Precise Positioning of Nonsmooth Dynamic Systems Using Fuzzy Wavelet Echo State Networks and Dynamic Surface Sliding Mode Control

Abstract

This paper presents a precise positioning robust hybrid intelligent control scheme based on the effective compensation of nonsmooth nonlinearities, such as friction, deadzone, and uncertainty in a dynamic system. A new adaptive fuzzy wavelet echo state network algorithm is proposed to improve performance in terms of approximating unknown uncertainties in conventional neural network algorithms. A strict feedback controller and adaptive laws for estimating the unknown friction and deadzone parameters were developed using the recursive steps of dynamic surface control and a sliding mode control based on Lyapunov stability theory. Lyapunov stability analysis confirmed the boundedness and convergence of the closed-loop system. The performance of the proposed control scheme was validated experimentally by applying it to the precise positional control of a robot manipulator.Index Terms-Deadzone and friction, dynamic surface sliding mode control, fuzzy wavelet echo state networks (FWESNs), robot manipulator.

A novel model of leaky integrator echo state network for time-series prediction

“…Thus, ESN can not only provide a simple and distinctive learning method, but also enables the learning results to obtain higher accuracy than that of RNN. ESN has recently attracted much attention and has been applied to some fields, for example, time-series prediction [5,[7][8][9][10][11][12][13][14], filtering or control [15][16][17][18][19], dynamic pattern recognition [20][21][22], system modeling [23], and so on. More and more scholars are committed to improving the performance of ESN, for instance, improving the training method of ESN [7][8][9][10][24][25][26][27][28][29], modifying the state update equation of reservoir [7,8], and so on.…”

Decentralized neural network control for guaranteed tracking error constraint of a robot manipulator

“…Among the four links of the Scorbot robot manipulator, only two links (upper arm = link1 and forearm=link2) were selected. From (1), the dynamics and parameters for two DOF (degree-of-freedom) links of the Scorbot robot manipulator are described in [16].…”

Fuzzy Approximation-Based Adaptive Backstepping Optimal Control for a Class of Nonlinear Discrete-Time Systems With Dead-Zone

“…To overcome the matching condition, several adaptive neural or fuzzy control approaches were designed in [32][33][34][35][36] for nonlinear systems in strict and non-strict feedback with nonlinear dead-zone. Two outstanding adaptive control techniques and applications using fuzzy echo state neural networks were developed in [37,38] for nonlinear dynamic system with dead-zone nonsmooth inputs. The control methods in [30][31][32][33][34][35][36][37][38] were designed for nonlinear continuous-time systems.…”

“…Two outstanding adaptive control techniques and applications using fuzzy echo state neural networks were developed in [37,38] for nonlinear dynamic system with dead-zone nonsmooth inputs. The control methods in [30][31][32][33][34][35][36][37][38] were designed for nonlinear continuous-time systems. For nonlinear discrete-time systems with dead-zone, Campos and Lewis in [39] developed a dead-zone compensation using the adaptive fuzzy systems.…”