Why Autonomous Public Transit Could Solve Problems that Robotaxis Cannot

Jason Lomberg recently reached out to me to discuss the race to launch robotaxis. He interviewed me for PSDcast (a podcast) which was published here.

Below is more background about our position regarding the importance of autonomous buses as opposed to robotaxis when looking for ways to address mobility concerns in city centers. These are the questions posed by Jason, and my responses.

With robo-taxis’ autonomous capabilities (and even their ability to take flight), why won’t they solve the congestion issue?
Robotaxis are simply replacing one passenger vehicle (a privately-owned one) with another. They only carry 1-2 passengers, unlike buses that can carry up to 50 passengers and thus take away many cars. Furthermore, where will robotaxis stop when they are picking up or dropping off passengers? Most likely this activity will cause even more congestion.

And I understand Imagry is pushing autonomous public transit as a solution…so right off the right, what advantages would that have over robo-taxis?
If we can agree that there enough Uber, Lyft, and yellow cabs around to handle the desire of 1-2 passengers at a time to move around the city, it would seem that the only problem robotaxis are solving is how to politely let your driver know that you aren’t in the mood to chit-chat…something easily accomplished with an autonomous bus!

Autonomous buses (especially category M3, which are the ones we use) carry many more passengers than a robotaxi. In spite of the worldwide bus driver shortage—currently at 16% and rising by 1% each year—they can increase their hours of operation, the frequency of their arrivals, and the extent of their routes, all factors which will encourage drivers to leave their cars outside of city centers and therefore relieve inner-city congestion. Public transportation operators can’t meet today the high demand for their services, given urbanization and aging. Furthermore, public transportation is less expensive than taxi service, so autonomous buses promote mobility equity (allowing all socio-economic levels and age groups to enjoy an opportunity to move efficiently from place-to-place).

So how exactly would autonomous public transit help reduce congestion (especially with the low speed of buses and their frequent stops)?
Unlike the pods and shuttles that were used in the early days of autonomous driving, self-driving buses are able to maintain the posted city speed limits, integrating smoothly with varied traffic on public roads. Regarding “frequent stops”, typically bus stops are positioned alongside traffic lanes, so autonomous buses pull outside of the lane of travel to off/onboard passengers with little disruption to traffic flow. Note that there is a growing trend in the most congested city centers worldwide to impose a “congestion tax” on private vehicles (e.g., London, Singapore, Stockholm, New York, Milan, etc.) which also decreases the number of private vehicles overall, and which will eventually lead to a decrease in on-street parking, further opening up more space for autonomous mass transit. Across the EU, more cities are pushing hard towards becoming “green cities”, limiting passenger vehicles in city centers as much as possible.

One of the ongoing problems with autonomous vehicles, in general, is their difficulty in predicting the behavior of human drivers who share the road. How would autonomous public transit account for that?
Our technology is “bio-inspired”, meaning that we use the principles of brain science to develop a system that drives like a human, only better. With eight cameras positioned around the bus for “eyes” and artificial intelligence powered by advanced computers for the “brain” the vehicle perceives the environment in real time (360 degrees, up to 300 meters from the vehicle) and makes decisions about how to advance given the objects that have been identified. Our buses maintain a safe distance from objects on all sides, reducing the ability of the human driver to “surprise” us. In all the years we have been driving autonomously on public roads we haven’t incurred so much as a scratch!

I read that Imagry has been operating autonomous vehicles using their AI-based HD-mapless driving system on public roads since 2019. What’re some of the most important lessons the company has learned in that time?
We’ve learned to rely on Moore’s Law with respect to computing power, which is a fundamental dependency for an AI-based autonomous driving system. When we started developing our AI-based technology, the available computing power was suitable to operate a self-driving car, but insufficient to operate an autonomous bus. However, technology has made tremendous progress since 2019, and now there are at least a dozen manufacturers producing computing platforms that can address the current requirements, and beyond.

Another lesson learned is the importance of gaining public acceptance for the solution. Remember how the first Cruise robotaxis were subjected to vandalism in the neighborhoods they frequented? We’re convinced that involving the public in the decision to launch the new technology would have prevented that negative reaction. We are proud members of PAVE (both in the U.S. and Europe), a not-for-profit organization dedicated to educating the public about autonomous driving. PAVE’s efforts on the industry’s behalf reduce the fear of the unknown and encourage communities to support the advancement of self-driving vehicles as a means to improve their quality of life.