German car manufacturer, Audi, unveiled its development plan for future smart cars at the Conference and Workshop on Neural Information Processing Systems event in Barcelona, Spain. The development, still in its concept stages, introduces technology to that will give the car the ability to maneuver in tight spaces and to park itself without human intervention. They used a 1:8 model car and named it the Audi Q2 deep learning concept.
The Audi Q2 Deep Learning Concept
The Audi Q2 deep learning concept was placed and tested in a small area and had it autonomously seek out an available parking space, and had it park itself. They placed the model car in different areas of the small field and had it seek out the spaces, some with obstacles that simulate other cars in a real-life parking lot. The Audi Q2 was successful in seeking out spaces and was able to properly park. The program uses algorithms to learn by means of trial and error and make the best judgment call based on the scenario. Two cameras were placed on the model car, one in front and another at the back, and were equipped with 10 ultrasonic sensors to mounted around the body to detect what's in its surroundings.
Audi Plans To Bring The Technology To Real-World Cars
Audi plans to use the same deep learning concept on a full sized car in the near future. Deep learning software will also be developed with Mobileye and will be equipped to the next generation Audi A8 sedans. The A8 is expected to be a semi-autonomous vehicle that is include features such as traffic-jam assist, this would allow the car to intelligently maneuver through stop and go traffic with minimal human intervention. A self-parking system is also expected to be featured for the next generation Audi A8 sedan.
