《智能控制智能控制智能控制 (7).pdf》由会员分享,可在线阅读,更多相关《智能控制智能控制智能控制 (7).pdf(16页珍藏版)》请在得力文库 - 分享文档赚钱的网站上搜索。
1、Several Significant Branches of ICSeveral significant branches of ICIntelligent Control can be divided into the following major sub-domains:Fuzzy Control Neural Network Bsed Control Expert Control Evolutionary OptimizationFuzzy ControlClassical control methods are based on the precise math model of
2、the controlled objects.People find that it is possible to control complex systems by simulating a controller that simulates the structure of the brains thinking.Fuzzy Control In 1965,Fuzzy Set was firstproposed by Lotfi A.Zadeh,which lay the foundation forFuzzy Control.In 1974,E.H.Mamdani built the
3、worlds first fuzzy logic controller,creating a history of Fuzzy Control.Fuzzy ControlFuji electric started the first application of fuzzy controlwater purification in Japan in 1983.Fuji used fuzzy control technology on Sendai subway in1987.In 1989,they made Japan a dominant state on fuzzy.当前无法显示此图像。
4、Fuzzy ControlThe development of fuzzy control can be divided into three phases:From 1965 to 1974 is the first stage of development of fuzzy control,that is the fuzzy mathematics development and formation stage;From 1974 to 1979 is the second stage,this period the fuzzy controller come into being.Fro
5、m 1979 to the present is the third stage,namely high-performance fuzzy control period.Fuzzy ControlFuzzificationFuzzy InferenceDefuzzificationRule-baseInputOutputNeural Network Based ControlFeaturesIt has good robustness and fault tolerance.Parallelism Pattern recognition Memorizing Self-learningNeu
6、ral network control is an intelligent control method simply simulating brains physiology system structure.Neural NetworkResearch on neural network has last for the past decades.In 1943,McCullouch and Pitts created the neural mathematical model.In 1975,Albus put forward the brains memory model CMAC n
7、etwork.In 1976,Grossberg promoted self-organizing network used for pattern classification in unsupervised learning.Neural NetworkIts period of rapid growth came when enter 1980In 1982,Hopfield solved the learning problems in regression networks.In 1986,the PDP research groups BP network achieved sup
8、ervised network learning.Neutral network control is formed by introducing neutral network into control field.Expert systemExpert system is one of the systems that can handle real world,complex problem requiring an experts interpretation.Expert system is one of the most active application research ar
9、eas in artificial intelligence.Expert systemThe functions of expert system depend on its massive knowledge contained.The key problem in the developing and using the expert system is how to present and use the expertise.UncertaintyIncompletenessimprecisionExpert ControlThe expert system can be applie
10、d to many control scenarios such as:space traffic controlautonomous robotics controlExpert ControlFight managementFault detection and diagnosisExpert Control:An exampleUnderwater robotSummary Data,Information,Knowledge,Intelligence Automation&Artificial Intelligence Challenges Conventional Control Faces Definition,Development&Characteristic of IC Structural Theories&Several Significant Branches of IC