A Proposal of Autonomous Distributed 3 Layers System for Mortion Control

Abstract

In this paper, we propose a new control system that has three layers for a complicated robot. The first layer calls a Paraneuron Reinforcement Algorithm (PRA). This algorithm simulates a simple neural coupling of creature. The PRA has a self-organizing function to input signals. The second layer memorizes a state of PRA System using the Neural Network with 3 layers Back Propagation law. The third layer make a halfway target that advances to possibility of following the last target.
By using the method (3 Layers System) to motion control, a flexible control of complicated plant was made possible. But the problem is that PRA can't realize a highly precise performance.