Neural Active Disturbance Rejection Output Control of Multimotor Servomechanism

Abstract

In this brief, the problems of stability and tracking control for multimotor servomechanism with unmodeled dynamics are addressed by neural active disturbance rejection control. For realizing output feedback, an extended state observer based on high-order sliding mode (HOSM) differentiator is designed to estimate the unmeasured velocity. Moreover, HOSM differentiator is introduced to modify the traditional dynamic surface control method. The designed controller solves the contradiction between rapidness and overshoot, which comes from the traditional proportional-integral-derivative that deals with a large number of practical systems with unknown disturbances. In addition, unknown functions, including friction and disturbances, are approximated by Chebyshev neural networks (CNNs), in which adaptive laws are provided by Lyapunov method. Especially, steady state and transient performance of closed-loop system are maintained by performance function in theoretical analysis. Finally, extensive experimental results are provided to illustrate our proposed approach.