Adaptive P‐type iterative learning radial basis function control for robot manipulators with unknown varying disturbances and unknown input dead zone

Abstract

This article proposes an adaptive iterative learning radial basis function (RBF) scheme to solve the trajectory-tracking problem for perturbed robot manipulators with unknown iteration varying disturbances and unknown dead-zone input. It is well known that the presence of the dead zone in actuators and mechatronics devices gives rise to extra difficulty due to the presence of singularity in the input channels. Hence, it is interesting to take this problem into account when synthesizing a controller. This synthesis is made here. In addition, the control design is very simple in the sense that we use, only, the proportional gain. Therefore, the considerable amount noise caused by the sensors for velocity measurements of robot manipulators is avoided. Another advantage of this work is that the unknown disturbances are assumed to be time varying and also varying from iteration to iteration. Thus, the RBF neural network is used to approximate these unknown nonlinear functions. Using the Lyapunov theory, the analysis of the stability of the closed-loop system is guaranteed when the iteration number tends to infinity. Finally, simulation results on the PUMA 560 arm are provided to illustrate the effectiveness of the proposed method. In order to evaluate the performance of our controller, a comparison of our results with other method is also given. K E Y W O R D SIterative learning control, Radial basis function, Dead-zone, Adaptive, Robot manipulators, Lyapunov theory 1 Indeed, input uncertainties such as dead-zone, saturation, backlash, and hysteresis are kind of nonsmooth and nonaffine in input factor widely existing in actuators and sensors. In fact, in many practical applications, the presence of Int J Robust Nonlinear Control. 2020;30:4075-4094.wileyonlinelibrary.com/journal/rnc

Iterative learning tracking control for second‐order nonlinear hyperbolic impulsive partial differential systems

“…It was of great significance for nonlinear, complex, intricate modelling and perfect trajectory control problems. ILC was widely used in many fields such as robotics [21]- [23], rehabilitation medicine [24], chemical batch processing technology [25], and tracking control of piezoelectric systems [26]. In addition to practical applications, ILC has also been widely used in switching systems [27,28], time-delay systems [29], and stochastic systems [30].…”

Adaptive boundary observer network design for the consensus on the estimation of a class of parabolic partial differential equation systems

“…Recent years have witnessed a rapidly growing interest in control system design using neural networks (NNs) together with the adaptive technique. As one of the most popular intelligent computation approaches, NNs have the inherent learning ability and can approximate nonlinear continuous functions to arbitrary accuracy 14‐16 . For the uncertain systems, radial basis function neural networks (RBFNNs) have been efficiently employed to approximate the ideal controller 17‐19 .…”

Distributed iterative learning temperature control for large‐scale buildings

“…However, none of the above studies considers such inherent characteristics of repeatability for control. In consideration of repeatability for control, it is reasonable to explore input and output information of previous iterations to gradually improve the control performance 12 . This idea motivates the proposed iterative learning control (ILC), which is a data‐driven intelligent control strategy 13‐15 .…”

“…In consideration of repeatability for control, it is reasonable to explore input and output information of previous iterations to gradually improve the control performance. 12 This idea motivates the proposed iterative learning control (ILC), which is a data-driven intelligent control strategy. [13][14][15] Different from other control methods, ILC uses input and output information from previous iterations to generate input for the current iteration.…”