Redundant sensor system for detecting an autonomously driving vehicle's environment
Feb. 01, 2017 to Jan. 31, 2020
In the KameRAD project, a redundant sensor system for detecting an autonomously driving vehicle's environment in urban and agricultural environments was developed. The project partners implemented a fail-safe sensor system for all-round protection during autonomous driving. They built a highly integrated electronic system consisting of a camera and radar sensor that combines the advantages of optical monitoring of the environment with radar measurement technology. In the future, these camera-radar systems will be flexibly linked with each other and with GPS and Car2X information in a decentralized network. The overall system should be able to be used in autopilot through the use of artificial intelligence and be used for autonomous vehicle control.
In the project, scientists at Fraunhofer FOKUS worked on sensor data fusion between the radar and camera systems. Various calibration and registration procedures have been developed for this purpose. Based on this, the scientists implemented a low-level sensor fusion, which matches the pixels or feature points between the 3D stereo camera and the radar, through unified coordinate systems and optical flow analysis.
To verify the results and to determine the accuracy, a comparison using a high-precision LIDAR system was conducted.
5.52 million euros. The Federal Ministry of Education and Research (BMBF) funded the project, with about 60 percent of the total funding amount.
- AVL Software and Functions GmbH, Regensburg
- Fraunhofer-Institut für Zuverlässigkeit und Mikrointegration IZM, Berlin
- Fraunhofer-Institut für offene Kommunikationssysteme FOKUS, Berlin
- InnoSenT GmbH, Donnersdorf (Verbundkoordinator)
- Jabil Optics Germany GmbH, Jena
- John Deere GmbH & Co KG, Kaiserslautern
- Silicon Radar GmbH, Frankfurt an der Oder
- Technische Universität Berlin