Font Size: a A A

Study On Chaotic Mechanism In Traffic Flow

Posted on:2008-05-13Degree:MasterType:Thesis
Country:ChinaCandidate:N MaFull Text:PDF
GTID:2132360245478379Subject:Road and Railway Engineering
Abstract/Summary:PDF Full Text Request
The exiting problems in traffic system can be studied by nonlinear chaotic theory, whichcan help to grasp the essences and the rules of the traffic system and provide new methods tosolve the hard problems. Chaos exists in the traffic system. This paper aims at recognizing chaosas early as possible and predicting the traffic flow so that the control step can be adopted timely.According to an intelligent method for real-time recognizing of chaos based on BP network,an intelligent method based on LS-SVM was presented, compared with conventional recognizingchaos, which can't meet this requirement because it needs a large number of samples. Thecomposition of the intelligent system, the algorithm of Lyapunov exponents, and the methods ofextracting characteristic pattern of chaotic time series were discussed. Identifying chaos modelwas established based on LS-SVM, whose input was characteristic pattern decomposed bywavelet packet and output was 0-1, by analyzing the relationship between initial condition andchaos when chaos generated. The experiment result shows that this method is feasible, meets thereal-time requirement of identifying chaos efficiently with a few of samples, and provides thetheory basis for the chaos control of traffic flow.According to BP network forecasting the short-term traffic flow combined with phrasespace reconstruction, LS-SVM forecasting method was presented based on chaotic theory,compared with former forecasting method which can not reveal the uncertainty or reducestochastic disturb factors. The parameters of phrase space reconstruction and the chaos timeseries forecasting methods of the traffic flow were discussed. The forecasting model based onLS-SVM was established, whose input was the time series of k time traffic flow reconstructedand output was the k+1 time traffic flow. The experiment result shows that the generalizationperformance and stabilization of the traffic flow chaos time series forecasting model based onLS-SVM are outstanding.According to ALINEA in freeway ramp control, the method of delay feedback control waspresented in freeway chaos control. The object and the principle of controlling were discussed.The controller was designed. Its input is the difference of the delayed traffic density and trafficdensity, and its output is the ramp metering rate. The experiment result shows that the controllercan make the traffic flow from chaos to masterstroke in optimal status. Delay feedback control issimple, the calculating speed is fast.
Keywords/Search Tags:intelligent transportation system (ITS), traffic flow, chaos, real-time recognizing, short-term traffic flow forecast, phrase space reconstruction, chaos controlling
PDF Full Text Request
Related items