By Mike Schmidt, AEM Industry Advisor Editor

The world eagerly awaits a future where autonomous vehicles are the norm, and the idea of transit is defined by the freedom for humans to travel from one place to the next at the speed in which they please.

But in meantime, an important question regarding self-driving vehicles requires a definitive answer: How does autonomous technology evolve and develop into a usable solution that places the human – as opposed to the vehicle itself – at the center of transportation?

“Self-driving vehicles are actually solving their own problem,” said Steve Vozar, chief technology officer for May Mobility, a Michigan-based self-driving fleet-as-a-service company. “If you set out to make a self-driving car, what you're going to get is a car that is going to drive itself. You’re not going to get a transportation solution that works for people.”

Vozar outlined his vision for an autonomous future and shared some key insights on the value proposition of self-driving technology as a viable microtransit solution at AEM’s most recent Thinking Forward event, held earlier this month at the Henry Ford Museum in Dearborn, Michigan

And those in attendance came away from Vozar’s presentation with a key takeaway regarding the potential of autonomous vehicles: No single technology is poised to serve as the tipping point that accelerates the self-driving revolution. A number of factors – the vehicle itself, software, infrastructure, customer demand, community development, among others – all will play integral roles in determining when autonomous vehicle technology ultimately takes off.

Sign Up for AEM's next Thinking Forward event in a city near you. 

Use Cases for Self-Driving Vehicles Today

According to Vozar, the reality of self-driving technology has yet to catch up to the concept of level-five autonomy. However, he was quick to suggest there is value in putting self-driving vehicles on the road today, and it can be found in embracing self-driving microtransit in urban areas.

“The value proposition for self-driving microtransit in urban areas is this,” he said. “It’s larger numbers of smaller vehicles we can put on the road for the same cost, and that reduces both latency and wait time.”

The May Mobility CTO said his company has targeted low-speed, urban environments where people live and work to implement self-driving microtransit fleets for transporting them from one place to another. However, he noted, other major companies –  namely General Motors and Tesla – look at autonomous vehicle technology a little bit differently.

“They’re saying put passenger-driven cars on the road that are autonomous some of the time,” said Vozar. “We’re saying we want to have non-passenger-driven cars that are autonomous all of the time – but only in some places. And when it comes to speed, we shouldn’t be going 45-miles-per-hour. The sensor technology, it’s just not there yet. And we want to get started today.”

Vozar said he looks forward to a future in which the number of high-speed vehicles found in urban areas is drastically reduced, commuters are afforded access to a mixture of transportation modalities, and speed doesn't matter. Why? Because in this world, traffic flow would be optimized much more effectively than it is today.

Urban Self-Driving Microtransit

So what does self-driving microtransit in urban areas look like today? Well, May Mobility recently launched the first deployment in the company's history in downtown Detroit. May Mobility replaced a 30-person diesel bus that arrived at stops on its route every 20 minutes with a trio of self-driving vehicles capable of carrying five passengers at a time. And while the maximum number of passengers able to be transported per hour was reduced, demand was such that a bus with the capacity to hold 30 people wasn’t necessary.

“The throughput is the same, but passengers aren’t waiting,” explained Vozar. “As soon as they get to a stop and the shuttle is filled, they can move on. They're seeing reduced wait times, which makes them happier riders as a result. And to us, that’s the No. 1 thing.”

Check out AEM's Thinking Forward podcast to learn more about how disruptive tech trends are impacting the industry. 

Based on the results of the Detroit deployment, May Mobility is moving forward with others in the future. One in Providence, Rhode Island consists of a 12-vehicle deployment working a 5-mile route, with six vehicles on the road at any given time. Another, in Grand Rapids, Michigan, aims to augment existing public transportation by extending both its hours and capabilities.

Lessons Learned

A number of valuable lessons regarding autonomous vehicle technology and its usefulness as an urban microtransit solution have been learned as a result of launching the deployment in Detroit.

Chief among the most notable takeaways, said Vozar, is data is overrated.

“People talk about collecting terabytes of radar data, lidar data and camera data,” he explained. “We use it, log it and keep it all. And we’re interested in it, because it helps us be better. It helps us tune our algorithms, and we do regression testing. And that’s all well and good.”

But, he continued, the type of data May Mobility really values most is rider feedback.

