Innovusion: Bringing Safety and Reliability to Self-Driving Cars

A common trait among new industry segments fueled by innovations, the pitfalls outweigh the degree of reliability. Case in point: autonomous driving—the new wave in the automotive industry—and the roadblocks car manufacturers and early adopters of self-driving cars are faced with. In the existing landscape, the automotive industry has yet to bring to market a self-driving car that meets criterions such as safety and affordability. Conversely, high-tech companies hoping to venture into autonomous driving haven’t successfully hatched a business model that can generate revenue through self-driving cars. Most of the challenges stem from Level 3 and Level 4 self-driving cars and not the fully-automated Level 5 variants—currently too expensive for mass production. In Level 3 vehicles, the autonomous driving system alerts the driver to take control in certain scenarios that are beyond its capacity. In Level 4, the system can overcome pre-defined scenarios. Either way, safety, and reliability leave a lot to be desired in self-driving cars, and industry experts cite a common denominator for the shortcomings—expensive technology. “When a Level 3 or Level 4 self-driving car runs into an accident in the future, would the manufacturer concede that they didn’t utilize safer methods, sensors, and expensive technologies? In many ways, safety is being compromised, and that is not acceptable,” says Junwei Bao, co-founder and CEO of Innovusion, a startup based in Silicon Valley—the epicenter of the advanced driver-assistance systems (ADAS) industry.

On a quest to make self-driving cars safer, Innovusion has designed the world’s first image-grade LiDAR (Light Detection and Ranging) system intended to create a high-density point-cloud for self-driving perception, decision, and control. Designed to provide crucial information about dynamic environments, a high angular resolution allows the LiDAR to detect cars for up to 270 meters, small objects, and even reading signage.The image-grade technology makes it seamless to reliably detect vehicles that are hundreds of meters away. “In Level 3 self-driving cars, our goal is to allow drivers to be hands-off and minds off—meaning they can afford to be distracted rather than being prompted to take control. The transition to manual driving could take up to 10 seconds due to psychological and physical limitations.