A Potential Game Approach to Target Assignment in Heterogeneous Manned/Unmanned Aerial Team With Incomplete Information

This brief proposes a potential game method in a distributed manner to address the target assignment problem with incomplete information in heterogeneous manned/unmanned aerial teams. Firstly, a heterogeneous system is formulated as a graph and different utility functions are designed to characterize the utilities of the two kinds of members. Secondly, a local information matrix is developed for each member to infer the actions of others. Finally, an improved learning algorithm is proposed to achieve the optimal solution. Simulation results show the feasibility and effectiveness of the potential game approach with information estimates.