基于修剪策略的D-FNN直接逆控制算法研究  

Research on D-FNN Direct Inverse Control Algorithm Based on Pruning Strategy

作　　者：张彩霞 刘国文 ZHANG Cai-Xia;LIU Guo-Wen(College of Automation, Foshan University, Foshan 528000;Guangdong Province Smart City Infrastructure Health Monitoring and Evaluation Engineering Technology Research Center, Foshan 528000)

机构地区：[1]佛山科学技术学院自动化学院,佛山528000 [2]广东省智慧城市基础设施健康监测与评估工程技术研究中心,佛山528000

出　　处：《自动化学报》2019年第8期1599-1605,共7页Acta Automatica Sinica

基　　金：广东省教育厅省级重大科研项目(2014KZDXM063);广东省智慧城市基础设施健康监测与评估工程技术研究中心;国家自然科学基金青年基金(61803087,kg33201,61703104)资助~~

摘　　要：神经网络是模拟人脑结构,它具有大规模并行及分布式信息处理能力,但是不能处理和描述模糊信息.模糊系统具有推理过程容易理解,但它很难实现自适应学习的功能.如果结合神经网络与模糊系统,可以取长补短.基于此,本文提出了一种新型动态模糊神经网络(Dynamic fuzzy neural network, D-FNN)学习算法.因为它具有结构和参数同时调整且学习速度快等优点,所以既可以在模糊逻辑系统中包含低级的神经网络学习和计算功能,也可以为神经网络提供高级的类似人的思维和推理的模糊逻辑系统.此外,本文还开发了生物医学工程应用算法程序,针对药物注射系统的直接逆控制案例进行了仿真,结果表明:D-FNN具有实时学习和控制能力强、参数估计和结构辨识同时进行等优点.The neural network simulates the human brain structure with the capabilities processing large-scale parallel and distributed information, which cannot process and describe fuzzy information. The inference processing of the fuzzy system is easy to understand, but it is difficult to realize the adaptive learning. If combining neural networks with fuzzy systems, they can learn from each other. This paper proposes a novel dynamic fuzzy neural network(D-FNN) learning algorithm. Because it has the advantages of simultaneous adjustment of structure and parameters and fast learning speed, it can not only include low-level neural network learning and calculation functions in fuzzy logic systems, but also provide the neural network with the high-level fuzzy logic system which is similar to human thinking and reasoning. In addition, this paper also develops a biomedical engineering application algorithm program, which simulates the direct inverse control case of the drug injection system. The simulating results show that D-FNN has many advantages such as real-time learning, robust control ability, simultaneous parameter estimation and structure identification.

关 键 词：动态模糊神经网络 神经网络 模糊逻辑 模糊规则 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP13[自动化与计算机技术—控制科学与工程]