Aerial Pursuit-Evasion Game Using Nonlinear Model Predictive Guidance

Aerial pursuit-evasion games pose a great challenge in obtaining the analytic solutions for highly nonlinear dynamics of the pursuer and evader. This paper proposes a nonlinear model predictive method as a tool for obtaining the numerical solutions to the pursuitevasion game. In the nonlinear model predictive method, the current control action is obtained by solving a finite-horizon optimal control problem based on nonlinear models of the pursuer and evader. The gradient descent method is implemented for the on-line optimization. Two types of performance measures, distanceand vector-based, are employed for the pursuit-evasion game. The performances under the two payoff functions are compared, and the capture sets for the vector-based payoff function are computed with respect to the varying initial flight path angles of the pursuer and evader.