According to foreign media reports, self-driving cars are not yet ready for large-scale commercial deployment, but as technology advances, most companies hope to develop multi-layered safety technology for self-driving cars. That includes Amazon-backed self-driving startup Aurora, which has developed a remote monitoring system akin to an air traffic controller that could help increase public acceptance of robo-taxi.
Headquartered in Palo Alto and co-founded by three veterans of self-driving technology from Google, Tesla and Uber, Aurora is developing a system called “teleassist” to improve software, computers, sensors and vision systems, allowing the company’s test fleet in the San Francisco Bay Area and Pittsburgh to handle a variety of road conditions. The technology is designed to allow trained technicians in remote facilities to access the vehicle’s sensors when needed and provide advice and guidance for unusual situations, such as pulling the vehicle over safely.
While companies like Phantom Auto and self-driving truck startup Starsky Robotics are testing remote control systems, their teleassist systems work differently, said Sterling Anderson, Aurora co-founder and chief product officer.
“Our system alerts remote assistance personnel when the need arises, either from the driver of the vehicle or another user of the vehicle in the logistics network, or because the vehicle itself has issued a distress signal,” he said. , as in “I see something that is not in my comfort zone, tell me what you think I should do. “Even so, the recommendation for remote assistance cannot be part of the functional safety of the car.”
Sterling Anderson declined to say how much Aurora has invested in the monitoring system, or details such as when the teleassist equipment will be turned on and when it will be operational.
Led by Chris Urmson, the former head of Google’s self-driving car project, Aurora has raised about $700 million so far since it came to the public in 2018. In February 2019, Amazon participated in Aurora’s Series B financing, which raised more than $530 million (about 3.716 billion yuan). In addition to self-driving passenger cars, Aurora has already begun testing self-driving delivery vans and self-driving trucks, though Anderson declined to discuss whether there is a self-driving partnership with Amazon in such projects.
Teleoperation has long been used to operate space probes and drones, not fully robotic cars. Waymo, a commercial self-driving car company founded by Google in 2016, also employs human monitors at its facilities outside Phoenix to monitor and assist its self-driving minivans, though such human monitors cannot be controlled remotely vehicle. While Tesla CEO Elon Musk has said Tesla will be an early leader in self-driving technology, using its Autopilot system to collect vast amounts of road data from drivers, the company It also doesn’t appear that remote instructors are being used. Anderson, who helped design Tesla’s Autopilot system, was also reluctant to talk about his former employer’s approach.
Anderson said the delays in cellular network signals and the difficulty of “situational awareness” for remote personnel to understand all the important situations the vehicle is facing create too many challenges to safely take over driving control. However, as in-vehicle software continues to improve, methods such as teleassist may help self-driving fleets get operational faster. (All pictures in the text are from Aurora)