DESIGN OF NONLINEAR COMPENSATOR FOR POSITIONING CONTROL IN A MASS-STORAGE SYSTEM - (A NEURAL-NETWORK-BASED CONTROLLER)

Mass storage systems (MSSs) with automatic mediahandling mechanisms have become widely used in the field of information processing, and faster media handling is desired. Although there are various handling mechanisms, in all cases, disturbances such as solid friction and inertia variations are major obstacles to faster handling. This paper investigates a neural-network-based controller that compensates for these disturbances. A direct-type neural controller is proposed. The capability and characteristics of the controller are numerically and experimentally investigated by testing the positioning control of a rotary storehouse mechanism in an MSS. Experimental results, compared with those obtained using a conventional disturbance observer, indicate that the proposed neural-network-based controller can compensate effectively for nonlinear disturbances in the mechanisms.