Designing a fully automated hierarchical fuzzy logic controllers using evolutionary algorithms

Hierarchical approaches and methodologies are commonly used as a promising technique for control system design and synthesis. This paper focuses an automatic way of evolving hierarchical fuzzy Logic controller which is a combination of cascaded fuzzy modules. The control algorithm works in two phases simultaneously: Firstly, evolutionary programming (EP) employed to identify the optimum hierarchical architecture and secondly, genetic algorithm (GA) employed to tune the fuzzy sub controllers (SC) involve in hierarchical controllers. The proposed control algorithms coupled both EP and GA optimizations. The fine tuning of the antecedent and consequent parameters of fuzzy control rules and their corresponding membership functions (MFs) of SCs encoded in the structure is accomplished using genetic algorithms (GAs). In this method, the total number of rules increases only linearly with the number of input variables. The automatically generated hierarchical fuzzy controller consists of a number of low-dimensional fuzzy systems in a hierarchical form. The proposed hierarchical control algorithm is evaluated using well known benchmark applications namely trailer-and-cab, a nonlinear motion control of nonholonomic robots. Simulation results are presented at different operating conditions and under various disturbances to verify the effectiveness of developed adaptive small size hierarchical controller.