You can enable subtitles (captions) in the video player
[MUSIC PLAYING]
Self-driving cars are everywhere now in Silicon Valley. But to operate safely these vehicles must clearly "see" their surroundings. Most autonomous vehicles use a combination of cameras, radar, and LIDAR, which stands for Light Detection And Ranging. LIDAR is considered by many to be the most important of these sensors and also the hardest to build.
To understand how LIDAR works, think of a bat using echolocation to perceive its surroundings. LIDAR does the same thing but using laser beams. The LIDAR sensor emits pulses of light, then measures how long it takes the photons to return, revealing the distance to the objects around it.
However, the rise of self-driving car companies has created a huge new demand for LIDAR sensors. It is also pushing the LIDAR technology to new levels. Today, a top-end LIDAR from Velodyne, the biggest LIDAR maker in Silicon Valley, contains 128 lasers inside a scanner that rotates rapidly, providing the car with a three-dimensional, 360-degree picture of the world. A single one of these complex sensors can cost tens of thousands of dollars.
A competing approach is a simpler design called solid-state, which does not rotate but instead uses mirrors and lenses to flick the lasers across the landscape. Start-ups like Luminar, Innoviz, and Quanergy, as well as Velodyne, are all beginning to produce these cheaper designs. High demand means Silicon Valley has been experiencing a LIDAR shortage, with wait lists that stretch for months. If driverless cars are ever going to become truly widespread and economic, developing an affordable high-definition LIDAR will have to happen first.