标题: 以CMAC类神经网路控制器实现机器人控制
Realizing Robot Control by CMAC Neural Network Controller
作者: 林桓宇
Huan-Yu Lin
陈福川
Fu-Chuang Chen
电控工程研究所
关键字: 小脑模组关节控制器;类神经网路;PID控制器;机械臂;两轴机械臂;CMAC;PID controller;robotics;CMAC control system;2-link robot
公开日期: 1999
摘要: 本论文的目的在研究CMAC类神经网路控制系统的控制行为并分析系统的稳定性,然后和PID控制器作比较。为了知道受控系统的输出/输入特性,在这里使用Runge-Kutta法。我们企图以CMAC控制系统来扮演机械臂的pd-plus-gravity控制效果。不过CMAC控制系统不需要关于机械臂的任何资讯,而且可以应付很大的负载变化。当输入向量缓慢的移动的时候,CMAC控制器会产生庞大的积分动作。但是当输入向量快速变化的时候,CMAC控制器也能有效的产生"遗忘"效应。论文中也证明CMAC控制系统能够在目标位置收敛。然后描述回授讯号杂讯对输出误差的影响,并且说明控制器参数改变时杂讯影响的能力。此外我们也将CMAC控制系统应用在两轴机械臂上,以两个独立的CMAC控制系统对各轴独立控制。最后比较CMAC控制系统及PID控制器的控制效果。
The object of this paper is to study the behaviors of CMAC control system and to analyze the stability of the system, and then to compare with PID controller. To find out the characteristic between output and input, the Runge-Kutta method is used. The CMAC controller is proposed to emulate the pd-plus-gravity control for robotics. However, the CMAC requires no information about the robot, and can deal with large variations in load. The CMAC produces enormous integration action when the input vector moves slowly in the space, but it can also forget efficiently when the input vector moves fast in the space. It is shown that the CMAC control system can converge into the target position. The effect of feedback signal noise has been described, the effect of noise when controller parameters are changed is also illustrated. We also use CMAC control system for 2-link robot control. Each link of the robot has an independent CMAC controller. The controller performance between CMAC control system and PID controller has been compared.
英文摘要.....................................................ii
志谢........................................................iii
目录.........................................................iv
表目录.......................................................vi
图目录......................................................vii
第一章 绪论..................................................1
1.1 研究动机、背景与目的.....................................1
1.2 论文内容介绍.............................................2
第二章 控制对象及PID控制结果.................................4
2.1 四阶Runge-Kutta法求解常微分方程式........................4
2.2 单轴机械臂...............................................6
2.2.1 PID控制结果............................................8
2.3 二轴机械臂..............................................10
2.3.1 PID控制结果...........................................14
第三章 CMAC控制器...........................................18
3.1 CMAC控制系统............................................18
3.2 CMAC之计算架构..........................................19
3.3 CMAC工作原理............................................23
3.4 记忆体配置及存取........................................24
3.5 CMAC与Traditional Adaptive Control System之比较.........25
第四章 单轴机械臂控制.......................................29
4.1 CMAC控制结果分析........................................29
4.2 稳定性分析..............................................34
4.2.1 控制对象线性化模型....................................34
4.2.2 只使用CMAC控制器......................................37
4.2.3 只使用PD控制器........................................38
4.2.4 CMAC控制系统的稳定条件................................39
4.2.5 不同操作点的稳定情形..................................40
4.2.6 PD控制器控制参数D的影响...............................40
4.2.7 Sampling Rate的影响...................................42
4.3 Tracking控制结果........................................44
4.4 系统杂讯................................................46
4.4.1 杂讯灵敏度............................................46
4.4.2 PD控制器控制参数KP的影响..............................48
4.4.3 PD控制器控制参数KD的影响..............................49
4.4.4 取样间隔(Sampling Interval)的影响.....................50
第五章 二轴机械臂控制.......................................52
5.1 控制系统架构............................................52
5.2 CMAC之计算架构..........................................52
5.3 CMAC控制结果............................................54
第六章 结论及未来展望.......................................57
参考文献.....................................................58
URI: http://140.113.39.130/cdrfb3/record/nctu/#NT880591028
http://hdl.handle.net/11536/66259
显示于类别:Thesis