Trajectory Control of Manipulator Based on Learning Control with the Ability to Use Already Acquired Knowledge in Other Problem

Abstract

The trajectory control of the manipulator based on the learning control can not generalize its target trajectory. Because the input data for a certain trajectory which has been acquired through the learning operaion can not be applied to the other trajectory. In this paper, we propose new learning control system with a Neural network. The Neural network can acquire and memorize the dynamics of the system based on the data obtained from the conventional learning processes. With the system, we can obtain well-approximated input data for any trajectory at first trial.