Cam mechanism is widely used in interactive movement systems because of its excellent movement characteristics. But the changing torque and inertia moment make the speed stabilizing control of the systems with cam mechanism very difficult. The changing load torque is the primary one in all the factors that influence the speed control accuracy. This thesis makes full use of the approximating to nonlinear functions capability of neural networks and establishes the load torque neural networks model of the system with cam mechanism. Based on the whole compensating theory

a speed stabilizing control system with neural networks compensatory loop is designed. Experiment results convince that the method works very well and can be used in other control systems with changing load torque.