FREEWAYTRAFFICFLOWMODELINGBASEDONRECURRENTNEURALNETWORKANDWAVELETTRANSFORM基于递归神经网络和小波变换的高速公路交通流建模6Conclusions结论Thehighlynonlinearanddynamiccharacteristicsofthemacroscopictrafficflowrequireamodelingapproach,whichcandealwiththecomplexnonlinearrelationshipsamongthespeed,flowanddensity.宏观交通流的高度非线性和动态特性,需要一个建模方法,它可以处理的速度,流量和密度之间的复杂的非线性关系。Atthesametime,effectivemeasureshavetobetakentoeliminatetrafficnoiseanddisturbance.Inthispaper,amethodofwaveletdenoisingandtrafficflowmodelingbyanElmanrecurrentneuralnetworkispresented.同时,必须采取有效措施来消除交通噪声和干扰。在本文中,一种小波去噪和交通流的Elman神经网络建模这工作了。Simulationresultsshowthatthewavelettransformcaneffectivelyeliminatethenoisesignal,andthattheElmannetworkcanaccuratelydescribetherealbehavioroffreewaytrafficflow.仿真结果表明,小波变换可以有效地消除信号中的噪声,和Elman神经网络能够准确地描述高速公路交通流的真实行为。Thismethodisofgreatimportancetorealizeon-linemodelingandcontroloffreewaytrafficflow.此方法对实现高速公路交通流的在线建模与控制具有重要意义。FuzzySelf-AdaptivePIDControllerforFreewayRampMetering高速公路匝道控制的模糊自适应控制器Abstract:Aimingatthenonlinearandtime-varyingcharacteristicsoffreewaytrafficsystem,afuzzyself-adaptivePIDcontrollerisdesignedandappliedtofreewayrampmeteringinthispaper.摘要:针对高速公路交通系统的非线性、时变特性,设计了一种模糊自适应控制器,并将其应用于高速公路匝道控制系统。Atrafficflowmodeltodescribethefreewayflowprocessisfirstlybuilt.Basedonthemodelandinconjunctionwithnonlinearfeedbacktheory,afuzzy-PIDrampcontrolleristhendesigned.TherampmeteringrateisdeterminedbythePIDcontrollerwhoseparametersaretunedbyfuzzylogicaccordingtothedensitytrackingerroranderrorvariation.首先建立了一个交通流模型来描述高速公路的流程。基于该模型,并结合非线性反馈理论,一个模糊的控制器,控制器,然后设计。斜坡根据密度跟踪误差和误差变化情况,通过模糊逻辑对控制器参数进行调整,确定了测量速度。Gaussandtrianglecurvesareusedforthemembershipfunctionsofthefuzzyvariables.Therulebaseincluding49fuzzyrulesisalsoestablished.Finally,thecontrolsystemissimulatedinMATLABsoftware.Theresultsshowthatthiscontrollerdesignedhasfastresponse,gooddynamicandsteady-stateperformance.Itcanachieveadesiredtrafficdensityalongthemainlineofafreeway,andcanmakevehiclestravelmoreefficientlyandsafely.Thisapproachisquiteeffectivetofreewayrampmetering.高斯和三角曲线用于模糊变量的隶属函数。还建立了包括49个模糊规则的规则库。最后,对控制系统进行了仿真研究的软件。结果表明,该控制器具有响应快、动态性能和稳态性能。它可以实现所需的交通密度,高速公路主线,并能使车辆行驶更安全、更安全。这种方法是相当有效的高速公路匝道控制。2.FreewayTrafficFlowModel高速公路交通流模型Consideramultiplelane(λ)freewaysectionwithasingleentryramp,asshowninfigure1.考虑多车道(λ)具有单一入口匝道路段,如图1所示。Assumethatattimeslicek,vehiclesflowintoagivensectionatarateofqu(k)vehiclesperhourperlanefromitsupperboundaryandr(k)vehiclesperhourfromtheentryramp.假设在时间片上,车辆进入一个给定的区段在一个区段(克)每车道从它的上边界和(钾)车辆每小时从入口匝道的速度。