Design of robust fuzzy-logic control systems by multi-objective evolutionary methods with hardware in the loop

Stewart, Paul and Stone, D. A. and Fleming, P. J. (2004) Design of robust fuzzy-logic control systems by multi-objective evolutionary methods with hardware in the loop. Engineering Applications of Artificial Intelligence, 17 (3). pp. 275-284. ISSN 0952-1976

Full text not available from this repository.

Official URL: http://www.elsevier.com/wps/find/journaldescriptio...

Abstract

Evolutionary development of a fuzzy logic controller is described and is evaluated in the context of hardware in the loop. It had been found previously that a robust speed controller could be designed for a DC motor motion control platform via off-line fuzzy logic controller design. However to achieve the desired performance, the controller required manual tuning on-line. This paper investigates the automatic design of a fuzzy logic controller directly on to hardware. An optimiser which modifies the fuzzy membership functions, rulebase and defuzzification algorithms is considered. A multi-objective evolutionary algorithm is applied to the task of controller development, while an objective function ranks the system response to find the Pareto-optimal set of controllers. Disturbances are introduced during each evaluation at run-time in order to produce robust performance. The performance of the controller is compared experimentally with the fuzzy logic controller which has been designed off-line, and a standard PID controller which has been tuned online. The on-line optimised fuzzycontroller is shown to be robust, possessing excellent steady-state and dynamic characteristics, demonstrating the performance possibilities of this type of approach to controller design.

Item Type:Article
Additional Information:Evolutionary development of a fuzzy logic controller is described and is evaluated in the context of hardware in the loop. It had been found previously that a robust speed controller could be designed for a DC motor motion control platform via off-line fuzzy logic controller design. However to achieve the desired performance, the controller required manual tuning on-line. This paper investigates the automatic design of a fuzzy logic controller directly on to hardware. An optimiser which modifies the fuzzy membership functions, rulebase and defuzzification algorithms is considered. A multi-objective evolutionary algorithm is applied to the task of controller development, while an objective function ranks the system response to find the Pareto-optimal set of controllers. Disturbances are introduced during each evaluation at run-time in order to produce robust performance. The performance of the controller is compared experimentally with the fuzzy logic controller which has been designed off-line, and a standard PID controller which has been tuned online. The on-line optimised fuzzycontroller is shown to be robust, possessing excellent steady-state and dynamic characteristics, demonstrating the performance possibilities of this type of approach to controller design.
Keywords:Fuzzy Systems, Genetic Algorithms, Multiobjective Optimisation, Control Systems, Hardware in the Loop
Subjects:H Engineering > H620 Electrical Engineering
H Engineering > H660 Control Systems
G Mathematical and Computer Sciences > G520 Systems Design Methodologies
H Engineering > H100 General Engineering
G Mathematical and Computer Sciences > G700 Artificial Intelligence
H Engineering > H360 Electromechanical Engineering
Divisions:College of Science > School of Engineering
ID Code:2190
Deposited By:INVALID USER
Deposited On:01 Mar 2010 23:01
Last Modified:11 Dec 2012 15:08

Repository Staff Only: item control page