Self-driving
Technology that lets a car perceive its surroundings with cameras and other sensors and drive itself without a person at the controls.
In 2005 Stanley, a car from Stanford University, completed the roughly 210-kilometre desert course of a driverless-vehicle race organised by DARPA, the US defence research agency[1]. Rather than being given hand-written rules for recognising drivable ground, it learned them from driving data, and that choice set the direction for self-driving development afterwards.
Today the dominant approach processes camera and radar data with neural networks. Seeing the surroundings, understanding the situation and deciding how to move is the same loop that robots in factories and warehouses need. That is why carmakers are moving into humanoid robots: they can reuse both the technology and their mass-production plant. Handling situations that occur only rarely, and deciding who is responsible after an accident, remain unsolved in both fields.
Related terms
- Humanoid robotA robot with a human-like body — torso, two arms and two legs. Its advantage is being able to use environments and tools built for people.
- Neural networkA system of simple computing units called neurons, connected in many layers, that learns from data how strong each connection should be and so becomes able to make complex judgements.
- Reinforcement learningLearning better actions by trial and error, guided by a reward given according to the outcome, rather than by being told the right answer.