Multilayer neural-net robot controller with guaranteed tracking performance

TLDR

The nonlinear nature of neural networks, reconstruction inaccuracies, and robot disturbances render standard backpropagation tuning insufficient for closed‑loop dynamic control. The study develops a multilayer neural‑net controller for a general serial‑link rigid robot arm and introduces novel online weight‑tuning algorithms that guarantee bounded tracking errors and bounded NN weights. The controller is structured via a filtered‑error/passivity approach, incorporates second‑order forward‑propagated correction terms, and introduces passive, dissipative, and robust neural‑net properties. The controller requires no offline learning, initializes weights easily, and achieves arbitrarily small tracking error bounds by increasing a feedback gain.

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.