Autonomous Mobility: Billy Riggs on Perception, Cities, and Advice for LA28

Billy Riggs, a national voice on autonomous mobility and Professor (University of San Francisco), explains why perception autonomy—the vehicle’s ability to truly see and interpret the world—is the real engine behind AV progress. Drawing upon his role leading the Autonomous Vehicles and the City Initiative at USF, as well as hands-on research with Level 4 deployments, Riggs argues that autonomy won’t solve congestion without major shifts in curb policy, workforce planning, and data governance. He challenges cities to stop chasing sci-fi and focus on “Monday-morning” solutions like microtransit, dynamic curb management, and realistic public–private models. Looking ahead to LA28, Riggs warns that LAX and demand-responsive mobility expectations pose the Games’ biggest transportation risks unless the region embraces fast, flexible deployments now.

“The vehicle’s ability to sense its environment and translate that into safe, predictive action is what makes autonomy possible. That’s the real breakthrough: machines performing a task that humans have always done.” — Billy Riggs

Billy, introduce yourself to our readers. How did your career evolve, and what academic and professional credentials ground the work you do today?

I'm Billy Riggs, a professor at the University of San Francisco, where I direct the Autonomous Vehicles and the City Initiative. My career sits at the intersection of transportation planning, engineering, technology, and urban development, and I’ve spent more than 25 years working across those domains. I’m trained as both a planner and an engineer from UC Berkeley, though I originally hail from Louisville, Kentucky.

I began my career with the U.S. Coast Guard doing pier-side planning—everything from housing development to land-use work—before moving into private-sector consulting. After completing my PhD at Berkeley, I fell in love with transportation innovation just as companies like Lyft and Uber were emerging.

Since then, I’ve increasingly worked in venture and advisory roles with automakers, cities, and AV companies. Today, I split my time between research, fieldwork, and policy advising, working with the European Commission and governments around the world, alongside teaching and collaborating with students and folks like you.

Address your professional perspective on the evolution of autonomous mobility, as autonomy, EVs, and now AI accelerate on parallel tracks.

I didn’t start my career working on autonomy. I originally was interested in technology and cities, but the “tech” that excited me early on was the bicycle. I wanted to create streets that were safer for cycling, and the tools I used were basic planning and analysis tools like GIS, street design models—nothing like what we talk about today. Around 2011–2012, AI-based tools and perception algorithms started coming of age, and that grabbed my attention. These systems had the potential to reduce collisions by addressing the horrible driving decisions humans routinely make. We’re all piloting 4,000–6,000-pound machines while making irrational, overconfident choices...technology offered a chance to correct for that.

I think my early work was pretty bullish on how electrification and autonomy could reshape cities—reduce collisions, improve environmental outcomes, even cut congestion. And paired with good land-use policy, I still think they can contribute meaningfully.

Over time, though, I’ve grown more skeptical about autonomy’s ability to reduce congestion. Vehicle miles traveled keeps going up. A lot of that isn’t “induced demand” so much as latent demand—people who have long been under-resourced in their transportation options. Many of those people are lower-income residents or people of color who’ve been marginalized by historic transportation decisions, including decisions embedded in legacy transit systems.

At the end of the day, what I’ve come to believe is that autonomy and EVs only work when they augment the human systems we’ve already built. Not when we expect them to replace those systems outright. The goal is to use these tools to make our cities function better.

And the term I used earlier—perception—refers to the foundational layer of autonomous driving: the ability of these vehicles to perceive their environment. That’s the game-changer. Perception systems use lidar, cameras, radar, and other sensors to create a real-time, high-definition understanding of a vehicle’s surroundings. That data informs the AI’s decisions and ability to do prediction and path planning. In many ways, perception is everything; without it, the vehicle can’t act safely or intelligently.

In the context of automation, what is perception? Elaborate and contextualize its applications to advanced mobility.

In the context of autonomy, perception is really the foundation of how autonomous vehicles work. The first job of any AV, whether it's a Waymo, Zoox, or Pony.ai vehicle, is to perceive its environment.

Here’s how that works: these cars are covered in sensors that spin and continuously map the environment in real time and in extraordinary detail—cameras and radar. The system identifies every object around the vehicle and compares that live map to a highly detailed pre-existing map stored in what you could call the vehicle’s “brain.”

The perception algorithms on board use all of that incoming data to make dynamic decisions, second by second. That’s where sensor fusion comes in—combining inputs from lidar, optical cameras, radar, and ultrasonic sensors into a single, coherent understanding of the roadway.

Sometimes I tell my students that humans drive with essentially one sense—vision. Meanwhile, these vehicles ingest a whole symphony of data streams: temperature, roadway traction, braking patterns, whether there’s moisture on the windshield, and even micro-changes in vehicle stability. Every bit of that is fed into the car's multiple GPUs, where perception data becomes everything needed to make automation work. 

The vehicle’s ability to sense its environment and translate that into safe, predictive action is what makes autonomy possible. For me, that’s the real breakthrough: machines performing a task that humans have always done, but doing it with far more data and potentially far greater consistency.

You direct the Autonomous Vehicles and the City Initiative at USF. Apply what you’re studying to making cities smarter—and, how’s that work going?

