旋轉式倒立單擺之系統辨識及順滑模控制器實現

Abstract

本篇論文中，我們研究如何穩健控制旋轉式倒立單擺的手臂和單擺問題。首先導出旋轉式倒立單擺的動態模型並且透過最小平方近似法去判別系統的未知參數，本論文中提出了兩種控制法運用在旋轉式倒立單擺的平衡控制上，狀態回授控制與順滑模控制。狀態回授控制被設計在旋轉式倒立單擺的線性模型上，而順滑模控制法不僅使用在旋轉式倒立單擺的線性模型而且也被運用非線性模型上。此外順滑模控制理論可以抵抗系統結構的不確定項並且做有效的控制。模擬以及實驗的結果顯示順滑模控制法比狀態回授控制法有更好的強健性並做有效的控制。

This thesis studies the robust problems of the rotary arm and pendulum control of the rotary inverted pendulum. Its dynamic model is derived by Lagrange’s Method and identify the unknown parameters via Least Squares Method. This thesis proposes balance control for the rotary inverted pendulum base on the state-feedback control and sliding mode control. The state-feedback control is used to design the linear model of the rotary inverted pendulum. However, the sliding mode control not only is used to design the linear model but also is applied in the nonlinear model. In addition, the sliding mode control deals with the uncertainties which is produced from the system structure and effectively controls the rotary inverted pendulum. Finally, simulation and experiment result show that the sliding mode has excellent robustness than state-feedback control.