Wayve rewrites self-driving playbooks with deep learning on Azure

Machine Learning


LONDON – It was a busier-than-usual Thursday morning in Soho, with rain falling against a gray December sky. Traffic was mostly stopped, with some stops and stops. The sidewalks were also congested.

Finally, the flow of cars and trucks regained some momentum next to the imposing British Museum. Inside the self-driving four-door EV sedan, a safety operator sat passively but attentively behind the wheel, his palm resting on his thigh. The car skidded forward without his help on the way to Trafalgar Square.

Moments later, a man in a hurry steps into our path from behind a parked car. The AI-guided sedan braked firmly to a stop, and the four passengers inside felt a slight tremor. A careless pedestrian crossed the street without looking back. The safety operator did not touch the pedal. This car was acting alone.

Self-driving cars powered by AI are roaming the streets of many major cities these days, but the company behind our vehicle, Wayve, took a different path when it was founded in Cambridge, England, in 2017.

Essentially, Wave has developed an AI-powered driver that can be installed in any new car, regardless of make or model, and potentially driven in any country or city, with just two weeks of fine-tuning. This approach relies on a form of AI model inspired by the human brain known as a “neural network.” Wayve's AI driver primarily uses a camera to get you safely from point to point.



Source link