Observer-based position control of uncertain mechanical systems with measurement uncertainty
Song Liang, Zaihua Wang
Abstract: This paper studies the position control of uncertain mechanical systems with measurement uncertainty. Firstly, an observer using measurement signals is designed to estimate the unknown disturbances under an improved assumption of the uncertainties. The improved assumption can give smaller error estimation with same observer gain compared with the conventional assumption. Next, the essence of the observer and the method for selecting the observer gain are introduced. For the case when measurement uncertainties are not considered, the observer is allowed to have a large gain. As for the current case, it is shown that the observer with a large gain will lead to a failure. Then, by using the observed values, two switching controllers using measurement signals are proposed to realize the position control. The proposed control is a generalization of the sliding mode control without measurement uncertainty to the case with measurement uncertainty. Meanwhile, the features of the velocity change of the controlled system are discussed. Finally, as an application of the control design, the position control of a dynamic positioning system is investigated in details. The motion of the controlled system exhibits obviously three phases with different velocity features during the system moves to the target position: approximate acceleration phase, approximate uniform velocity phase, and approximate deceleration phase, and the velocity of the approximate uniform velocity phase can be easily adjusted with the change of the control gains.
原文链接: https://onlinelibrary.wiley.com/doi/abs/10.1002/rnc.4571
