For the first time, Audi has participated in the Conference and Workshop on Neural Information Processing Systems (NIPS) that begun from 5 December 2016 and will go on untill 10 December 2016. The event presents advances in the fields of machine learning and computational neuroscience and is regarded as one of the world’s most important specialist conferences for artificial intelligence. Audi is the only automaker represented at NIPS with its own stand and has brought some of their hi-tech toys to the Barcelona event to showcase its expertise.

With the help of a scale model Audi has demonstrated as to how a car develops intelligent parking strategies. The 1:8 scale model car used is dubbed as the ‘Audi Q2 deep learning concept’. It can autonomously search for and find a suitable parking space in the form of a metal frame, and then park itself there. All of this can be executed by the little Audi Q2 on an area measuring 3 x 3 meters. Self-learning systems are a key technology for piloted driving cars. That’s why Audi has already built up a wealth of know-how in machine learning. The company is the only automaker represented at NIPS with its own stand and a showcase.

The Audi Q2 deep learning concept is a pre-development project of Audi Electronics Venture (AEV). The developers at AEV are planning to transfer the parking space search process to a real car soon.

The tech on the Audi Q2 deep learning concept’s consists of:

  • Two mono cameras – one facing forward and the other towards the rear
  • Ten ultrasonic sensors positioned at points all around the model
  • Central onboard computer to converts data into control signals for steering and the electric motor

The concept executes its parking skill in the following process:

  • On the driving surface, the model car first determines its position relative to the parking space.
  • As soon as it perceives the position, it calculates how it can safely drive to its targeted destination.
  • The model car manoeuvres, steers and drives forward or in reverse, depending on the situation.

Deep reinforcement learning helps the model car in parking. The folks at Audi states that the system essentially learns through trial and error. To begin, the car selects its direction of travel at random. An algorithm autonomously identifies the successful actions, thus continually refining the parking strategy. So in the end, the system is able to solve even difficult problems autonomously.

Audi is working with partners including Mobileye who have an expertise in the field of image recognition. The two companies have teamed up to develop deep-learning-based software for environment perception systems and will use the software for the first time in 2017. Audi will be using the software in the central driver assistance controller (zFAS) in the new generation of the Audi A8. Nvidia was an important partner in the development of the zFAS.