A contest between high-tech models to develop autonomous driving

22.11.2018
A contest between high-tech models to develop autonomous driving

The fourth edition of the Audi Autonomous Driving Cup shows how it is possible to perform innovative research and development on autonomous driving.

Autonomous driving is undoubtedly one of the key issues for the near future, and Audi intends to take a leading role in this field as well. However, research and development does not just take place in the offices where specialists work or in dedicated testing facilities – they also go a different way, such as the Audi Autonomous Driving Cup: a contest aimed at university students and designed to stimulate their creativity and allow them to direct it in a concrete manner.  Participants are invited to develop completely automatic driving functions, including the underlying software architecture, and to prove their effectiveness in Audi Q2 scale models (1:8). The teams compete on a circuit where the self-driving miniature autos are required to avoid obstacles, to properly approach complex intersections and to recognise traffic signage, as well as maintaining the appropriate distance from the vehicle ahead.

Sensors talk to the computer

The level of difficulty increases with each edition, in line with technological progress in the field. And, just like real Audis, the models competing in the Audi Autonomous Driving Cup are becoming more and more “intelligent”. Artificial intelligence, through self-learning algorithms, helps solve even the most problematic traffic issues quickly, using devices such as front and rear cameras, laser scanners and ultrasound sensors. All the data from these instruments are aggregated in order to reconstruct the surrounding environment as accurately as possible. Moreover, a nine-axis acceleration-sensor platform records every vehicle movement and sends this information in real time to the central on-board computer equipped with a high-speed quad-core processor and a software known as “Automotive Data and Time Triggered Framework”, or ADTF. The vehicle positioning system “reads” map material developed and optimised by experts from the navigation specialist HERE. Combined with optical markers along the route, the cars are thus able to continuously recalibrate their calculated position.

{

A three-stage competition

The Audi Autonomous Driving Cup, held at the Audi Forum in Ingolstadt, involved ten technical universities from four countries: Austria, Great Britain, Germany and Italy. All participating teams were provided with two models and the base software. The competition took place in three stages: after the track testing, teams gave a scientific presentation on their technical developments. The third and final stage, the open challenge, spurred the students’ creativity: the teams made their mini Audi Q2 perform a task of their choice in the most original way, with the emphasis on the use of artificial intelligence (AI). The winners – team AlpaKa, from the IT Research Centre at Karlsruhe University – excelled - in each stage, being awarded the first prize of 10,000 Euros. The technology developed for the open challenge is a car-to-car communication system to handle confusing situations such as emergency braking. In this case the software gives information to the control unit and to the other surrounding vehicles in order to warn them in real time.

Source:  Volkswagen AG

{