Generation of movement trajectories in primates and robots

The design of brainlike machines for robot motor control ultimately requires a radical departure from the current robotics approach of implementing kinematics and internal space representation by means of software in synchronous, digital, and sequential ‘general purpose’ computers. In stead, neural net motor control of intelligent robots aims at hardware implementation of kinematics and internal space representation by means of self-organizing neural nets, which operate asynchronously, as fully parallel ‘special purpose’ analogue computers. On the basis of our present knowledge of the neural control of movements in the primate brain, recent concepts and models of artificial neural net motor control a rebeing presented.