The evasive navigation of an aircraft in the presence of enemy radar or SAM (surface-to-air missile) sites or AWACS is formulated using a time-varying Flexible Delaunay Triangulation (FDT) and an artificial-intelligence-based multi-objective search. It is implemented on a SUN 4 workstation. The system displays the circumdisks, the FDT, and the connected graph for generation of candidate paths. The multi-objective A* algorithm generates the path which satisfies the following requirements: 1) avoidance of the detection range of the moving obstacle; and 2) minimum deviation from the shortest path to the goal. The system then displays the motion of the vehicle and of the moving radar/SAM site in real time. In case of user interruption (for the purpose of inserting, in a simulation, new obstacles crossing the designated path), the FDT retriangulates locally. The closest line to the path is drawn, and A* search is once again performed for generation of the optimum path. Background material on relevant aspects of computational geometry, intelligent control and heuristic search is also given, providing a survey character to this work.
Bibliographical noteFunding Information:
This research was supported by AFOSR under Grants No. 870252 and 899158. presented at the Fifth AAAI Conference, Dayton, OH, U.S.A., October 1989.