Adaptive ANN-based control for constrained robot manipulators

Abstract

An Artificial Neural Network (ANN)-based approach is proposed in this paper for the motion and force control of constrained robot manipulators. The dynamic model of constrained manipulator is modified to contain two sets of state variables, where one describes the constrained motion and the other describes the reduced, unconstrained one. An application related to a disk cutter robot is considered. The design objective is that under a prescribed radius of the disk and a desired force to be applied by the effector of the manipulator, an ANN-based control system is to be designed to develop the requested torques on the manipulator actuators. The suggested controller is simple and can be implemented easily. The main feature of employing the ANN here is that we obtained a force sensor-less control system and a good adaptation with environmental unmodelled dynamics. Simulation results are presented to validate the proposed approach.