“We want to understand what passengers want, and what they think about the service,” Vozar added. “And all other data we collect is in service of that.”

 

 “If you set out to make a self-driving car, what you are going to get is a car that is going to drive itself. You’re not going to get a transportation solution that works for people,” said Steve Vozar, chief technology officer, May Mobility. 

 

Another lesson Vozar and his colleagues at May Mobility have learned from their time spent developing a self-driving fleet-as-a-service company? It's going to take more than one seminal moment or cutting-edge technology to harness the immense potential of autonomous vehicles.

“We think there will be four or five ‘Eureka!’ moments that will unlock the autonomy for everyone,” he said.

Above all else, Vozar noted, the key to a future in which self-driving vehicles are the norm is a concerted effort on the part of key stakeholders – both in the public and private sector – to develop vehicles, infrastructure and operational domain sooner rather than later. Because simply creating technology for technology’s sake isn’t a particularly useful – or lucrative – endeavor.

The Tech Behind Self-Driving Vehicles

The key to May Mobility’s self-driving technology is its emphasis on the concept of behavior behind the wheel. Behavior refers to what occurs between long-term planning – how to get from one place to another – and the actual brake-throttle-steer connection with the vehicle.

“It’s what I’m going to do in the next 10, 15, 30 seconds to advance my progress toward a goal,” said Vozar. “That’s informed by the path planning, and it informs control.”

While sensor technology in the form of lidar, radar and onboard cameras tasked with providing a 360-degree of what’s occurring all around a self-driving vehicle are crucial to informing behavior, Vozar said it’s now possible for a machine to actually simulate a driving environment in real time and employ what he refers to as “simulated hindsight” to affect the vehicle’s behavior. It’s more or less an online and onboard simulation, and it helps provide better reaction on the part of the machine. More importantly, however, is it means more efficient and effective transportation.

Advancements in self-driving technology also allow for onboard sensing to be augmented with offboard sensing. Whether it comes in the form of looking at traffic lights or seeing around corners, the redundant communication between onboard and offboard sensors provides self-driving vehicles with the capability to make more informed – and, therefore, better – decisions than human operators.

“Sensors go into behavior perception,” said Vozar. “High-level control moves are evaluated against ride quality, against safety, against progress toward a goal and out comes – with the benefit of simulated hindsight – what should be done.”

According to the May Mobility CTO, election cycles occur twice per second, are evaluated by the vehicle, and then brake-throttle-steer actions take place.

“What it allows for is more and smoother human-like behavior,” he said. “And when you put all of it together, you get emergent driving behaviors in very complicated environments. Always on alert, always looking at things.”

The Future of Autonomous Vehicles

The future is merging with the present. The technology exists to make autonomous vehicles useful in certain scenarios – and it’s being employed today. However, hurdles to growth and widespread adoption exist. Questions related to the amount of human intervention necessary to make self-driving vehicles operate safely and properly, as well as questions related the insurability of companies that deploy self-driving vehicles in certain environments, have not yet been definitely answered.

But the two most significant – and challenging questions – deal with society’s lingering uncertainty regarding level-five autonomy and how it affects the ability of companies like May Mobility to scale their businesses.

“People aren’t ready for completely unmanned vehicles, and we do see a role for an ambassador in the vehicle right now,” admitted Vozar. “And by far the biggest challenge for companies like ours moving forward is scale – more cars in more places, scale of manufacturing, how many cities, how many people. Because every new location brings unique challenges with it.”

The future remains uncertain. Level-five autonomy is nothing more than a concept today. What seems likely, at least from Vozar’s standpoint, is the world will see a steady increase of urban, low-speed use cases over time. Eventually, autopilot and supercruiser vehicles will become more capable. Then – someday, in the not-too-distant future – humans will rarely have to actively operate a vehicle’s controls.

“That, though – is maybe 15 or 20 years away” added Vozar.

AEM members learned about this and other topics at a Thinking Forward event in Dearborn, Michigan on May 14 Visit aem.org/think to learn about more of these upcoming events in a city near you, including one at the Cisco Innovation Centre in Toronto on Sept. 10. 

Stay plugged in to AEM’s Thinking Forward initiative for events, articles, podcasts and insights to keep your business on the cutting edge.
• Sign up to get the AEM Industry Advisor in your email.
• Subscribe to the AEM Thinking Forward podcast on your favorite podcast app.

 

×