用CPN网络实现的自学习模糊控制器

Self-learning Fuzzy Controllers Based On CPN Network

采用计数传播网络 (CPN)作为框架 ,结合一种简单的模糊控制算法 (SFCA) ,实现模糊神经元控制 .这种方法有能力自组织、自学习控制过程所需的控制知识 ,控制规则库初始为空 ,逐渐地被自构造 ,来满足预先设定的性能要求 .方法能实现任意维控制规则库的自学习 .仿真研究体现了系统的适应能力。

A fuzzy neural confrol was realized by using counter propagation network(CPN) as frame work with a simple fuzzy control algorithm(SFCA).This way may has self organizing and self learning the control knowledge which is used in control process. At the beginning , control rule-base is empty and generally is self constructed, in order to satisfy the performance that is specified in advance. The way may realize self learning multi dimension control rule base. Simulation results demonstrate its adaptable capability and learning capability and reasoning capability.