Tesla is trying to teach robotaxis not to hit cats at night

Bringing self-driving cars onto public roads involves many challenges. One of them, according to Elon Musk, has a mustache, paws and a tail. Due to the increase in robotaxi operating time to 23 hours, the development team is trying to solve the most difficult problem - teaching cars not to hit cats at night, writes Business Insider.

“The main goal we are trying to achieve is to avoid running over pets that are hard to see in the dark. We are literally trying not to run into gray kittens on gray asphalt,” Musk said. Cats and dogs, which often unexpectedly run out onto the road, can be more dangerous than the flow of cars, the head of the company believes.

According to Tesla, Cybercab uses cameras and sensors for navigation. Tesla vehicle operating instructions note that low or limited lighting, including unlit or poorly lit roads at night, may reduce the camera's ability to provide accurate visual information.

Musk cited artificial intelligence image processing as a way to improve visibility at night. “Visible cameras see very well in the dark when artificial intelligence analyzes the number of photons,” he said.

The leader in robotaxi, Waymo, uses a different system than Tesla - LiDAR (laser radar), which scans the environment using laser beams. This technology did not help prevent a fatal incident: last year, one late evening, a nine-year-old tabby cat named KitKat fell victim to a Waymo robotaxi. The accident turned out to be resonant.

A year ago, a Tesla Model Y, working as a robotaxi service, hit a dog that ran onto the roadway. At the time, a report filed with the National Highway Traffic Safety Administration (NHTSA) stated that the dog ran away from the scene, likely without injury. The regulator will now study Tesla's camera system for fully self-driving cars and its effectiveness in detecting road hazards in poor visibility.