same approach with AVs, we are virtually guaranteeing that the same situation will unfold.
Facing the Infrastructure Challenges Head-On
The real infrastructure challenges for AVs are not high-tech. They’re low-tech. The most significant one my co-panelists and I have heard is striping, or painting clean lines on the roads and keeping them maintained. AVs perform best with clean and maintained basic roadway infrastructure like painted lines, consistent curbs, and no potholes.
This type of infrastructure challenge is not new: Clearly painted lines and no potholes also make driving for humans much safer and easier. And yet, many communities are struggling to keep these basics up to par. What about AVs make us think that we can avoid the systemic policy and funding issues that drag down our ability to maintain the fundamental infrastructure of our roadways?
Instead, cities seem to be distracted by companies asking them to spend extraordinary amounts of money on “connected IoT” and other infrastructure investments that are far less critical - to AVs and to everyone else - and way more expensive to implement. But if we don’t put our collective "house" of infrastructure into order first, AV deployment won’t be successful.
Don’t Start With Consumers
The vision for AVs does not have to be focused on personal vehicles taking drunk people to home from a bar! Some of the most interesting applications are in commercial vehicles, utility vehicles, and delivery vehicles. These commercial market segments are also easier to manage from a regulatory perspective: There are fewer actors in the market (when compared to consumers), and these actors are already accustomed to regulatory compliance.
Where Do We Go From Here?
Bringing this together, here’s a proposal: I think the first AVs should be maintenance vehicles that make the infrastructure and road network better for all. They can start as super-Roombas that sense potholes and chipped line paint. Maybe they can send messages that deploy fix-it teams, or maybe they can fix issues themselves. The streets will get better and safer for everyone, and they will also become AV-friendly. As our fleets of self-driving super-Roombas fix our highways, the industry can work with manufacturers to ensure our data collection systems are robust and seamless. We can test the data that these maintenance teams sends to ensure the information provided is complete. With an effective data collection system in place prior to consumer adoption, we will be ready to measure the impact of widespread AV deployment as it occurs.
I know this may result in jobs lost for people working in road maintenance. I suggest that we already have a lot of unfilled jobs and work in this field. Plus, the companies that deploy these maintenance AVs as pilots could also be asked to provide AV maintenance or management training in the community as part of the pilot.
The conversation about AVs has just begun, it’s complex – but it’s starting to seem like we’re running out of time to get this right. For example, the New York Times magazine was almost completely about AVs this weekend. To that end, we welcome more ideas and feedback about AVs, and in particular about how data can help. Let us know in the comments below!