Jay Gierak at Ghost agrees with the company's overall viewpoint and is impressed by Wayve's demonstrations. The way to do this is not with the robot.
He doesn't like Wayve's commitment to deep learning. Ghost trains hundreds of smaller models each with a specialism. The self-driving system is told which models to use in certain situations. The approach taken byGhost is similar to the one taken by Autobrains. Autobrains uses more than one neural network to learn the rules.
It is easier to establish that a vehicle is safe when it is split into smaller pieces. Something will happen at some point. If there is a person in front of you, you have to stop. There needs to be a piece of code. It would be difficult to find the code in a large model like Wayve's.
Ghost wants to make consumer vehicles that can drive themselves on freeways while Wayve wants to be the first company to put self-drive cars in 100 cities. Wayve is working with two UK grocery giants.
Both firms are behind the market leaders by a wide margin. The two companies have racked up hundreds of hours of driving without a human in their cars and are now offering a service to the public.
“I don't want to diminish the scale of the challenge ahead of us,” says Hawke. “The AV industry teaches you humility.”
All Rights Reserved © 2024
