Learning control system for manipulators with the ability to use already acquired knowledge in other problem

Abstract

The conventional learning control system doesn't have the capability of generalization, because input data for one trajectory, acquired by learning, cannot be applied to other trajectories. In this paper, we propose the new learning control system with the neural network. The neural network can acquire and memorize the dynamics of the system based on the data obtained from conventional learning process. By using this system, we can obtain the well-approximated input data for any trajectory.