一种具专家能力的模糊神经网络自组织控制器

随着现代科学技术的迅猛发展，人们所面临的问题日益复杂多变。传统的控制技术与信息处理技术对这些复杂问题时常无能为力，因而产生了智能控制技术。近年来智能控制的研究主要集中在模糊系统、神经网络以及二者结合的模糊神经网络技术方面，特别是模糊神经网络已成为研究者们倾注的焦点。因此提出了一种具有专家调整策略的模糊神经网络自组织控制器的设计方案。其特点是：用神经网络代替传统模糊控制器的隶属函数和权值，实现了模糊规则的自动更新；采用一种对量化、比例因子进行专家调整的策略。仿真结果表明，这种智能控制器具有良好的控制性能。

A New Kind of Self-Organization Fuzzy Controller Based on Neur al Network with Expert Adjusting Ability

With the rapid development of modern science and technology, people are facing more and more complicated and varied problems. Traditional control and information disposal technique can’t deal with these complicated systems, so came the intelligent control technique. Recently, the reserch of intelligent control focuses on fuzzy system, neural network and fuzzy neural network techniques. And fuzzy neural network techniques are becoming people’s main research point. A new kind of self-organization fuzzy controller based on neural network with expert adjusting ability is proposed. It’s characteristics are: the reasoning method of traditional fuzzy controller is replaced by the neural network method which can improve the real-time dynamic performance. The controller is able to modify the membership function on-line, which makes it possible to update the fuzzy rules automatically. The quantization factors and the proportion factors are regulated automatically by expert system. The simulation results show that this intelligent controller has features of response fast, overshoot small, good stability and robust performance.