Road curb tracking in an urban environment

Publication Type:
Conference Proceeding
Proceedings of the 6th International Conference on Information Fusion, FUSION 2003, 2003, 1 pp. 261 - 268
Issue Date:
Full metadata record
Road detection and tracking is very useful in the synthesis of driver assistance and intelligent transportation systems. In this paper a methodology is proposed based on the extended Kalman filer for robust road curb detection and tracking using a combination of onboard active and passive sensors. The problem is formulated as detecting and tracking a maneuvering target in clutter using onboard sensors on a moving platform. The primary sensors utilized are a 2 dimensional SICK laser scanner, five encoders and a gyroscope, together with an image sensor (CCD camera). Compared to the active 20 laser scanner the CCD camera is capable of providing observations over an extended horizon, thus making available much useful information about the curb trend, which is exploited in mainly the laser based tracking algorithm. The advantage of the proposed image enhanced laser detection/tracking method, over laser alone detection/tracking, is illustrated using simulations and its robustness to varied road curvatures, branching, turns and scenarios, is demonstrated through experimental results. © 2003 ISlF.
Please use this identifier to cite or link to this item: