The new Audi A8 is the first car in the world developed for level 3 autonomous driving. With Audi AI traffic jam pilot, the A8 can drive itself in slow-moving traffic at up to 60 km/h. However, a fundamental requirement for automated driving is a mapped image of the environment that is as precise as possible, and at all times. Artificial intelligence is a key technology for this.

As a company at the forefront of autonomous driving technology, Audi is set to exhibit a new pre-development project at the world’s most important symposium for artificial intelligence (AI) – the Neural Information Processing Systems (NIPS) conference.

Audi Electronics Venture (AEV), a subsidiary of Audi and the team responsible for this project, says the new pre-development project utilises a mono camera equipped with AI to generate an extremely precise three-dimensional model of a vehicle’s environment.

The system comprises a front camera that acts as a sensor, which has a viewing angle of 120 degrees and is capable of capturing up to 15 images per second at a resolution of 1.3 megapixels. This set of pictures then gets processed in a neural network where semantic segmenting occurs.

Each pixel is classified into one of 13 object classes, enabling identification and differentiation of other cars, trucks, houses, road markings, people and traffic signs. The system also uses neural networks for reading distance, and works by way of ISO lines (virtual boundaries that define a constant distance). In essence, the system makes it possible to capture the exact surroundings of the car.

With the help of “unsupervised learning” (a method of learning from observations of circumstances and scenarios that does not require pre-sorted and classified data), the neural network has learned to independently understand rules and holds great potential for computed interpretation of traffic situations.

Separately, two partners from the Volkswagen Group will also present their own AI topics at NIPS, one of which is the Fundamental AI Research department from the Group IT’s Data:Lab, focusing on unsupervised learning and optimised control through variational inference – an efficient method for representing probability distributions.

Next, an Audi team from the Electronics Research Laboratory of Belmont, California, will demonstrate a solution for pure AI-based parking and driving in parking lots and on highways. Here, lateral guidance of the car is completely done through neural networks. The AI learns to independently generate a model of the environment from camera data and steer the car (in real-time!), all without the use of pre-existing map data.

In 2016, Audi became the first automobile manufacturer to participate at NIPS with its own exhibition booth. This year, it makes a return as co-sponsor of the event, which will take place from December 4 to 9.

GALLERY: 2018 Audi A8