本文提出用人工神经网络和常规ＰＩＤ控制器构成复合控制器，控制多指灵巧手的运动．系统在跟踪阶段用神经网络控制器，而在稳态阶段用ＰＩＤ控制器，利用ＰＩＤ算法的积分特性去消除稳态误差．实验证明这种复合形式的控制器，满足机器人伺服控制的速度和精度要求 In this paper, artificial neural network and conventional PID controller to form a composite controller to control the movement of more than one dexterous hand. The system uses the neural network controller in the tracking phase, and uses the PID controller in the steady state phase, using the integral characteristic of the PID algorithm to eliminate the steady state error. Experiments show that this kind of composite controller meets the speed and accuracy requirements of robot servo control