Identification Method Of Chaos In Traffic Flow From Small Data Sets

For the traffic control based on chaos, chaos in traffic flow must be identified as soon as possible, in order to take the control measure in time and make the traffic flow tend towards in order. But the existing chaos identifying methods demand large amount samples and calculate slowly, they can't meet the request for real-time character. Therefore the research of real-time identifying of chaos has not only important academic significance, but also has important practical value.The chaos phenomenon in the traffic flow of qualitative identification and quantitative estimation was researched, and the validity of the proposed method was verified through the simulations.In the qualitative identification the chaos traffic flow, one kind of identification method is presented, the method finds the corresponding relation of the chaos and initial condition through support vector machine, then the question of identifying the chaos can turn into the question of weather the system enter the initial condition. An overall frame of qualitative identification method of the systematic chaos phenomenon based on small amount data is given, and the implementation method of each subsystem in detail is introduced.In the quantitative estimation to the chaos in the traffic flow, a method based on the wavelet network and calculating of Lyapunov exponents was proposed. In the mothod, the one-dimensional serial of the system data is reconstructed, and the dynamical equation of the reconstructed system is estimated by the wavelet network. Then Lyapunov exponents table of the reconstructed system are calculated, studying the unknown system through the reconstructed system. In the experiment part, the Lyapunov exponent tables of Logistic and Henon through the one-dimensional serial were calculated, the results indicate that the wavelet network has good nonlinearity and approaches ability, and the final result of calculation is high in precision, the demand of sample is counted few, thus this method can meet the request for real-time character. The use of the method for identification of chaos in the traffic flow was simulated.