Tesla's approach to achieve full autonomy is different from that of other companies. Whereas Waymo, Cruise, and other companies are relying on highly detailed (centimeter-scale) three-dimensional maps, lidar, and cameras, as well as radar and ultrasonic sensors in their autonomous vehicles, Tesla's approach is to use coarse-grained two-dimensional maps and cameras (no lidar) as well as radar and ultrasonic sensors. Tesla claims that although its approach is much more difficult, it will ultimately be more useful, because its vehicles will be able to self-drive without geofencing concerns. Tesla's self-driving software has been trained on over 20 billion miles driven by Tesla vehicles as of January 2021. Tesla also designed a self-driving computer chip that has been installed in its cars since March 2019.
How is Tesla's approach to full autonomy is different from the other companies?
Tesla uses e coarse-grained two-dimensional maps and cameras (no lidar) as well as radar and ultrasonic sensors whereas its competitors rely on highly detailed  (centimeter-scale) three-dimensional maps, lidar, and cameras