Remote Operators for Self-Driving Cars
Self-driving cars with remote human operators could be tested in California starting from April, VentureBeat reports. It’s not certain. The new regulations need to be approved by California’s legal compliance agency. Then the DMV will open a 30-day public notice starting from March 1.
A few companies, such as Nissan, Waymo, Zoox, Phantom Auto and Starsky Robotics, have been working on remote control technology. It is an interesting strategy that allows using autonomous vehicles in more use cases sooner. Vehicles can drive autonomously in simpler situations, and a human takes control when the autopilot can’t operate safely.
It seems to me, the largest obstacle to get this working will be qualities of communication channels.
Do we have enough mobile network coverage to have 100% connectivity for major use cases like urban driving and highway driving?
What bandwidth will we need to give the human operator enough information to make decisions? At least, the operator will need a high definition video stream from a few cameras. Data from radars, sonars, lidars, and telemetry from other vehicle sensors won’t hurt either. Perhaps, some of this can be pre-processed an compressed on the vehicle, but even after compression, it will, probably, be over 1 megabyte per second.
Latency is important as well. Currently, U.S. mobile networks have average latencies (measured as time to ping google.com) in the range about 80–170 milliseconds. What would the impact be if we add this time to the average driver reaction time?
Security is a concern. The connection between the vehicle and the human operator will need to be secure. Access control will be important too.
From the positive side, remote operators can get more information, and information presented in a better way, comparing to human drivers inside the car. It’s easier to read lidar and radar data and overlay it on the image from the camera. It’s easier to highlight important things like dangerous objects on the road. Ergonomics of the driver’s workplace will be very important.
As one operator can work with multiple vehicles, the companies will be able to use more vehicles for testing and learn faster.
Read next:
