Heuristic Near-Optimal Uas Path Planning Algorithm for Convoy Overwatch


The optimal path to fly a small unmanned aerial system (SUAS) for convoy overwatch was calculated, heuristically approximated onboard a SUAS autopilot, and demonstrated with hardware in-the-loop simulations and flight test. The optimal path minimized a cost functional consisting of the SUAS’s control effort and deviation from a desired slant range. Due to several hardware and software limitations, the SUAS autopilot was incapable of implementing the optimal controller onboard. This paper introduces a novel heuristic-based algorithm developed in three steps and implemented onboard the autopilot to approximate the optimal solution. The first step manipulated the autopilot loiter logic to allow target tracking. The second step identified three parameters in the loiter logic that were tuned using a Design of Experiments (DOE) methodology. Lastly, a finite state machine (FSM) was created based on the DOE results to further optimize the real-world convoy overwatch algorithm. Each step was tested to evaluate how well it approximated the optimal path. The final algorithm using the FSM exhibited a 65% improvement in tracking performance and demonstrated an implementable, near-optimal convoy overwatch algorithm.

PDF icon Download Article