[If you are driving, you will slow down and stop when you cannot see the road ahead clearly due to the external objective environment or your own physical conditions], this is basic common sense. The risk of breaking basic common sense at the expense of personal safety is too high! The traffic accident of [Model S and Uber autonomous driving] shows that the [logical model] of [autonomous driving] has serious flaws.
It is wrong to compare the accident rate of human driving with the accident rate of autonomous driving in isolation: it is impossible to design "drink driving mode", "fatigue driving mode", "inattention mode", " Overspeed mode”, etc., as long as no common sense mistakes are made, the safety of autonomous driving is far greater than that of human driving; because the accuracy of conventional sensors (cameras, ultrasonic sensors, ordinary radar) has exceeded that of human beings, so the fundamental of not making common sense mistakes lies in automatic driving. The rationality of driving the [logical model]; in the system where the [logical model] is unreasonable, focus on higher precision and more forms of sensors such as lidar and infrared cameras. I believe that continuous patching will eventually lead to It will definitely solve the problem, but it's not a serious work ethic.
Based on the existing mature technology, with reasonable hardware cost, through [remodeling, process reengineering], I can solve the problem within 3 to 5 months.