CORRECTION ALGORITHMS SYNTHESIS FOR THE MOTION PATH OF THE HEXAPOD ROBOT OUTPUT LINK BASED ON THE THEORY OF ARTIFICIAL NEURAL NETWORKS

Abstract

The article considers the problem of the correction algorithm synthesis of hexapod robot output link trajectory, based on the solution of the direct problem of kinematics. The structure of a robot control system, including units for direct and inverse kinematics problem. Shown an algorithm for the management of the direct problem solution of kinematics and the application of neural networks. As a method of training it is recommended a combined method of back propagation and annealing. Shown the comparative performance characteristics of neural networks with different structures. It was found that the use of corrective neural networks is the best option.