Sliding mode controller with time-varying hyperplane

Abstract

A new adaptive robust sliding mode controller with a time-varying hyperplane is proposed far robot manipulators. Adaptive sliding mode controllers, especially those designed for robot manipulators, are known to have robust stability characteristics against inaccurate system parameter values, external disturbances, and relatively slow variation of system characteristics. They achieve this desirable characteristics through recursive identification of system parameters and on-line estimation of upper-bounds of system uncertainties. The robustness of such controllers depend heavily on their ability to keep the system characteristics on the sliding hyperplane. This is achieved via the use of high control gains. There exists a trade-off, however, of exciting the "chattering", a high frequeney oscillation, if control gains are set too high. The proposed controller alleviates this problem by providing high gains only where they are warranted and using low gains elsewhere to suppress chattering. Superior characteristics of the proposed controller is demonstrated first by simulation and then in an actual implementation to a two-link PUMA type robot.