A Game Theoretic Data Fusion Aided Path Planning Approach for Cooperative UAV ISR

Cooperative and intelligent path planning is important for UAVs to carry out coordinated intelligence, surveillance and reconnaissance (ISR) in adversarial environments. In this paper, we propose a game theoretic data fusion aided platform routing algorithm for cooperative ISR. Our approach consists of three closely coupled components: 1) closed-loop data fusion. The Level 1 (Object), Level 2 (Situation) and Level 3 (threat) data fusion form a closed-loop structure, in which Markov game theoretic intent inferences will execute from the results of Level 1 and Level 2 results. The estimated threat intents will be fed back to the Level 2 fusion to improve the performance of the entity aggregation. 2) Cooperative platform routing based on Pareto-optimization, social foraging, and cooperative jamming. Given the threat information including the threat intents from the data fusion module, a Pareto-optimal problem is formed and graph-cut based fast solution serves as a reference trajectory for a foraging algorithm, which further dynamically refines the reference path to avoid pop-up obstacles detected along the planned path. 3) display/monitor module, in which relevant threats and constraints information are indicated, the terrain data are shown, and current real route and planned route are highlighted, compared, and evaluated. The commander's suggestions can be inputted in this mode.