If you missed Innoviz’s June 16, 2020 webinar, “LiDAR vs. Camera: Can Autonomous Vehicles Really Waive One?”, you’ve come to the right place. Mati Shani, VP Product at Innoviz, and Assistant Professor Dr. Raja Giryes from the Electrical Engineering department at Tel-Aviv University gave an engaging and informative presentation, which you can watch the recording of and read about here.
The presentation addresses the ongoing industry debate about whether autonomous vehicles (AVs) could be safe with just camera technology’s sight capabilities. Spoiler Alert: they would not. The presentation makes a strong case for the need for multiple sensors through three sections:
- Understanding the three main AV sensing technologies (camera, radar, and LiDAR)
- Comparing the sensing technologies
- Sharing the advantages of fusing the sensing technologies together
As the presentation started, Mati Shani thoroughly explained the three technologies, including how they work and embody their capabilities. Then, Dr. Raja Giryes used a map of Sensing Functions to show the various and differing advantages of camera, radar, and LiDAR technology. Through vivid visual examples, taking place from the 18-28-minute mark, Dr. Raja Giryes showed how LiDAR fills in camera’s sight gaps based on conditions including color, depth, sun, rain, low light, rapidly-changing light, and more. Finally, Dr. Raja Giryes offered a brief summary about how fusing the three sensors will lead to safety benefits that far outweigh the costs.
The presentation made a clear case: the AV accidents we’ve seen to date are not coming from glitches. They reflect a flaw in the design that the industry can only fix by bringing the different sensing technologies together. Mati articulated this point at the start of the presentation well. “Vision impaired and hearing-impaired people have many capabilities since they have other senses to use,” he reminded us. “Similarly, what happens when a camera or radar fails? What replaces them? Bottom line: multiple senses increase certainty and redundancy, which lead to much greater confidence in the performance and safety of the AV system.”