Also in this session we support the Formula Student racing team “Elefant Racing” of the University of Bayreuth with our stereo vision technology. Since the last year, our SceneScan Pro and the Karmin3 stereo camera are already in use in the team’s racing car. Our 3D stereo vision system mainly supports the discipline of Driverless and Autonomous driving.
Formula SAE is an international student competition that is held at various locations around the world. It is about the design and production of racing cars by interdisciplinary student teams. The teams compete with their racing cars in various disciplines. In addition to the technical implementation of the car and its performance in various races, the competition also focuses on non-technical areas such as cost structures, business plans and efficiency.
The Racing Car FR20 Ragnarök
The Elefant Racing Team of the University of Bayreuth has now been active for over 20 years and is therefore no longer a newcomer to the racing circuit. With the FR20 Ragnarök, they have constructed the first autonomous racing car in the club’s history.
In the 2020/2021 season, the Elefant Racing Team has already successfully participated with the car in several events for the Electric and Driverless discipline.
The FR20 was constructed entirely by the team itself and has a Monocoque made entirely of carbon. Part of the success story are the self-developed components such as the battery, the electric motors and the gearbox. Thanks to the four electric motors and the integrated all-wheel drive, an impressive 190 hp and 1,564 Nm can be safely brought onto the race track.
How the Stereo Vision System is Used
Nerian’s stero vision system is attached to the roll bar, also called the main hoop, of the race car. Together with two other mono-ocular cameras, it forms the race car’s perception systems and thus the basis for the system. Connected to the powerful SceneScan Pro stereo vision sensor, the perception system is directly connected to the driverless computer via Ethernet. Another advantage of the Nerian system is the fact that the SceneScan Pro stereo vision sensor allows simultaneous triggering and synchronization with another pair of cameras.
The system mainly provides the basis for recognizing the race track. The course of the race track at international Formula Student events is defined and delimited by a large number of traffic cones. The task of the stereo vision system is to recognize these and also to estimate their relative position to the car. The distance of the objects plays a decisive role here, as distant objects present greater challenges. The recognition of traffic cones is best achieved in color camera images. However, position estimation, especially distance, with mono-ocular cameras is inaccurate and unreliable. Therefore, this option is only suitable for close objects. So, to measure the distance of more distant cones as well, the combination of Karmin3 and SceneScan Pro is used.
The particular added value of Nerian’s 3D stereo vision system is the ease of use and integration of the system thanks to the software and API provided, specifically for ROS. Thus, the use of the stereo vision system also saved a considerable amount of development effort in the construction of the vehicle.
In addition, the simultaneous calculation of the stereo matching on the SceneScan Pro is much more efficient than on the computer. This reduces power consumption and significantly increases the performance of the entire vehicle. Thanks to the low processing latency and the high frame rates, we were able to make the FR20 Ragnarök drive faster and more efficiently through the respective race tracks.
Nerian we are thrilled with what the Elefant Racing Team has already accomplished and the success they have had. We are proud to support such a great team and look forward to an exciting season. We wish them good luck and success in the upcoming races and the further development of the race car.
We are already looking forward to the presentation of their new race car. Details will follow!