AI should help increasing road safety for robot taxi allowing to put them in function soon
Robot taxis are under testing in many cities such as Phoenix Arizona. Globally it works but the issue is still the rate of accident. Human driver has an accident in average about every 70 000 km and there are about 3 deads every billion km.
This very high performance of human driver makes it difficult to automate driving, especially in urban traffic where road infrastructure characteristics are very complex. This explain why tests long more that expected.
A French sartup (NEXYAD) recently lauched the AI module SafetyNex. SafetyNex is a real time driving risk assessment system that estimates (in particular) risk that comes from a driving behaviour (speed, accelerations) that is not appropriate to driving context (in particular : infrastructure complexity).
When driving risk (estimated by SafetyNex) rises, the robot taxi slows down (smoothly) and then avoids to be in situation of a potential emergency/accident.
The use of SafetyNex for autonomous robot taxis, should contribute to decrease accident rate and launch their deployment in cities.
See more : https://www.youtube.com/watch?v=24nCV5zWcws