ROBUST NONLINEAR CONTROL AND ESTIMATION OF A PRRR ROBOT SYSTEM

Abstract

In this paper, a newly proposed implementation of an unscented Kalman (UK)-based smooth variable structure filter (UK-SVSF) is introduced. The method is combined with a sliding mode controller (SMC) to compensate for modelling uncertainties. The robustness and tracking accuracy of the proposed controller and estimation strategy are demonstrated on a four-degree-of-freedom robotic system with one prismatic and three rotary joints. The effectiveness of the proposed combination is proven through comparisons with three different non-linear estimation algorithms: the standard unscented Kalman filter, SVSF, and a previously published UK-SVSF. The robot's trajectory following accuracy and efficiency are used as the performance parameters to study and compare the different strategies. Modelling uncertainties are added to the system to provide a more thorough evaluation of the robustness of the different non-linear control and estimation strategies.