I. Introduction
This paper describes a self-organized method for control of robots with redundant degrees of freedom based on a neural model. Unlike prior work in the area of robotics [[1]–[2]],this model draws its inspiration from the analysis of how visual, spatial and motor representations are formed and combined for the control of goal-oriented behaviors in humans [3]–[11]. We argue that it is not coincidental but a deliberate evolutionary adaptation of nature to have redundancy in the design of motor systems wherein goals can be reached using multiple motor means. This phenomenon known as motor equivalence [3] addresses how animals and humans can correctly choose among the alternative means that are available to perform a given goal in different situations.