Friction compensation of a Flight simulator based on disturbance observers and neural networks

The Flight simulator is a typical high-performance servo system. When moving slowly, machines are likely to exhibit stick slip, a periodic cycle of altering motion and arrest, caused by nonlinear frictional dynamics. To solve this problem, this paper studies the compensation methods of nonlinear factors when the servo system is operating at low speed. Based on the description of dynamic and static models of a nonlinear Stribeck friction model, this paper puts forward a compound controller composed of adaptive neural networks compensation and conventional PD controller with a disturbance observer. The simulation results show that the proposed method is effective to suppress the influence of nonlinear factors of the system and it can significantly improve the tracking ability of flight simulator servo system.