In this paper, a novel approach incorporating neural networks, genetic algorithms and fuzzy logic is used for modelling of a flexible manipulator. The neural network structure and weights are optimised through a genetic algorithm. Selection of a chromosome for further evolution is determined by using fuzzy logic. Fuzzy logic selects chromosomes on the basis of feasibility and priority of the constraint. Results show that modelling of a flexible manipulator, considering dominant modes and all of its features is successfully obtained. Fine local tuning of the GA is obtained by considering the local feature at the output of the neural network as well as dynamic selection by fuzzy logic.
展开▼
