Source: https://www.autopilotreview.com/tesla-autopilot-accidents-causes/ The NTSB investigated the collision as they would a plane crash, looking into the nuances of the accident to determine the cause. Because planes are years ahead of cars in terms of autonomy, the autonomous car market can learn a lot from aviation. Pilots and crew members have been responsible for not engaging hands-on operation because they were distracted or overly confident in the autopilot features. Because of this, many people advocate to skip “level 3” autonomy, where a driver takes over when the car’s autonomy can’t handle a situation, for “level 4” autonomy, which is equipped to handle situations even when a driver ignores the recommendation to take over. I follow aviation accidents and this is a real problem. More than a few cases have the cockpit crew discussing something confusing them when if they'd turned off the autopilot, they would be able to deal with the problem. Typically a sensor fails and some of the alternate sensors are reporting what doesn't make sense to the crew. It was a bad sensor. Bob Wilson
Autopilot will cover most of driving conditions, but not all of them, especially for the very special cases. It may not be the sensor issues, could be confused by autopilot system. The responsible time is too short to give the correct decision. Based on these, a very few accidents, all of are very spacial cases. They may not be properly covered by autopilot. Just like the Boeing 737 crash, it may not properly handle the case caused accident.
Safety systems are not expected to be perfect but better than not having them: seat belts - there were so many now dead and injured who claimed they caused deaths and injuries. air bags - there were so many now dead and injured who claimed they caused deaths and injuries. Safety systems improve survival odds which is how they pay for themselves. Dealing with the 'edge cases' is a much smaller problem than the absence of safety systems. Bob Wilson
Yes! Perfect safety system will be long time to reach. Tesla just eventually is working on approach that point. We are still on the way to that perfect point. Tesla autopilot has covered most accident cases. I believe Tesla has already saved a lot of accidents by its autopilot. 90%, 95% or 99%...
Each quarter, Tesla reports the rates of Autopilot vs non-Autopilot accidents. The ratio is roughly 2-to-1: https://www.tesla.com/VehicleSafetyReport Compared to NHTSA statistic, about 6-to-1. Bob Wilson
Thanks for the information. Very useful. If we study the data in more details, the ratio on autopilot should be even better. Those accidents with autopilot may include at lease two cases that should be removed from statistic. One is the driver did not turn on autopilot or fully turn on autopilot functions. The other is the accident caused by other car's fault. R.C. siway