Tesla autopilot learned to recognize motorcyclesJanuary 4, 2019
The developers of the Tesla autopilot “taught” him to see motorcyclists who ride between the rows. In terms of safety, the update is very important, as motorcycles are one of the biggest problems for autonomous vehicles.
By itself, the motorcycle has small dimensions, and also often moves along the road relatively chaotically. Electronic system is more difficult to track a motorcycle than any other vehicle. To solve the problem, Tesla developers took up an advanced neural network that is able to detect multiple vehicles, both large and small.
The video below shows Tesla Model 3 with Version 9. The car is capable of seeing motorcycles moving in the aisle. The system works imperfectly. She sometimes takes a motorcycle for a car or places it in the next row … If the cameras did not work, there are always sound sensors that warn the driver when trying to change lanes.
Of course, there is still a lot of work, but the problem has been identified correctly and is moving in the right direction.
- Hyundai has developed a new type of airbag
- Bentley made a cowboy Bentayga
- The launch of the new Genesis G80 is delayed until the beginning of 2020
- Ford plans to use Volkswagen MEB platform
- Analysts say BMW should buy Jaguar Land Rover