屆本科畢業(yè)論文（設計）論文題目：神經(jīng)網(wǎng)絡PID控制系統(tǒng)設計與仿真學生姓名：所在院系：所學專業(yè)：應用電子技術教育指導老師：完成時間：摘要PID控制是最早發(fā)展起來的控制策略之一，由于其算法簡單、魯棒性好和可靠性高，被廣泛應用于過程控制和運動控制中，尤其適用于可建立精確數(shù)學模型的確定性控制系統(tǒng)。但是常規(guī)PID控制器參數(shù)往往整定不良、性能欠佳，對運行工況的適應性很差。而神經(jīng)網(wǎng)絡具有很強的非線性映射能力、自學習能力、聯(lián)想記憶能力、并行信息處理方式及優(yōu)良的容錯性能。本設計提出了一種單神經(jīng)網(wǎng)絡的PID控制器，使人工神經(jīng)網(wǎng)絡與傳統(tǒng)PID控制相結合互相補充，共同提高控制質量，并利用Matlab軟件進行仿真。仿真結果表明，神經(jīng)網(wǎng)絡PID控制器具有較高的精度和很強的適應性，可以獲得滿意的控制效果。關鍵詞：PID控制器；單神經(jīng)網(wǎng)絡PID控制器；BP算法NeuralNetworkPIDControlSystemDesignandSimulationAbstractThePIDcontrolisoneofcontrolstrategieswhichdevelopsmostearly,becauseitsalgorithmissimple,robustnessisgoodandthereliabilityishigh,iswidelyappliedintheprocesscontrolandthemovementcontrol,especiallyitsuitableinthequalitativecontrolsystemofestablishingtheprecisemathematicalmodel.ButtheconventionalPIDcontrollersparametertuningisoftenpoor,performanceofpoor,operatingconditionsofthepooradaptability.Buttheneuralnetworkhastheverystrongnon-linearmappingability,capabilityoflearningbyitself,associationmemoryability,thewayofparallelinformationprocessingandthefinefault-tolerantperformance.ThisdesignproposedonekindofPIDcontrollerofmononeuralnetwork,causestheartificialneuralnetworksandthetraditionalPIDcontrolunifiessupplementedmutually,improvesthecontrolqualitytogether,anditcanusethesoftwareofMatlabcarriessimulation.Thesimulationsresultindicatedthat,PIDcontrollerofneuralnetworkhasthehighprecisionandtheverystrongcompatibility,mayobtainthesatisfactioncontroleffect.Keywords:PIDController；SingleNeuralNetworksPIDController；BPAlgorithms目錄1緒論.11.1常規(guī)PID控制的特點.11.2神經(jīng)網(wǎng)絡和PID控制相結合的研究現(xiàn)狀.22神經(jīng)網(wǎng)絡的基本原理.32.1PID神經(jīng)元的基本模型結構.32.2PID神經(jīng)元的計算.43神經(jīng)網(wǎng)絡PID的系統(tǒng)設計.83.1單神經(jīng)網(wǎng)絡PID的結構.83.2單神經(jīng)網(wǎng)絡PID控制器的算法（BP算法）.93.2.1算法的基本思想.93.2.2正向計算方法.93.2.3反傳學習計算方法.113.3單神經(jīng)網(wǎng)絡PID的等價系統(tǒng).124系統(tǒng)仿真.134.1神經(jīng)網(wǎng)絡工具箱概述