Command, Control and Communications Systems/Subsystems

Air defense Systems of distributed ground and airborne tracking radars are beset with problems when fusing track data at a central command site. This occurs when one or more of the tracking radars produces a track scene with component track offsets (or biases) from the true target location by a constant amount. The offsets are created by a number of mechanisms, including imprecise knowledge of the radar's absolute location and electromagnetic (EM) propogation anomalies such as ducting. The result is that two track scenes from different radar sites will be incorrectly fused if the track offsets are not properly accounted for. This proposal will attempt to circumvent these problems by using a novel approach called "noncooperative target sensor registration" (NCTSR). This is in contrast to the cooperative target sensor registration (CTSR) algorithms in existence today. The technical approach is a two step problem. First, a track scene-to- scene matching is performed using an optimal assignment algorithm (like Munkres') to identify common targets in the two track scenes. Next, the track pairs that pertain to common targets are passed through another optimization procedure to generate the offset state vector to correct the tracks. This approach has the virtue of operating on the actual targets that are present -- it does not rely on scheduling a known and cooperating target to fly calibrated trajectories for it to operate properly. Thus, it is possible to obtain near continual updates on the radar track offsets without having to expend considerable resources to compute these offsets. The bottom-line is we obtain more accurate track scenes when performing data fusion in a distributed sensor environment.