基于观测器和全局收缩映射的设定点控制  

Set-point Control Based on Observer and Contraction Mapping

作　　者：方浩[1] 吴建林[2] 范瑞霞[1] 

机构地区：[1]北京理工大学自动控制系模式识别与智能系统,北京100081 [2]北京邮电大学计算机科学与技术学院,北京100876

出　　处：《机械科学与技术》2007年第5期567-571,共5页Mechanical Science and Technology for Aerospace Engineering

摘　　要：针对具有参数不确定性并且仅有位置信息可测的机器人系统,本文首先研究了带有速度近似的■+有偏标称重力补偿的设定点控制方法。该方法利用线性观测器理论对不可测速度信号进行近似,基于此伪速度信号设计的设定点控制虽然可以保证闭环系统全局稳定,但由于无法对重力精确补偿,因此系统存在静差。为消除静差,本文借助全局收缩映射理论,每到闭环系统输出达到稳态时,通过重复修正线性观测器参数使静差逐步收敛到零,保证系统全局渐进稳定。利用所设计的控制器对并行机器人进行设定点控制,仿真实验表明,虽然不同的设计参数对控制效果有所影响,但都能够确保系统输出收敛到标称点。In order to control a robot manipulator system with parameter uncertainties, we propose a set-point control method that has P^^D plus estimated nominal gravity compensation with velocity approximation. The method applies a linear observer to approximate velocity signals and ensure the closed-loop system to be globally stable. However, the system cannot compensate for gravity accurately and has steady-state errors. To eliminate the steady-state errors, we resort to the global contraction mapping theory and make the closed-loop system iteratively learn the parameters of the linear observer whenever the system reaches its steady state. As a result, the steady-state errors converge to zero and the system achieves global stability. Finally we did experiments to take set-point control of a parallel manipulator. The experimental results show that the controller designed according to the proposed method can ensure the system＇s output to converge to the desired equilibrium point.

关 键 词：观测器 设定点控制 收缩映射 并联机器人 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]