ABSTRACT
The presence of nonlinearities in robotic manipulators is the main factor in the selection of control algorithms for their real-time regulation and tracking. Control strategies based on a linearized model are usually grossly inadequate in meeting the precise performance requirements, and one must resort to techniques that are patently nonlinear. Adaptive control provides such nonlinear class of techniques that are particularly suitable for application in robotic manipulators. Adaptive control refers to a set of techniques that provide a structured or systematic approach for automatically and continuously adjusting the parameters of a controller in real time, either directly or indirectly, in order to achieve or to maintain a certain desired level of system performance when the parameters of the plant dynamic model are unknown and/or change in time. An adaptive controller is one that can modify the system’s closed-loop behaviour in response to changes in the dynamics of the manipulator due to changes in the configuration, external constraints and disturbances. One approach to adaptive control is the two-step or indirect adaptation procedure. In the first step, a parameter identifier is used to estimate all the parameters that characterize the system dynamics. Once these are known, a classical algorithm, such as pole placement, optimal control, feedback linearization with or without dynamic inversion or sliding mode control, is used to synthesize the control law. Conventional sliding mode controllers may result in an appreciable amount of chattering. One method of reducing the chattering is by continuous adaptation of the parameters. Continuous parameter adaptation schemes are one of the most popular approaches to adaptation, which directly result in chattering-free control action.
