"Humans make more mistakes than machines," said Krister Kalda, head of industry cooperation at Tallinn University of Technology (TalTech). "If the bus is made smart enough, it's safer."

He noted that many car crashes are caused by drivers checking their phones, for example, while passengers in autonomous buses could even work on laptops safely.

Self-driving buses in Estonia have remained limited to pilot projects and short trial routes. For his recently defended doctoral thesis, "Integration of Autonomous Last-Mile Minibuses into Urban Space," Kalda's research examined human-bus interaction, a Rae Municipality mobility pilot just outside Tallinn, and scenario awareness tests in Kadriorg Park.

In Rae Municipality, people's receptiveness to the buses doubled. "Before the trial, 43 percent said they'd use autonomous buses daily; after, that figure rose to 86 percent," he said.

Demographic differences in trust

In one study, Kalda's team studied how passengers respond to driverless vehicles and how they communicate with them. Their results painted different pictures along various demographic lines.

"Estonians don't even want anyone in the vehicle; they trust it and would prefer no one supervising," he highlighted. Older folks and those with a Russian background, meanwhile, definitely want an attendant on board to provide a greater sense of security.

When it came to human-bus interaction, passengers had no trouble understanding red Xs and green circles displayed on the windshield, but stumbled when it came to the white stripes displayed to represent waiting at a crosswalk.

"It turned out they don't always even look at the bus, so some kind of audio or other form of communication is needed," Kalda said.

Kindergarten bus of the future?

Just outside the capital, a Mobility as a Service (MaaS) pilot in Rae Municipality let people summon a driverless bus by app or online to reach local stops, where they could catch county buses.

Instead, many used it for local errands, including trips to the doctor's office, pharmacy or store.

There were even regular bookings for kindergarten dropoffs. "The kids enjoyed the ride, and the parents could get back home and still drive into the city afterward," Kalda said.

His team surveyed 1,000 municipal residents before and after the pilot, and confirmed that support for the autonomous buses doubled, surpassing even Kalda's predicted 30 percent growth.

"People's trust grew as they saw just how safe [these buses] can be," he said. Even local officials began seeing autonomous buses as a potential way to make getting around the area easier.

"The unequivocal 'no' dropped from 17.1 percent to 6.7 percent," he added.

Challenges and safety testing

In Tallinn, Kadriorg Park trials nonetheless highlighted situational awareness issues. Falling leaves, for example, triggered sudden breaking as the bus detected obstacles in its path.

"Cyclists even started trying to sabotage the autonomous bus," Kalda added. The team's research documented these challenges and more to improve ride quality.

Other tests looked at last-mile travel, such as passengers with luggage taking the bus to a tramp stop and then the tram to the port before riding a driverless bus up to their terminal. They also studied how factors like weather or passenger load affect travel time.

"I even served as a safety operator on some days in all of these pilots, and overall things went smoothly," Kalda said, although he admitted to a near miss one day in Kadriorg.

"The bus turned from a park path onto a bigger road," he recalled. A car on the road was moving too fast, nearly causing a collision with the bus along a sharp curve. Even equipped with a 360-degree lidar, he added, this incident showed the need for speed regulation.

These risks and more all need to be addressed before the more widespread adoption of driverless buses.

At their current safety levels, Kalda said, autonomous buses are best suited for shorter, supervised routes or enclosed areas, "like from airport gate to plane or from the bus stop to the ticket entrance at the zoo."

Next steps for adoption

To date, three major groups have piloted self-driving buses in Estonia, one of which has ceased operations. The second, Auve Tech, is a TalTech offshoot, while the third consists of the school's own researchers, Kalda among them.

The country's first autonomous bus debuted in September 2018, to coincide with the university's centennial; it was also Europe's third and the world's fourth driverless bus.

Globally, manufacturers like Toyota are testing buses, which were even used to transport athletes at the 2020 Summer Olympics in Tokyo, but large-scale public deployment has been cautious as reputations remain on the line.

In Estonia, Kalda recommends longer and broader trials, including in winter conditions, standardized testing that accounts for cultural differences, and more sensors to detect hazards, obstacles or fallen passengers.

"If you only do testing in very safe, controlled conditions, you won't uncover the weaknesses that need addressing," he noted. Cybersecurity.would also need to be addressed.

On-demand driverless buses could complement personal vehicles, reduce pressure on city parking, integrate with other forms of micromobility, and provide independent travel for blind and visually impaired passengers — all while reducing emissions.

"There's an economic upside too," Kalda continued. "In areas that need a driver and a regular scheduled route but see little traffic, a self-driving bus could cover three to four kilometers efficiently."

Krister Kalda defended his dissertation, "Integration of Autonomous Last-Mile Minibuses into Urban Space," at TalTech on January 27.

--

Follow ERR News on Facebook and X and never miss an update!