Performance Evaluation of a Track to Track Sensor Fusion Algorithm

The development of autonomous vehicles requires algorithms responsible for constructing the local surroundings of the vehicles. The algorithm responsible for creating a common model from the sensors' detection is the so called sensor fusion method. The performance of these methods should be compared to each other, using a common metric, in order to select the proper algorithmic approach, sensor set or sensor placement. This paper describes the evaluation of a track to track sensor fusion algorithm on a dataset recorded on highways.