Using grid maps to reduce the number of false positive measurements in advanced driver assistance systems

In Advanced Driver Assistance Systems (ADAS), object tracking is a crucial method to foresee dangerous situations. The Joint Integrated Probabilistic Data Association (JIPDA) offers the advantage, that existence and association uncertainties are considered in multi-target tracking. Recent- ly, real-time capable implementations have been presented. However, the real-time capability is only given, if a certain number of tracked objects is not exceeded. Thus, so called false positive object detections yield a problem. To mitigate this issue, additional information about the vehicle's environment is used to identify measurements that are not relevant. The idea is to focus on moving objects for tracking. As an example, an Occupancy Grid Map is used to distinguish between stationary and non-stationary objects. The approach is evaluated using real-world data of a research vehicle.