A Neural Network Compensation Technique for an Inertia Estimation Error of a Time-Delayed Controller for a Robot Manipulator

摘要

In this paper, a neural network is added to compensate for the deviation error of an estimated inertia matrix of the time-delayed controller for a robot manipulator. The time-delayed control (TDC) method is known as a simple and practical control method for controlling robot manipulators. The previously sampled information is used to cancel uncertainties for the current control using a time- delay. One of the problems of TDC is the constant inertia selected for simplicity and how to deal with the error of the inertia model estimation. In this paper, a neural network is used to compensate for the deviated inertia error. Simulation studies of position tracking control performances of a three link rotary robot are presented.