Adaptive sliding mode control for robotic systems using multiple parameter models

In the face of large scale parametric uncertainties, the existing single model (SM)-based sliding mode control (SMC) demands high controller gains to achieve good transient tracking performance. The main practical problem of having high-gains based design is that it amplifies the input and output disturbance as well as excites hidden unmodeled dynamics causing high-frequency switching and infinitely fast control action. To deal with the problem associated with high-frequency control chattering phenomenon, we introduce multiple parameters model based Lyapunov switching strategy to reduce the level of parametric uncertainty so as to reduce the control gains. This method extends the SM-based SMC approach by allowing the controllers to be reset among the finite set of the candidate controllers. The key idea is to distribute the compact set of the unknown parameter into a finite number of smaller compact subsets. Then, we design a family of candidate controllers corresponding to each of the smaller compact parameter subsets. We then use a Lyapunovbased switching-logic to identify a controller from a family of the candidate SMC controllers. Finally, implementation results on a 2-DOF robotic system is shown to demonstrate the effectiveness of the proposed method.