Abstract: We present a radial basis function neural network (RBFNN)tuning based PID control strategy.It is then applied to the tracking control of flexible joint inverted pendulum. It acquires Jacobian information of flexible joint inverted pendulum by the identification of neural network,and adaptively adjusts PID control parameters by gradient descent method.Simulation results show that the method has rapid response and small overshoot, as compared with other conventional PID control methods. It better solves the difficulty inparameters tuning of PID control method,and implements effective tracking and control over flexible joint inverted pendulum.

Key words: radial basis function neural network, tracking control, flexible joint inverted pendulum