Neural network adaptive control of high-precision flight simulator: Theory and experiments

This paper developed a control scheme of neural network based on feedforward and PD (proportional and derivative) control for high-precision flight simulator. A radial basis-function neural network (RBFNN) controller was used to learn and to compensate the unknown model dynamics, parameter variation and disturbance of the system on-line. The iterative algorithm of RBFNN parameters is got by Lyapunov stability theory. The effectiveness of the proposed control scheme is evaluated by simulation results and a real-time flight simulator system experiment. It is found that the proposed scheme can reduce the plant´s sensitivity to parameter variation and disturbance and high precision performance of flight simulator can be obtained.