Optimal allocation of multi-platform sensor resources for multiple target tracking

This paper presents graph algorithms for computation of the exact optimal solution of the Radar Resources Distribution Problem. These algorithms can be applied for single and multiple vehicles on a target field and for single and multiple target tracking. This problem involves the computation of the optimal flight paths for a number of observing vehicles (aircrafts or UAV) on a target field for radar observations of multiple targets (or track clusters for multi-target trackers). Each vehicle had restricted resources of time or fuel and fixed start and finish position. Each target or cluster was assigned a priority and position on the target field. Optimization criteria were: maximize number of observed targets or clusters with highest priorities for all vehicles on target field with the constraint that each target could be observed by only one vehicle. The exact optimal solution was found as the result of a fixed number of iteration steps on a multi-level tree, which represents the full partially-oriented graph of flight paths with time-weighted connectivity matrix and priority-weighted vertices. The algorithms are designed to work in real-time: optimal flight paths are recalculated for each vehicle each time that the situation on the target field changes — targets or tracks can be added, removed, moved or their priorities and positions can be changed. These algorithms can be applied for stationary and moving targets, as well as for combinations of ground and airborne targets. Separate algorithms were designed for cases of single and multiple vehicles on the target field. Each target can be observed by a vehicle once or several times during flight path. Algorithms can be applied for single and for multiple-target tracker radars. Closely spaced targets or tracks can be grouped in clusters to speed up computations. Also, for multiple targets tracking of ground targets these algorithms can be applied for computation of optimal flight path to cover intersections with high numbers of track ambiguities. The same algorithms can be applied for the Combat Management Problem: to optimize flight paths of vehicles (aircraft or UAV) on the combat field to distribute troops, weapons, and ammunition to different areas of the combat field with different priorities. The algorithms were implemented in C-code. Results are shown to illustrate the methods.