A genetic approach to design a fuzzy controller for a nonlinear DC motor drive is described. The genetic algorithm optimises the input/output membership functions and gains, by experimenting new solutions on-line, i.e. directly on the physical hardware. The multi-objective fitness function is defined to achieve a good trade-off of performance across different operating conditions including a variable load.
Published in:
Electronics Letters
(Volume:39
,
Issue:
5
)
Date of Publication: 6 Mar 2003