Learning control systems using computer with application to remote manipulation

Remote manipulators and teleoperators have a unique potential to satisfy coming needs in space station observation, maintenance and repair of unmanned spacecraft and exploration of the moon and planets.1 Although teleoperators solve the problem of life support by allowing the operator to remain protected from the hazards of the task environment, they complicate the operator's control greatly. The associated control problems fall in one of three categories: feedback information loss, control information overload, and transmission time delays.