基于神经网络的水下灵巧手阻抗控制  

Impedance control of underwater dexterous hand based on neural network

作　　者：王华[1] 孟庆鑫 何晋[1,2] 王立权[1] 

机构地区：[1]哈尔滨工程大学机电工程学院,黑龙江哈尔滨150001 [2]哈尔滨工业大学机电工程学院,黑龙江哈尔滨150001

出　　处：《哈尔滨工程大学学报》2006年第B07期124-128,共5页Journal of Harbin Engineering University

基　　金：黑龙江省自然科学基金资助项目（E0301）.

摘　　要：水下灵巧手抓取物体时，物体与指尖存在力控制问题，但是由于动力学模型、被抓取物体位置和刚度的不确定性，采用传统阻抗控制方法不具有鲁棒性．文中基于位置型阻抗控制方法，提出采用神经网络对手指动力学模型、物体刚度和物体位置误差进行补偿．对补偿策略进行了详细的推导，并通过仿真实验验证了该方法的补偿效果，结果表明基于神经网络的位置型阻抗控制器具有较强的鲁棒性．The contract force between the fingertip and object need to be controlled when the underwater dexterous hand grasp the object. The traditional impedance control is known to lack robustness due to unknown dynamic model, position, and stiffness of the object grasped. Based on the position-based impedance control method, this paper introduces a new method adopting neural network, which can compensate the uncertainty of finger dynamics, object position, and stiffness. The compensating strategy is deduced in details, and the simulation studies have been carried out to test the compensation effect in these three conditions. The results show the position-based neural network impedance control is a robust method fitting for the control system of the underwater dexterous hand.

关 键 词：水下灵巧手 神经网络 阻抗控制 力控制 

分 类 号：TP241[自动化与计算机技术—检测技术与自动化装置]