A Model-Based Joint Detection and Tracking Approach for Multi-Vehicle Tracking With Lidar Sensor

摘要: This paper presents a method for joint detection and tracking of vehicles with scanning laser rangefinder. The lidar measurements an object have the particularity to be spatially distributed, which generally leads step before any tracking. Differently, proposed relies on raw measurement processing without step, improves overall performance in multiobject while providing good estimation accuracies. solution uses sequential Monte Carlo methods by incorporating geometric invariant objects interest (vehicles). approach also offers efficient problem multitarget integrating naturally track management filtering process.