Design of a Friction Self-Compensation Control Law for Multi-Link Robot Using Adaptive Sliding Mode Control

点击翻译按钮获取中文摘要

(1) In order to apply a sliding mode controller to robot control, a modified sliding mode controller isﾟCdesigned to follow a reference input. The modified controller is applied to robot in experiment. In robotﾟCcontrol, to track input exactly, the input model is included as an inner model in controller. An adaptiveﾟCservo controller is proposed to identify the model and to include the model adaptively. To cancel frictionﾟCdisturbances, an adaptive servo controller is given by a similar principle. Also to suppress disturbances, aﾟCstrongly stable self-tuning controller is derived from two-compensator scheme.(2) It is shown from experimental results that chattering, of which reduction is the most significantﾟCsubject in designing sliding mode controller, is caused by noise in estimating robot angle velocity and isﾟCreduced by improving the estimation using an observer. Designing the observer is shown to be dual toﾟCdesigning a sliding hyper plane in sliding mode controller. Using the duality, a new design scheme forﾟCobserver is given and the observer designed by the new method is shown to be effective in an experiment.(3) A large gain of switching law in sliding mode controller is another cause of chattering but the gainﾟCmust be larger than the upper bound of disturbances. To design a smaller gain an adaptive sliding modeﾟCcontroller to identify the bound and adaptively adjust the gain has been proposed by other authors. In thisﾟCproject, it is proposed to use P1-type adaptive law in the adaptive sliding mode controller and it is shownﾟCthat the proposed method reduces the chattering and the closed-loop stability is proved mathematically.ﾟCAlso proposed to use the PI-type adaptive law in an adaptive observer used in the adaptive sliding modeﾟCcontroller and a reduction of chattering is shown.