Abstract: As the increase of input dimensions in fuzzy system, fuzzy rules and identification parameters will increase according to exponential growth. To solve the problem, the use of hierarchical fuzzy system is an excellent solution. However, the identification parameters in hierarchical fuzzy system haven’t clear physical meaning, it is hard to be designed and explained reasonably. This paper proposes a new hierarchical fuzzy system by using centralized TSK fuzzy model. The new fuzzy system retains the structure advantages of hierarchical fuzzy system. It not only includes less fuzzy rules and parameters, but also can give a reasonable explanation for the parameters of the system. According to the example simulation, the new fuzzy system has good approximation performance and more simple structure.

Key words: hierarchical fuzzy system, TSK fuzzy model, explanation, fuzzy rule, identification parameter

摘要： 模糊系统随着输入维数的增加，其中模糊规则和辨识参数的数量将按指数级增长，针对这一问题，采用分层模糊系统是一种很好的解决方法，但分层模糊系统中各层的辨识变量没有明确的物理含义，无法进行合理的模糊化设计和解释。基于一种分层模糊系统，引用中心性TSK模糊系统思想，从而构造了一种新型的模糊系统。这种新型模糊系统保留了分层模糊系统的结构优势，极大地减少了模糊系统的模糊规则数量和辨识参数数量，又能对用到的内部参数进行很好的解释。并通过实例仿真表明基于中心型TSK模糊模型的分层模糊系统具有较好的逼近性能和更简单的结构。

关键词: 分层模糊系统, TSK模糊模型, 解释性, 模糊规则, 辨识参数