Pseudospectral Method for Autonomous Attack Trajectory Planning of a Fixed-wing UCAV

This paper studies the problem of generating optimal guided bomb delivery trajectories on-board for a fixed-wing Unmanned Combat Aerial Vehicle (UCAV). First, a high-fidelity nonlinear 3 Degree-Of-Freedom (DOF) aircraft model is built, which takes platform kinematic and dynamic constraints into account. Second, for the sake of integration of weapon delivery constraints into the problem formulation, an approximate elliptic Launch Acceptable Region (LAR) model is introduced. Third, a new optimal control method - Gauss Pseudospectral Method (GPM) is used to transform the trajectory planning problem into a nonlinear programming problem. To improve computation efficiency and reduce the complexity of initial guess, a multistage iterative optimization strategy is presented. Finally, simulation results demonstrate the efficiency of the method.