Self driving car, the robot way.
Over six hundred experts in the driverless mobility space converged on San Diego in late July to share progress and debate what the future should look like on our roads. As noted in Part One of this article series, the Automated Road Transportation Symposium (ARTS) is a unique assembly of tech geeks, social scientists, government regulators, and the occasional skeptic. At the event, leading companies deploying driverless vehicles laid out their latest progress and plans for expansion. Participants from overseas provided their unique viewpoints and rolled out their plans for the age of autonomy.
ARTS is famed for having an audacious array of breakout sessions, which are designed to stimulate debate and address the hard questions. These sessions are organized by a dedicated army of academics and practitioners. I reviewed the overall event in Part One. In this Part Two, I dive a little deeper to highlight some viewpoints and debates heard in the breakouts, as well as other key voices. That, plus some intriguing emerging areas of innovation.
Will robotaxis deploying custom shared-ride vehicles fulfill the urban dream of motorists ditching their cars?
Many researchers attending ARTS have focused their careers on improving urban mobility and see robotaxis as a key part of the answer – as long as they’re supporting ride-sharing. In this space, May Mobility and Beep are leading operators for deploying services under contract to city mobility agencies. When public money is involved, ride-sharing is fundamental (buses, trolleys, etc.), with the goal being to enhance personal mobility while reducing traffic congestion and greenhouse gas emissions.
But public money can be scarce. Funding processes are typically achingly slow. Robotaxi services risking their invested capital need to move fast. Here, the profit imperative is such that what their passengers want, they get. Based on experimentation by Uber and Lyft in recent years, customers really want to have their ride-hailed car all to themselves. No sharing.
Nevertheless, some of the startups uphold the larger vision. Waymo, highly active in ARTS, has been a leading voice in affirming the basic goals of enhancing urban mobility.
Despite significant skepticism in past years from some city officials and the more strident researchers, in ARTS discussions I sensed an implicit acceptance that the tech is at a high level of performance and that robo-based personal mobility has a role to play going forward. Compared to prior years, the conversation has definitively moved from “if” to “how.” Shared mobility with automated vehicles is gaining steam in parallel with robotaxi rollouts. In the case of Zoox, for instance, their robotaxi is designed specifically for shared mobility. Will these two models increasingly overlap such that the distinctions diminish?
And by the way, no one at ARTS was talking about the vaporous Tesla robotaxi (but to be fair, Tesla is probably not talking about ARTS either).
Shared Rides: May Mobility’s Approach
In an interview at ARTS, May Mobility’s Director of Product, Kurtis Hodge, provided an update on their automated shared-ride deployments so far. Their first driverless operations were launched in Sun City, Arizona, last December. Using safety drivers, they are active in several other cities, including Ann Arbor, Michigan, and Grand Rapids, Minnesota. He noted the variety their vehicles face, from urban operations in Ann Arbor to more of a small-town vibe in Grand Rapids.
An interesting validation of their use case comes with Grand Rapids, which has never had public transit services. As Tom Tang, May Mobility’s Chief People Officer, put it in an ARTS keynote address, “Low-density rides are harder for public transit systems. It’s also difficult to design a good route that meets the needs of many people. Bus costs are also high; when the vehicle is underutilized, it becomes an incredibly expensive proposition. All this, plus the challenges of labor shortages, especially for bus drivers, are worsening.” These types of challenges can result in no transit at all, but in this case, Grand Rapids went for May’s driverless offering.
May Mobility’s tech isn’t scared of snow either; they’ve been running during the winter months in all of their snow-weather deployment sites. They haven’t removed safety drivers yet in these areas; when it happens it will be quite the milestone.
May Mobility has shown the ability to successfully navigate the challenges of contracting with public sector transit agencies. My hat is off to them.
The company has more deployments planned and not just in the U.S. May Mobility has long had a deep partnership with Toyota Motor Company, and in previous news they’ve signaled that a deployment in Tokyo is in their plans. I was excited to hear from Mr. Hodge that this will occur within a year, which would be the first-ever ongoing commercial AV deployment in Japan.
Will cities resist or welcome driverless mobility?
At ARTS 2023, the vitriol was fierce from city officials and first responders; at least those from the host city of San Francisco. This was when news articles about “robotaxis doing dumb things” were almost a weekly occurrence.
