ADAPTIVE FUZZY-WAVELET NEURAL NETWORKS-BASED REAL-TIME MODEL GENERATION FOR INCREASING TRACKING PRECISION OF MULTIVARIABLE SERVO ACTUATOR

Sadeq Yaqubi, Mahdi Homaeinezhad, and Mohammad R. Homaeinezhad

View Full Paper

Keywords

Disturbance observer, fuzzywavelet neural network, discrete sliding mode control, uncertain systems, simultaneous torque and position control

Abstract

This paper proposes a new method for simultaneous torque and position control of uncertain servomechanisms under the effect of unknown external disturbance signals. Realistically, it cannot be expected that mathematical models of gear-drive systems express its dynamical behaviour accurately. Multiple control algorithms are proposed to overcome the aforementioned issue. First control algorithm is constructed considering worst-case scenario to ensure closed-loop stability in all possible configurations of modelling and disturbance uncertainty. However, this control mode is potentially conservative. To address this issue, the second control algorithm constitutes of fuzzy-wavelet neural network (FWNN) disturbance observer to pinpoint uncertain dynamical effects. Based on the FWNN estimation of predicted states, a separate sliding controller is designed in this mode.

Important Links:



Go Back