Well, the Autonomous Vehicles in the City Initiative really came together when I left Cal Poly San Luis Obispo, where I’d been a city planning professor for six years. I loved that work, but I wanted to build a center focused specifically on how automation intersects with real urban environments. The best home for that turned out to be the University of San Francisco, right in the heart of the city, with a founding grant from Cruise Automation.

Cruise was incredibly generous. Their CEO, Kyle Vogt, and their policy team, led by Rob Grant, let us study their early Level 4 systems and actually put our students into those vehicles. That gave us a unique chance to understand, from a planning and policy perspective, what happens when autonomous vehicles start interacting with the built environment—not in theory, but on real streets, with real constraints.

What did those early deployments reveal about how people actually travel when automation is introduced—not hypothetically, but in practice?

As we began studying travel behavior, interactions with other vehicles, and early policy friction points, it became clear that most writing on autonomy was still theoretical. We had the rare opportunity to work with Cruise and Waymo as these systems were operating at true Level 4. That allowed us to do curb-management research, transit-integration research, and even observe Waymo testing transit incentives for pickups and drop-offs near stations. There were some genuinely surprising findings about how you can nudge riders into first/last-mile behaviors that support rather than cannibalize transit.

But another part of the AV & the City Initiative is about governance itself. We created it because cities desperately need “third spaces,” or some kind of neutral ground where public agencies, private companies, and researchers can actually talk to each other about the future of mobility. Our civic systems are so politicized that there are almost no places left where those conversations can happen productively.

Is a robust public transit system “realistic" post the full adoption of autonomous mobility?  

Right now, the financing picture for public transit isn’t healthy. And I do believe the future of transit will integrate some level of automation, and likely some form of public–private partnership. That doesn’t mean abandoning public transit’s role; it means reinventing it. The system we build for the next 50 years won’t look like the system from the past 50, and we have to create a safe space to acknowledge that.

That’s why we host our annual conference and bring industry, government, and academia together. Not to speculate in the abstract, but to test ideas in the field, evaluate what works, and start reshaping policy frameworks based on real evidence rather than wishful thinking.

Bay Area voters will be asked to approve a $400 million mobility and street-safety bond to rebuild aging transit corridors. From your perspective, is a regional tax-and-bond measure the most effective way to finance meaningful mobility improvements, or are there better funding tools the Bay Area should be considering?

Transportation finance needs a refresh. Across California and the country, we keep putting forward the same kinds of local tax measures—property taxes, sales taxes—levied on the same local user base. They can be sustainable in theory, but they’re incredibly vulnerable to economic cycles. Sales-tax-based transit funding collapses in a recession. 

When downtown San Francisco emptied out, sales tax dropped, and so did the revenue that underpins the entire transit finance system.

In San Francisco’s proposed regional measure, you also see equity issues. Most of the revenue would come from just two counties, San Francisco and San Mateo, because they generate the bulk of the region’s sales tax. Whether that’s “fair” is up to those residents, but functionally, you end up with two counties subsidizing regional transit for everyone.

Returning to the consulting you do for cities, share about feasibility and deployment.

Well, cities need solutions that aren’t science fiction. They need things that can work next Monday morning, not 10 years from now. They need tools that procurement officers, operators, and mayors can actually deploy within an election cycle. That’s why our lab recently brought on an EU expert, Dr. Henriette Cornet, who has spent her career scaling multimillion-euro autonomous deployments across European cities. A lot of those systems are working today, and there’s a lot we can learn from them.

Cities need help not just deploying autonomous pilots, but managing the basic infrastructure around them—data-sharing frameworks, data-management strategies, and, increasingly, curb-space allocation. Because when you look at consumer behavior in the rideshare era, the demand isn’t for traditional transit. It’s for on-demand, demand-responsive mobility.

The rideshare revolution shifted consumer expectations. People aren’t demanding the same old transit service—they want demand-responsive mobility. And when you start talking about the future of mobility through that lens, public trust becomes a major factor. Cities can either amplify suspicion about new technology or help reinforce a safety-first, innovation-forward narrative. We’ve seen both. Some cities have been excellent partners to private-sector innovators; others have successfully stirred up public disenfranchisement around automation.

If that’s the consumer shift, what does “demand-responsive transit” look like in a city that’s still structurally built for 9-to-5 commuting and fixed-route buses?

You know, we need cities to be able to deploy not just autonomous solutions, but solutions that are real to their citizens—things they can put on the ground immediately. They need data-sharing strategies, data-management strategies, and ideas for how they’re going to allocate and rethink their curb space for pickup and drop-off.

And I highlight curb space because when you look at the rideshare revolution and what consumers are demanding right now, it’s not the same transit demands we saw ten or twenty years ago. People want demand-responsive transit. That’s the expectation.

That’s why I always push for a glass-half-full mentality. We don’t want to stall innovation—we want politically positive solutions. And that includes difficult, even radical conversations. 

