Adaptive inversion control of missile based on neural network and particle swarm optimization

As the nonlinear effect and coupling character of the flight dynamics became a big problem to the blended aero and reaction jet flight control system of missile, dynamic inversion was used to make the system decouple and linearize. Because of the effects of actuator saturation, pseudo-control hedging (PCH) was introduced to reduce the level and duration of actuator saturation. Considering fitting characteristics of neural network, we designed an adaptive neural network (NN) controller with a modified particle swarm optimization (PSO) to account for the dynamic inverse error. Meanwhile, the inertial weight of exponential decay was applied to enhance the performance of the PSO. The simulation result proves that the new flight control system conquered the aerodynamic modeling inaccuracies and the external disturbances; the PSO avoided the local optimization of NN and improved the learning efficiency. The compensation of the inverse error is effective and the robustness of the control system is improved greatly.