3D neural net for learning visuomotor-coordination of a robot arm

An extension of T. Kohonen's (Biol. Cybern., vol.43, p.59-69, 1982; vol.44, p.135-140, 1982) self-organizing mapping algorithm together with an error-correction rule of the Widrow-Hoff type is applied to develop an unsupervised learning scheme for the visuomotor coordination of a simulated robot arm. Using input signals from a pair of cameras, the closed robot arm system is able to reduce its positioning error to about 0.3% of the linear dimensions of its work space. This is achieved by choosing the connectivity of a 3D lattice between the units of the neural net.<<ETX>>