A Design Method of Model Error Compensator using Meta-Heuristics and LMIs

Abstract

Control systems achieves desired performance with model based controller if the dynamical model of the actual plant is given with sufficient accuracy. However, if there is difference between actual plant dynamics and its model dynamics, the model based controller does not work well and does not achieve intended desired performance. To overcome this problem, a model error compensator (MEC) is proposed in our previous study. Attaching the compensator for the model error to the actual plant, the output trajectory of the actual plant is made close to that of its dynamical model. Then, the difference of dynamics on appearance from controller to observed output can be smaller and performance degradation caused by the modeling error is reduced by using MEC. Previous studies proposed the design method of filter parameters in MEC based on frequency domain. However, we have to design the gain of MEC appropriately for many type of systems such as MIMO system, non-minimum phase system and so on. In addition, it is necessary to design controller based on time domain method. In this paper, we propose a time domain design method of the filter parameters in MEC using particle swarm optimization (PSO). First, we show an analysis method about H ∞ performance of MEC, which is used as evaluation function in PSO. Then, we simulate proposed method for MIMO system. Using proposed method, it is expected that we can design MEC more efficiently and robust.