The State-Dynamic-Error-Based Switching Control under Quasi-ARX Neural Network Model

Abu Jami’in, Mohammad and Sutrisno, Imam and Hu, Jinglu (2015) The State-Dynamic-Error-Based Switching Control under Quasi-ARX Neural Network Model. The Twentieth International Symposium on Artificial Life and Robotics.

Text 2015 ISAROB.pdf Download (636kB)

Abstract

In our previous research, an error-based switching control has already used for controlling nonlinear system. However, the switching function is not work efficiently, because it is difficult to obtain more information from error vector to determine the stability of the control system. Hence unnecessary switching to linear controller will be longer and more often that causes the accuracy of the control system become poor. In this paper, a new switching rule based on Lyapunov stability theorem is proposed which is derived from the state dependent parameter estimation (SDPE). Not only error but also one up to p-th differential error will be available as the switching variable. Thus the proposed control method is able to keep the stability and improve the accuracy of the control system. A numerical simulation reveal that the proposed control gives satisfactory tracking and disturbances rejection performances. Experimental results demonstrate its effectiveness

Actions (login required)

View Item