中文摘要：

超声电机的小质量、大转矩、响应快等特点，是其成为小型直接驱动机器人执行器的基础。通过对超声电机的伺服特性研究及关节型机器人动力学分析，提出了超声电机驱动的机器人的模糊神经网络控制的复合控制方法，该方法中，模糊子控制器实现了关节定位的初步控制，神经网络子控制器起到降解稳态误差、提高控制精度作用。文中较详细地研究了复合控制器的结构及两个子控制器的设计问题。为了检验控制效果，对控制系统进行了仿真，结果表明，采用这种复合控制器可获得较高的位置精度。

英文摘要：

The characteristics of the USM such as light weight, fast response speed and high torque/mass ratio make it an ideal actuator for the direct drive robot. Based on the servo characteristics of the USM and the dynamics of the manipulator, a hybrid control system composed of a fuzzy controller which makes overall system stable along a trajectory and a neuron network controller which aims to reduce the stable error for a planar robot driven by USM is proposed. The structure of the hybrid contrller and the design of the two sub-controllers have been studied. In order to validate the effect of the control method, a simulation of the control system has been made. The result shows that a high position precise will be obtained with such a hybrid controller.