The biggest news lately has been about Waymo cars honking too much in a San Francisco parking lot late at night. Not nearly as exciting.
This year, officials from other cities spoke of a much more orderly and relatively uneventful experience. In a panel discussion titled “City’s Experiences with Automated Driving Systems on their Streets,” representatives from the Arizona Commerce Authority, the City of Austin, the District of Columbia, and the Open Mobility Foundation examined the evolving situation since last year’s froth. My takeaway was that, while feelings are still raw among some in the overall community, deployment appears to be proceeding in a workable manner. Plus, there is an understanding that awkward, confusing, or risky behaviors can be diagnosed and fixed quickly by responsible companies. It was also clear that the interactions between the cities and AV deployers are more collaborative now.
Marc Scribner, Senior Transportation Policy Analyst at the Reason Foundation, attended ARTS and participated in many robotaxi discussions. From the business side, he saw another angle. “There are advantages to leveraging existing ride-hail networks for robotaxi services, so that introducing technology to the public occurs in lower-stakes environments where robotaxis can excel. This allows for incremental scaling, while ride-hail network expertise in pricing and marketing is also a major plus.” This has been underscored recently by Uber partnering with Cruise and Waymo, aiming to be the leading portal for the public to access robotaxis.
Is Smart Infrastructure A Thing?
It is fundamental that automated vehicles must fulfill their safety case based on on-board sensors, software, and actuators. In other words, achieving safety cannot rely on third-party communications or other supporting infrastructure, because the Automated Driving System (ADS) deployer has no direct control over support provided by a third party. This principle is baked into all of the deployments happening now, whether robotaxis or trucks.
But this doesn’t mean external data can’t be useful to enhance safety and operational efficiency. The Chinese government is all over this approach, planning a massive investment in supporting infrastructure for AVs over the next decade. Other governments appear to be in a wait-and-see mode.
But the private sector is also in the game. Companies deploying roadside sensing to collect and transmit relevant data to AVs were part of the discussion at ARTS.
Cavnue has so far been the leading player advocating for supporting roadside infrastructure. Since its inception a few years ago, the company has attended several ARTS events and was present this year. Just before the event, Cavnue announced the completion of its three-mile-long Automated Vehicle Corridor Pilot along I-94 in cooperation with the Michigan DOT.
I asked Cavnue CEO and Co-Founder Tyler Duvall for his views on how the conversation about roadside sensing is evolving.
“We’ve observed a notable shift in how ADS developers, cities, and states perceive the value of our I2V solutions,” he said. “While the initial focus was heavily on vehicle-based technologies, there’s now a growing recognition that infrastructure plays a critical role in enabling effective AV operations. Cavnue’s work on connected corridors is at the forefront of this shift, providing essential infrastructure that delivers real-time data and enhances the decision-making capabilities of autonomous vehicles. This evolution reflects an understanding that integrating vehicle and infrastructure data is key to building safer, more reliable, and scalable autonomous transportation systems.”
I was originally skeptical of the feasibility of roadside sensing to support AVs, but not because it wouldn’t be useful. The complexity of installing vast sensor arrays on public right of way gave me pause; this could be a hornet’s nest of regulatory barriers.
But in private conversations at ARTS, some new lines of thinking arose. While industry-wide action still has a long way to go, there is indeed an uptick in interest in external support across a variety of AV use cases. This will be a very interesting space to monitor going forward.
Will U.S. Federal Regulations Ever Catch Up?
The Autonomous Vehicle Industry Association is a leading voice advancing the safe and effective deployment of self-driving tech. Of its 19 members, four are robotaxi deployers and six are driverless trucking developers.
AVIA has no lack of issues to focus on. At ARTS, AVIA CEO Jeff Farrah emphasized the need for movement on a U.S. Federal Safety Framework for Automated Vehicles, which has been in discussion between USDOT and industry for several years. The AVIA comments on this topic were firm but polite.
As one of the speakers in a later session, I was more blunt. Despite diligent efforts by USDOT staffers, the lack of such a federal safety framework is an abject failure on the part of the U.S. government. The only reason the U.S. leads the world in AV deployment is that federal regulations are silent on driverless vehicles, so that by default the vehicles can operate on public roads. In parallel, 25 states have taken action to expressly allow these vehicles. For trucks, the only explicit holdout has been California, and after years of pressure from the industry, this is starting to evolve as well.
