The founder of the world's top LiDAR maker says there's one problem with Elon Musk's approach to self-driving

""There's been a lot of debate over whether a vision-only or multi-sensor approach is better when it comes to self-driving vehicles in the past 10 years or so," Qiu told Business Insider on the sidelines of the FutureChina Global Forum held in Singapore in September. "But by now, it is clear that everyone understands that a vision-only approach is not safe enough. There are a lot of corner cases that a vision-only system cannot account for," he added."

"Qiu told Business Insider that vehicles would not be able to achieve Level 3 or Level 4 driving automation capability with a vision-only system. He added that other sensors, including LiDAR, need to be added to the mix to do so. The standards organization SAE International ranks automation systems from Level 1 to 5. Level 1 systems can only provide basic assistance, like automatic braking and lane-keeping, while Level 5 systems can drive a vehicle in all conditions. Tesla's Full Self-Driving software requires human supervision and is a Level 2 system."

Steven Qiu is the founder and chief scientist of RoboSense, a Chinese LiDAR maker. Multi-sensor systems that include LiDAR offer greater safety and reliability for self-driving vehicles than vision-only systems and are necessary to reach SAE Level 3 or 4 automation. LiDAR operates by emitting laser beams and measuring return times to map environments and appears in Waymo robotaxis and consumer products like robot vacuums and smartphone cameras. Vision-only systems fail in many corner cases, such as distinguishing a stopped white car from a white cloud. Tesla's Full Self-Driving remains Level 2, and Elon Musk has called LiDAR "expensive and unnecessary."

Read at Business Insider