ADAPTIVE ROBUST BACKSTEPPING SLIDING MODE CONTROL OF A DE-ICING INDUSTRIAL ROBOT MANIPULATOR USING NEURAL NETWORK WITH DEAD ZONE

Van T. La, Shoudao Huang, Thi D. Tran, and Duc H. Vu

Keywords

Adaptive-robust-neural-network, industrial robot manipulator, unknown dead zone

Abstract

In this study, a combination of backstepping technique, neural net- work (NN), adaptive sliding mode control (ASMC) and adaptive proportional integral (API) control with dead zone is introduced to the industrial robot manipulator (IRM). In this controller, the input of NNs update rules, ASMC and API control are the error tracking filter and intermediate function. The unknown dynamics of IRM are approximated by a three-layer NN. The stability and sustainability under various environmental conditions of this method are guaran- teed by a sliding mode controller. The API is added as the third controller to optimize tracking performance, fast response and the overshoot that can be achieved. Simulation and experimental results show a high performance of this control method when compared to adaptive-fuzzy, fuzzy-wavelet-neural-network, robust-neural-fuzzy- network and proportional–integral–derivative. The proposed con- troller not only shows flexibility during parameter adjustment but also demonstrates the ability to compensate for approximate er- rors, thereby concluding that the suggested control is in accordance for adaptive-robust-neural-network controller and can be used as supplement and replacement of traditional backstepping control.

Important Links:



Go Back