Audi has exposed its proficiency of intelligent parking strategies, a key aspect in automated driving cars at the Neural Information Processing Systems (NIPS) conference in Barcelona.
With the help of a 1:8 scale model car called the “Audi Q2 deep learning concept”, the brand communicated their intelligent parking process over an area that covered 3x3 meters. The scale model would autonomously hunt and discover a suitable parking space before parking itself in the slot.
What makes this possible is the Q2 deep learning concept’s sensor technology which is made up of two cameras; one facing forward and the other towards the rear, along with ten ultrasonic sensors all around the scale model. A central on-board computer converts the information received into control signals for the steering and electric motor.
Thanks to the model’s ‘deep reinforcement learning’, the system simply learns through trial and error. So while the car selects its direction of travel at random, an algorithm autonomously identifies the successful actions which continuously refines the parking strategy, even if it is difficult.
Audi has stated that the deep learning-based software is expected to first feature on the next generation Audi A8 in 2017, and this will allow the car to be auto-piloted in traffic jam situations as well as parking.
