Robot control

Abstract

This review of the state-of-the-art in robot control has highlighted the great difficulties in synthesising a satisfactory controller. It has been established that the manipulator is a nonlinear, multivariable system with time-varying and/or inaccurately-known parameters. However, most of the controllers proposed to date make simplifying assumptions which treat the manipulator as a linear and/or single-variable system. Such assumptions allow reasonably satisfactory control at low speeds but cause gross errors in dynamic control at high speeds. Various avenues of current research around the world are aimed at developing a proper nonlinear and multivariable controller which is tolerant towards parameter inaccuracies and variations. Presently, results are being reported for simulated controllers and simulated robot models.