Mark Rober's recent video critiques Tesla's dependence on visual-only data for its Autopilot system, highlighting significant safety risks compared to competitors using LIDAR. In his demonstrations, Tesla vehicles fail to react appropriately to obstacles like a child mannequin, especially in adverse conditions such as fog or rain. Rober's experiments indicate that this cameras-only approach could easily mislead drivers, particularly with Tesla's impending unsupervised 'Full Self-Driving' rollout. Experts voice concerns about the dangers of eliminating LIDAR and radar and the overconfidence placed in the current Autopilot technology.
During a series of tests, Rober compares how a Tesla and a test vehicle equipped with LIDAR sensors, courtesy of the company Luminar, react to a child mannequin in the middle of the road.
Despite registering the kid in the software, the vehicle's emergency braking system refuses to jump into action in time, causing Rober to plow right through the mannequin.
Collection
[
|
...
]