Cooperative deterministic learning control for a group of homogeneous nonlinear uncertain robot manipulators

Abstract

This paper addresses the learning control problem for a group of robot manipulators with homogeneous nonlinear uncertain dynamics, where all the robots have an identical system structure but the reference signals to be tracked differ. The control objective is twofold: to track on reference trajectories and to learn/identify uncertain dynamics. For this purpose, deterministic learning theory is combined with consensus theory to find a common neural network (NN) approximation of the nonlinear uncertain dynamics for a multi-robot system. Specifically, we first present a control scheme called cooperative deterministic learning using adaptive NNs to enable the robotic agents to track their respective reference trajectories on one hand and to exchange their estimated NN weights online through networked communication on the other. As a result, a consensus about one common NN approximation for the nonlinear uncertain dynamics is achieved for all the agents. Thus, the trained distributed NNs have a better generalization capability than those obtained by existing techniques. By virtue of the convergence of partial NN weights to their ideal values under the proposed scheme, the cooperatively learned knowledge can be stored/represented by NNs with constant/converged weights, so that it can be used to improve the tracking control performance without re-adaptation. Numerical simulations of a team of two-degree-of-freedom robot manipulators were conducted to demonstrate the effectiveness of the proposed approach.

Event-triggered neural network control of autonomous surface vehicles over wireless network

“…Theorem 1. Considering the ASV kinetics being (6), the controller formed by (8), (11), (12), (16), and the event-triggered rules being (20)- (22), the resulting closed-loop system cascaded by (14) and 17is input-to-state stable. Moreover, the errors e andê are uniformly ultimately bounded.…”

“…To solve the effects of uncertainties, a quantity of control approaches have been presented from papers . In [3][4][5][6][7][8][9][10][11], neural networks (NNs) were used to approximate the uncertainties. In [12][13][14][15], nonlinear disturbance observers were employed to approximate the uncertainties caused by unknown external disturbances.…”

“…Compared with previous studies in [2-11, 15-17, 19, 22, 25, 26], the contributions of this paper are summarized below. Firstly, compared with the NN control approaches proposed in [3][4][5][6][7][8][9]11], where the systems work with fixed sampling and controlling period, the proposed control method updates the vehicle states and control inputs aperiodically. Network transmissions occur only at triggering instants.…”

Snapshot boosting: a fast ensemble framework for deep neural networks

“…Deep neural network (DNN) has been widely used and achieved great performance on many practical applications [1][2][3]. Usually, we train a single DNN until convergence, but the generalization ability of a single model is limited [4].…”

Sliding-mode-disturbance-observer-based adaptive neural control of uncertain discrete-time systems

“…The above-mentioned problems can not only degrade the performance of the controlled object, but also result in system instability. However, the issues of system uncertainties, input saturation, and external disturbances have seldom been considered in combination in the existing research on the control of continuous-time and discrete-time systems [1][2][3][4][5]. In recent years, the utilization of digital computers and samplers in actual control plants has made the controller design based on discrete-time system representation more reasonable than that based on continuoustime representation.…”