An overview of a real time sensor fusion vehicle detection and tracking system incorporating monocular video and lidar measurements has been presented. The state estimation part including time evolution and measurement update models was explained in detail. The benefit of sensor fusion on state estimation can be shown by comparison with a single sensor detection and tracking system. The presented system estimates 3D positions as well as the velocities and yaw rates of the other vehicles for applications as ACC stop and go, traffic jam warning, emergency braking, and cutting-in vehicle warning. Even standing vehicles are detected in distances up to 150 meters and the system can keep track of already detected vehicles up to 200 meters. The system is implemented in a Mercedes-Benz S-Class with serial night vision camera and a multi beam lidar sensor with 16 measurement channels. The current development phase allows non-synthetic invehicle demonstrations in real traffic scenarios.
Feature-level video and multibeam lidar sensor fusion for full-speed ACC state estimation
Sensorfusion zur Gegenstandserkennung mit Video und mehrstrahliger LIDAR für die Ortserkennung bei voller Geschwindigkeit im Rahmen der aktiven Fahrerunterstützung
2001
6 Seiten, 8 Bilder, 13 Quellen
Aufsatz (Konferenz)
Englisch
British Library Conference Proceedings | 1999
|Multibeam LIDAR ACC - Approaching the start of production
Tema Archiv | 2005
|