Design of a fuzzy neural network control system
Yang Siyu,Wang Jinjin,Li Xiuxi
DOI: https://doi.org/10.3969/j.issn.1001-4160.2011.06.012
2011-01-01
Abstract:A self-structured control system is often integrated with a neural network and fuzzy theories to simplify the shaping of the control system. The actions of this control system,however,are sequential and obey to the sequence in which parameter adjustment follows by structure adjustment. In addition,most of the self-structured control system has only the adding node operation resulting in the generation of redundant rule nodes in the network.The presence of the redundant nodes requires for additional computation and slows down the responsive time of the system.For the solution to the above two shortcoming of present control system,this paper gives rise to a novel fuzzy network neural control system(FNCC).In such system, differing to the conventional self-structured control system,a removing node scheme is included to remove the redundant rule nodes in a network. This additional operation controls the size of network and eliminates unnecessary computation that can slow down the responsive time.For solutions to the inherent sequential process,authors find a way to parallize the actions of structure adjustment and parameter adjustment.For case studies,the proposed control system is applied to a nonlinear system to testify its validity and demonstrates its features:1)no need for pre knowing the application model;2)no restriction on structure and parameter design.