Observer-Based Robust AILC for Robotic System Tracking Problem

基於觀測器架構之適應式迭代學習法則於機械臂追蹤控制之應用

陳介力(Chieh-Li Chen) ； 李凱笙(Kai-Sheng Li)

《中國機械工程學刊》 30卷6期 (2009/12) Pp. 483-491

https://doi.org/10.29979/JCSME.200912.0005

摘要

本文針對機械臂追蹤控制問題，提出一結合線性觀測器和動態適應疊代學習的控制架構。基於擴階的方式將積分器導入機械臂系統，因此可消弭由控制器設計而導致的高頻震顫現象，且不影響系統控制精度。控制器設計區分為回饋控制和學習控制兩個部分；在學習控制部份，其控制參數是由適應更新律沿著時間與學習次數軸所估測，可有效克服系統之不確定性和外擾。亦利用能量法來分析控制設計和閉迴路系統的學習收斂性。經由模擬的結果，可發現控制誤差將有效收斂，達成追蹤控制所賦予的任務。

並列摘要

A dynamic observer-based control scheme for trajectory tracking of robotic systems with parametric uncertainty is proposed in this paper. In the proposed control scheme, an integral action is augmented into the robotic system to avoid the chattering of iterative learning control without decreasing control accuracy. The learning control scheme uses a linear observer which is model independent. The proposed approach does not require any prior knowledge of parameter values on robot dynamics. The learning control algorithm includes a feedback controller and an adaptive learning term, to which a hybrid adaptive updating law is applied. The hybrid adaptive updating law updates the learning control parameter for each iterative operation and modifies the learning control parameter to compensate for closed-loop system uncertainty and to overcome external disturbances in the time domain. A Lyapunov-like energy function is utilized to analyze learning convergence and to derive the control algorithm. Simulation results are provided to illustrate the performance of the proposed dynamic learning control algorithm.

並列關鍵字

robotic system ； iterative learning control ； hybrid adaptive updating law ； linear observer