Multilayer neural-net robot controller with guaranteed tracking performance

Abstract

A multilayer neural-net (NN) controller for a general serial-link rigid robot arm is developed. The structure of the NN controller is derived using a filtered error/passivity approach. No off-line learning phase is needed for the proposed NN controller and the weights are easily initialized. The nonlinear nature of the NN, plus NN functional reconstruction inaccuracies and robot disturbances, mean that the standard delta rule using backpropagation tuning does not suffice for closed-loop dynamic control. Novel online weight tuning algorithms, including correction terms to the delta rule plus an added robust signal, guarantee bounded tracking errors as well as bounded NN weights. Specific bounds are determined, and the tracking error bound can be made arbitrarily small by increasing a certain feedback gain. The correction terms involve a second-order forward-propagated wave in the backpropagation network. New NN properties including the notions of a passive NN, a dissipative NN, and a robust NN are introduced.

Adaptive control of the SCORBOT-ER 4PC manipulator

“…The adaptive law (19) guarantees that estimates of parameters will be bounded without persistency of the excitation condition. According to a standard Lyapunov theorem extension [10], both s and p % are uniformly ultimately bounded.…”

Adaptive friction compensation of servo mechanisms

“…For ease of analysis and controller design later, we present only an LIP NN-based friction model: nonlinear multilayer NN-based friction models can also be investigated following the method of treating multilayer NN in [26,36]. Accordingly, we have the following LIP friction model in general form:…”

“…In this paper, only Gaussian RBF neural networks are discussed. If fact, other neural networks can also be used without any difficulty, and include other RBF neural networks, high order neural networks, and multilayer neural networks [26,36].…”

Distributed Adaptive Control for Networked Multi-Robot Systems

“…The Lyapunov analysis technique approach of (Lewis et al, 1999;Lewis et al, 1996) is used, though there are some complications arising from the fact that ( ) The following Fact gives two standard results used in neural adaptive control (Lewis et al, 1999) Fact 1. Let the nonlinearities () fx in (3) …”