Multi-Objective Controller Design for Stochastic Model Reference Systems via Sliding Mode Control Concept and LMI Approach

In this paper, we design a multi-objective controller ( ) u t to achieve the following three objectives simultaneously. They are pole placement, H∞ norm constraint and individual error state variance constraint for stochastic model reference systems. By using the invariance property of sliding mode control, the matched model reference input and the plant error term will disappear on the sliding mode of the error system. Combining the upper bound covariance control theory, pole placement skill, H∞ norm control theory and linear matrix inequality approach, the controller ( ) u t is derived in which the control feedback gain matrix G is synthesized for achieving the above multiple objectives.