Sensor Combinations in Heterogeneous Multi-sensor Fusion for Maritime Target Tracking

Safe navigation for autonomous surface vehicles requires a robust and reliable tracking system that maintains and estimates position and velocity of other vessels. This paper demonstrates a measurement level sensor fusion system for tracking in a maritime environment using lidar, radar, electrooptical and infrared cameras. The backbone of the system is a multi-sensor version of the Joint Integrated Probabilistic Data Association (JIPDA) with both existence and visibility probabilities. Using reference targets equipped with GPS receivers, the performance of different sensors and sensor combinations are evaluated for autonomous surface vehicles (ASVs), Several interesting observations are made, among them that passive sensors can help resolve merged measurements issues in radar tracking, and that the choice between radar and lidar may boil down to a trade-off between fast track initiation and large numbers of false tracks.