Theydischargeatarateofq(k)vehiclesperhourperlaneatitslowerboundary.他们以每小时的速度在其较低的边界处排放一个问(克)的车辆。Bytheconservationprinciple,thenumberofvehiclesinthisfreewaysegmentattime按养护原则,在这条高速公路的车辆数量在时间k+1,N(k+1)wouldbe)]()()([)()1(krkqkqtkNkNu(1)ΔxquρvqrFigure1Afreewaysection)]()()([)()1(krkqkqtkNkNu(1)Definetrafficdensityas,whereislengthoftheroadsegment.Equation(1)cannowbewrittenintermsofdensity定义交通密度,哪里是公路段的长度。方程(1)现在可以用密度来书写]/)()()([)/()()1(krkqkqxtkku(2)Empiricalevidencesuggeststhatarelationshipbetweenflowandtrafficdensityexists.Thisrelationship,denotedasisgenerallyreferredtoasthefundamentaldiagramoftrafficflow.Manyformsoffundamentaldiagramshavebeenproposed,allofwhichsharesomecommonfeatures:经验证据表明,流量和交通密度之间的关系存在。这种关系,表示为通常被称为交通流基本图。许多形式基本图已经被提出,所有这些都有一些共同的特点:(1)flowiszerowhendensityiszero;(2)thereisamaximumdensity(oftenreferredtoasjamdensity)thatcorrespondstobumper-to-bumpertraffic(atwhichflowisalsozero);and(3)thereexistsatrafficdensityatwhichflowismaximal(oftenreferredtoascriticaldensity).(1)密度为零时的流量为零;(2)有一个最大密度(通常称为果酱密度),对应于保险杠到保险杠的流量(在流量也是零);(3)存在流量最大的交通密度(通常称为临界密度)。Whentrafficdensityisbelowcriticaldensity,flowincreaseswithdensity;andwhentrafficdensityisabovecriticaldensity,flowdecreaseswithdensity.交通密度低于临界密度时,流量随密度增大而增大;当交通密度超过临界密度时,流量随密度的增加而减小。Equation(2)andthefundamentaldiagramconstituteadiscreteformofthewellknowntrafficflowmodelcalledLWRmodel[4].方程(2)和基本图构成的离散的众所周知的交通流模型的LWR模型[4]形式。TheLWRmodelprovidesasimpleandtractableframeworkfortrafficsimulationandcontrolstudies.LWR模型提供了一个简单的框架,交通仿真与控制研究。Itgivesgoodqualitativeresultsindescribingsuchtrafficphenomenaasshockwavesattrafficbottlenecks,progressionofatraffichump,andstartingandstoppingtrafficatintersections.它提供了良好的定性结果,在交通瓶颈,交通驼峰的进展,并停止交通在路口的交通瓶颈,描述这样的交通现象。TheLWRmodelisusedinthispaperasthesystemmodelfordevelopingtraffic-responsiverampcontroller.LWR模型作为本文发展交通响应匝道控制器的系统模型。Substituteinto(2),anddefine,weobtainequation(3).替代成(2),并定义,我们得到方程(3)。/)()()]([)()1(xrkqkfkku(3)Withinitialcondition,equation(3)givesacompletedescriptionofthefreewayflowprocess.初始条件,方程(3)给出了高速公路交通流过程的完整描述。FuzzySelf-AdaptivePIDControllerforFreewayRampMetering高速公路匝道控制的模糊自适应控制器Theoptimizationprocedureswithgeneticalgorithmareasfollows.遗传算法的优化程序如下(1)Determinationofoptimizedparametersandencodedmode.优化参数和编码方式的确定。TheparametersneededtobeoptimizedareproportionalcoefficientKpandintegralcoefficientKi,andtheencodedmodeisrealnumbercode.待优化参数比例系数KP和积分系数Ki和编码方式是实数编码。(2)Generationofinitialpopulationrandomly随机初始种群的生成Theinit