MULTI-OBJECTIVE RECONFIGURABLE OUTPUT FEEDBACK CONTROLLER FOR MIMO SYSTEM USING MOEAS |
| |
Authors: | S. Sutha S. Lavanya T. Thyagarajan |
| |
Affiliation: | 1. Department of Instrumentation Engineering , MIT Campus, Anna University , Chennai , Tamil Nadu , India sutha_mani@hotmail.com;3. Department of Instrumentation Engineering , MIT Campus, Anna University , Chennai , Tamil Nadu , India |
| |
Abstract: | This article deals with the design of a parametric eigenstructure assignment (PEA)-based multi-objective reconfigurable output feedback controller for a multiple-input multiple-output (MIMO) system against partial actuator failure. The significance of the PEA technique is that it provides more design degrees of freedom to obtain multi-objective functions, namely robust stability and improved transient response. Robust stability and transient response are ensured by a minimum condition number of the eigenvector and a minimum norm of controller gain. These two objectives are conflicting in nature, and, hence, the main aim is to minimize both objectives simultaneously by means of a multi-objective evolutionary algorithm (MOEA). In this study, a multi-objective optimization problem is formulated and is solved by non-dominated sorting genetic algorithm-II (NSGA-II) and strength Pareto evolutionary algorithm-II (SPEA-II). Four indicators are used to evaluate each algorithm performance: spacing, error ratio, generational distance, and hypervolume. Then a reconfigurable controller is designed using PEA against actuator failure, which guarantees closed-loop stability. The effectiveness of the proposed controllers is demonstrated by implementing them in an interacting three-tank benchmark system. |
| |
Keywords: | MIMO MOEA NSGA-II PEA PI SPEA-II |
|
|