权益分类	功能权益	普通用户	{{item.name}}会员
{{category.name}}	{{benefitItem.name}}

Study on the control of the position / force of the manipulator with the neural network model

神经网络模型对机械臂位置/力控制的研究

基本信息

批准号：
02650313
负责人：
OGAWARA Yoichi
金额：
$ 0.26万
依托单位：
Kyushu University
依托单位国家：
日本
项目类别：
Grant-in-Aid for General Scientific Research (C)
财政年份：
1990
资助国家：
日本
起止时间：
1990 至 1991
项目状态：
已结题

来源：
https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-02650313/
关键词：
Ocean Robot Manipulator Position Force Control Neural Network Model Inverse System Feedback-Error-Learning Control Learning Time Constant Learning Speed Hybrid Control System 位置制御小脳適応フィルタモデル高速学習方法

项目摘要

The manipulator for the ocean robot is required the ability to adapt its manipulation to various occasion and delicate movement. So some advanced control method are necessary to the control system of the manipulator, and it will be demanded the control of the position of the end effector and the force affected it. Recently, it is paid attention to the application of the neural network. Therefore, in this report, it is represented that the control method of the position of the end effector and the force affected it using the neural network model.(1) The feed forward controller which uses an adaptive neuro-filter employing the inverse system of dynamics of the manipulator is adopted as an internal model, and the feedback-error-learning scheme is used for the learning process. This control system enables precise operations compared to the conventional feedback control systems, and is adaptable to the change of the character of the plant.(2) By turning the time constants of the learning, the learning speed which is one of the difficult problem of the neural network is faster than the usual method.(3) By using present method, it is developed a practical hybrid control method of the position and the force of the manipulator.From these result, it would be established the foundation of the manipulation technique to the manipulator for the ocean robot.We are planning the study of the control method to reduce the energy of the manipulation, which considers the particular environment of the ocean robot.

海洋机器人的机械手要求具有适应各种场合的操作能力和精细的运动。因此，机械手的控制系统需要一些先进的控制方法，这就要求对机械手末端执行器的位置和力进行控制，近年来，神经网络的应用受到了广泛的关注。因此，本文提出了利用神经网络模型对机器人末端执行器的位置和作用力进行控制的方法。(1)前馈控制器采用自适应神经滤波器，采用逆系统的动力学的机械手作为一个内部模型，和反馈误差学习计划用于学习过程。与传统的反馈控制系统相比，该控制系统能够实现精确的操作，并且能够适应设备特性的变化。(2)通过改变学习的时间常数，使神经网络的学习速度比通常的方法快，这是神经网络的难点之一。(3)利用本文的方法，提出了一种实用的机械手位置和力的混合控制方法，为海洋机器人机械手的操作技术奠定了基础，并计划研究考虑海洋机器人特殊环境的降低操作能量的控制方法。