Predictive Control of an Electromagnetic Suspension System via Modified Locally Linear Model Tree with Merging Ability

A predictive control algorithm based on modified locally linear model tree (LOLIMOT) with merging is implemented to control of an electromagnetic suspension system. A self-construction LOLIMOT is used to predict the response of the plant in a future time interval. This modified algorithm could improve the accuracy with reduced computational times and fewer rules which is important in real-time input optimization. An evolutionary programming (EP) is used to determine the optimized control variables for a finite future time interval. This method is applied to an electromagnetic suspension system (EMS) and simulation results show the effectiveness of the proposed predictive control strategy