I, along with many others in the industry, attribute lack of progress to a lack of will on the part of USDOT at the political level.
The result is a clunky but workable regulatory environment for initial deployment. There was broad agreement among the AV developers attending ARTS that regulatory certainty provided by such a federal rule is needed for this industry to scale.
For trucking specifically, there are even more regulatory issues that I’ll discuss below.
Is Driverless Trucking Ready For Primetime?
I provided an automated trucking plenary presentation at ARTS which provided a broad overview of the breadth and depth of this very active space. In both long haul and short haul, scalable fully driverless operations are expected by year’s end from Aurora, Gatik, and Kodiak. ISEE and Outrider are already running driverless in logistics yards, with Forterra operations starting next year. Across a wide range of industrial operations, where driver shortages are endemic, truck platooning systems from Forterra, Hexagon, and Kratos are enabling a single driver to haul multiple loads.
From the presentation, my summary of automated truck deployment timing is shown below.
Timeline For Commercial Driverless Trucking Deployment
The Automated Freight and Trucking breakout session at ARTS this year provided an further depth on this immense amount of commercially oriented activity. ADS developers came together with trucking fleet executives, state DOTs, public safety stakeholders, and federal government officials. Talks from tech developers and regulators reflected the momentum of the industry while discussing no shortage of open issues.
Truck automation has been on the agenda since the start of the ARTS event twelve years ago. In the early years, ARTS was a great place for nervous state and federal officials to rub shoulders with the system developers, while also hearing from sometimes skeptical researchers. Over the years, despite some bumps in the road stemming from companies like Starsky Robotics and TuSimple, a significant level of trust now exists among the key stakeholders. Not that they agree, necessarily, but that each sees the others as straight shooters.
Coming back to achingly slow U.S. regulatory movement, in early 2023 industry players asked the Federal Motor Carrier Safety Administration (FMCSA) for a regulatory exemption to an antiquated law pertaining to a commercial truck which has pulled to the side of the road. When this happens, current regulations require that the driver jump out of the cab and place reflective warning triangles on the shoulder behind the vehicle. The AV industry request proposed a simpler approach: cab-mounted warning beacons, which automated trucks could activate when needed. Advocates assert that this approach can be adopted by human-driven trucks as well, so that the driver needn’t be exposed to danger on the roadway shoulder. FMCSA has been mum on their response to this request for the last 18 months.
Interestingly, at the breakout session, folks on the government side and those on the industry side were in the same room, but government officials couldn’t talk about the pending exemption request. Awkward, but they’re all used to it by now.
Notably, the breakout session underlined that the “competent and careful human” benchmark is gaining acceptance as a threshold for technical safety assurance and as an easy-to-explain target for an emotionally driven public. This is something to watch going forward. Will the ADS developers openly provide this type of information or just share it with their customers? If public, this will give federal and state officials some breathing room as they clear the way to fully allowing this technology on the open road.
Who determines that a vehicle is safe on public roads? As quoted from notes provided by the truck breakout organizers, the group was adamant that, “when it comes to policy and technology change in the U.S., we must remember that self-certification based on evidence is the foundational precedent for balancing care and progress,” while adding that “joint public and private determination of minimum performance is needed to keep ‘careful agents inside and careless agents outside’ the guardrails.”
Wrap Up
Late this year and into early 2025 is looking to be a breakout period for commercial driverless operations, building on the impressive level of deployment that Waymo has already achieved.
Even while the U.S. regulatory picture is cloudy, the startups have what they need to go for that cherished pot at the end of the rainbow, i.e. profitability.
Deployments overseas will likely create some interesting headlines next year as well.
As we see the maturing of this new tech capability, a new wave of tech enhancements may roll out. I’m keeping an eye on the infrastructure sensing topic in particular.
Meanwhile, enjoy the ride.
Note: Many of the presentations provided at ARTS 2024 are available for download on the TRB ARTS website.
Disclosure: Richard Bishop advises or holds equity positions for the following companies: Aurora, Forterra, Gatik, Plus, and Outrider.
