Self-organized learning and its implementation of robot movements

Abstract

The self-organizing map algorithm using an artificial neural network originally developed by Kohonen and extended and modified later provides a distributed and autonomous learning procedure in engineering modeling of the human sensory-motor mapping mechanism. Its extension and adaptation to a control problem of a robot manipulator has been intensively discussed in past years. In this article the application of the self-organizing map algorithm to the generation of a visuo-motor map is focused on. A task-oriented inverse kinematic solution to a redundant manipulator is formed and real-time implementation of the map on a mechanical manipulator is performed.