FYI, the video froze at e911. I rebooted my phone and restarted the video which then ran smoothly.
Another takeaway is that remote technical support is inadequate to assure safe robotaxi operation. The communication latency, the limited situational awareness provided by the vehicle sensor suite, the time it takes for the remote human monitor to adequately apprise the situation, (retrospectively assessed) inadequate training for their emergency response, and natural reluctance of a low-level employee to declare an emergency that should be handled by higher level management all contributed to the pedestrian’s injuries. All of those factors are inherent in remote vehicle monitoring and make safety claims hard to believe in any robotic vehicle fleet that does not have humans in the driver’s seat.
If remote support is inadequate (and it always will be) to assure safe robotaxi operation, then to assure safe operation all safety-critical automated logic, sensors, and hardware must be validated as safe in a vast set of circumstances before the vehicles are deployed and every time any of those components is changed. Unfortunately, in the analog world of matter and humans, everything changes continuously. Tough engineering problem.
Thanks Phil.
FYI, the video froze at e911. I rebooted my phone and restarted the video which then ran smoothly.
Another takeaway is that remote technical support is inadequate to assure safe robotaxi operation. The communication latency, the limited situational awareness provided by the vehicle sensor suite, the time it takes for the remote human monitor to adequately apprise the situation, (retrospectively assessed) inadequate training for their emergency response, and natural reluctance of a low-level employee to declare an emergency that should be handled by higher level management all contributed to the pedestrian’s injuries. All of those factors are inherent in remote vehicle monitoring and make safety claims hard to believe in any robotic vehicle fleet that does not have humans in the driver’s seat.
If remote support is inadequate (and it always will be) to assure safe robotaxi operation, then to assure safe operation all safety-critical automated logic, sensors, and hardware must be validated as safe in a vast set of circumstances before the vehicles are deployed and every time any of those components is changed. Unfortunately, in the analog world of matter and humans, everything changes continuously. Tough engineering problem.