Trying out autonomous shuttle

Not too long ago, I made a trek to San Ramon to try out the PRESTO autonomous shuttle operating inside the Bishop Ranch business park.

The service runs 2 battery powered autonomous vehicles between the transit center and the shopping mall. For now, each vehicle has an attendant that monitors the vehicle (and takes over with a game like controller if necessary). The vehicle has a footprint of a minivan but can seat up to 11, along with ability to accommodate a wheelchair. The service primarily operates inside the parking lots and driveways of the business park, but crosses a public street (at Camino Ramon). The vehicle operates at level 3 automation and currently at a max speed of about 13 miles an hour.

The service is not particularly fast, and if you’re on a bike (or for me, electric unicycle), this service wouldn’t be attractive at all, but according to my conversation with one of the attendants, people do use the shuttle to go to the shopping center and can get busy. I like the general interior layout of the shuttle, with easy access between the door and each of the seats. It has a feel of a proper transit vehicle versus personal vehicles like sedans and SUVs. The vehicle amazingly has a tight turning radius.

With being autonomous, the technology is not perfect. On 1 of my trial rides, the vehicle made a few sudden stops. Fortunately the vehicle features seat belts. Those quick stops may be an advantage to avoid collisions but the machine generated response of sudden braking seems too much for humans to handle compared to human driven vehicles.

A question is whether this technology has a place in the transit network. An obvious concern is how this technology would impact the existing transit workforce. In my opinion, this technology is unlikely to replace mainline buses and bus drivers, which is a majority of service transit agencies provide. Besides operating the vehicle, transit bus drivers have to monitor and assist boarding of disabled riders and provide security presence onboard. Autonomous vehicles cannot fulfil those duties. AV technology however can improve safety performance of bus drivers.

I think AV should be positioned to provide new service that doesn’t exist today. It’s important to note that AV shuttle is not labor free. Workers are needed on the ground to monitor and assist boarding, along with enforcing safety rules. The difference is that, unlike a regular bus, where there is 1 worker per vehicle. 1 ground based assistant can serve more than 1 AV vehicle.

A hundred years ago, transit vehicles (primarily streetcars) had more than 1 worker per vehicle. 1 of the workers would operate throttle and brakes and the other worker would control the doors, monitor boarding, and collect fares (those old work rule still applies on Muni’s cable cars and F line, whenever they run other than PCCs or Milan cars). Transition to bus and newer streetcar eliminated the need to have a second worker, but for decades productivity stayed mostly the same.

For traditional transit, if there’s high demand, capacity can be increased on the rail side by lengthening trains, or switch to articulated or double deck buses on the bus side, rather than adding service, which may be desirable but costly due to extra workforce involved.

With AV technology, it is possible with a same level of workforce, more frequent service can be provided with smaller capacity vehicles. Such service can serve as dedicated first/last mile feeder to high capacity transit. For riders, it means no longer having to worry about coordinating schedules or missing connections like with traditional buses. Small vehicles can travel in areas where traditional buses may have difficulty to operate and can turnaround in a much smaller space.

There are many opportunities to position such service to connect high capacity transit to popular areas where the walk is too long, even if it’s within biking distance. For example, BART’s 12th St/Oakland station is about 3/4 mile from Jack London Square. AV shuttles could provide non-stop service between those two locations every 5 minutes or on-demand. This would be an equivalent to extending BART but without the price tag. Passengers don’t have to wait for a full size bus coming from Richmond or Berkeley that may have delayed somewhere along the line. While people on bikes or myself would skip the service, there are many others who couldn’t use micromobility: seniors, disabled, those traveling with young children, would benefit.

Other places that may benefit from frequent AV shuttles include neighborhoods and campus environments that are close to rail systems like Park Merced and Mission Bay in SF. Service can be made rail or BRT like with better station amenities.

I think the transit sector should consider what roles AVs can perform for what traditional transit cannot cost effectively do, and done in the way that would complement traditional transit. Remember there’s another AV sector that pushes AV technology not only to compete with, but also disrupt transit (that’s what essentially what Uber/Lyft have done for more than a decade), despite all the negative impacts those services could impose. Transit needs to compete by offering more & better service at the same cost, and make AV technology work with the public transit model of scheduled service, fixed route, and shared rides.

Automated trains and people mover have been part of the transit ecosystem for decades, providing more frequent service than otherwise possible, and have coexisted with human operated buses and trains. AVs can provide the similar advantage without the cost and environmental impacts of developing a fixed guideway system.