For instance, can we create a narrative that makes it acceptable to free up on-street parking for pickup and drop-off? Politically, no one wants to walk into a merchant district and say, “We’re taking away your short-term parking.” No politician wants that fight. But in some places, reallocating curb space is the best thing you can do to improve turnover and mobility. We need to have those discussions, and this is the kind of work we do when we advise cities and companies, talking about governance, data sharing, curb space, all of that…and then we can’t forget the future of labor.

We spend a lot of time helping cities think through what the future of “the driver” looks like. It doesn’t have to mean job elimination. It can mean job retention, but in a different form. In fact, when you look at charging, fleet management, operations, depot management—in the AI and automation future, there’s a tremendous amount of blue-collar employment opportunities. 

Now, the white-collar side is a little different, and I don’t entirely know how we take that on. A lot of those roles are not protected labor positions in the same way. But we are trying to elevate that dialogue, because cities really do need to start thinking about these transitions now. And those are the conversations we’re beginning to have with cities and companies as we think about the future of AI and autonomy.

Relatedly, drawing on your private-sector experience, speak to the general market. Who is actually embracing innovation right now?

What’s interesting is that innovation isn’t being embraced uniformly—it’s being embraced selectively. Right now, the most aggressive adopters tend to fall into three camps. First, it’s logistics and freight operators. They’re under relentless pressure on margins, labor availability, and reliability, and that’s made them very pragmatic about automation, electrification, and AI. Warehousing, ports, and middle-mile freight are where some of the most meaningful gains are actually happening. The southwestern US has really led in this area, and Texas has really set the standard in logistics and freight. 

Second, we are seeing real momentum among small and mid-sized cities and agencies that are capacity rather than politically constrained in the US and Europe; places like Jacksonville, Florida, and Oslo, Norway. Mid-sized cities—often outside the coastal spotlight—are more willing to test demand-responsive transit, automated shuttles, or dynamic curb pricing because they don’t have the luxury of maintaining inefficient legacy systems.

Finally, it’s private sector companies that already operate fleets that are really moving in the innovation space right now—rideshare, delivery, transit operators, utilities. Innovation is much easier when you have aligned vehicles, depots, and data, and then maximize revenue and route efficiency per passenger mile/kilometer. This also helps with vehicle cost and production. The fantasy that private consumers would be the early adopters of autonomy was always misplaced, given the high cost of the vehicles. Fleet operators are where the economics actually work.

What’s also notable is the inverse—who isn’t moving fast: organizations that are structurally rewarded for maintaining the status quo. And in many cases, those are our most entrenched and slowest-moving organizations—large cities. Innovation today is sometimes less about enthusiasm for the tech and more about willpower and inertia.

And generally speaking, have the investments of public transit agencies in new technologies delivered the results they hoped for?

The honest answer is: sometimes, but far less often than promised. And it’s rarely because the technology itself failed. Many transit agencies invested heavily in new tools—real-time arrival systems, fare technology, electrification, even AI-based planning—but layered them on top of base tech that never really changed. The technology is an accessory rather than a catalyst. It’s not foundational. 

Where investments have delivered results is when agencies paired technology with operational reform. Things like bus-only lanes with enforcement, true signal priority, redesigned routes, or demand-responsive pilots that replaced low-ridership fixed routes. In those cases, tech amplified good policy.

Where it hasn’t worked is when agencies expected technology to solve structural problems—declining ridership, labor shortages, underfunded operations—on its own. No app fixes a bus that’s stuck in traffic.  I’d also add that agencies are often judged too harshly for not delivering “transformational” outcomes, when in reality the biggest wins are incremental: better reliability, safer operations, more flexible service. Those gains matter, but they don’t photograph well for press releases. I think the less isn’t that transit agencies shouldn’t invest in technology—it’s that technology only works when institutions are willing to change how they operate alongside it.

Because of time and space, let’s Zoom in: what advice would you give to the planners and staff working on LA28? 

Well, it’s going to be a super high-visibility event, and I think LA needs to zero in on making sure transportation is safe, sustainable, clean, and well-orchestrated. The real risk here is over-promising. I have two big fears. One is LAX. The other is the broader consumer preference for demand-responsive transit.

If you’ve been to LAX, you know what I mean. Trying to get an Uber or Lyft or any type of demand-responsive service is a mess. And I don’t think just throwing more buses at the airport is going to fix that. There are answers, but right now I’m not sure we’re thinking creatively enough about how you move massive numbers of people from the airport—and potentially between venues—using demand-responsive approaches.

And tied to that is curb space. That may mean we need a radical curb strategy for the Games. Deliveries, TNCs, micromobility…none of that goes away during LA28, and no amount of AVs or EVs is going to solve curbside congestion on its own. There are models out there. Boston and Seattle have done a lot around dynamic curb strategies, even ballet-style curb management. And we can look at microtransit shuttle solutions that have already been deployed. Paris, for example, used autonomous micro-shuttles at Roland-Garros during the French Open. We could imagine something similar for LA28.

The catch is that the French Open shuttles were Chinese-built, and I don’t know if a U.S. platform exists yet that could fill that role at scale. But the point is: there are radical, quickly deployable solutions out there that don’t require us to lay down rail between now and 2028. It could look more like a bus rapid transit or microtransit approach—something nimble, flexible, and fast to stand up.