Research Article  Open Access
Highway Traffic Flow Nonlinear Character Analysis and Prediction
Abstract
In order to meet the highway guidance demand, this work studies the shortterm traffic flow prediction method of highway. The YuWu highway which is the main road in Chongqing, China, traffic flow time series is taken as the study object. It uses phase space reconstruction theory and Lyapunov exponent to analyze the nonlinear character of traffic flow. A new Volterra prediction method based on model order reduction via quadraticlinear systems (QLMOR) is applied to predict the traffic flow. Compared with Taylorexpansionbased methods, these QLMORreduced Volterra models retain more information of the system and more accuracy. The simulation results using this new Volterra model to predict short time traffic flow reveal that the accuracy of chaotic traffic flow prediction is enough for highway guidance and could be a new reference for intelligent highway management.
1. Introduction
The prediction of highway traffic flow is a crucial part of transportation planning, traffic control, and intelligent transportation systems [1]. In particular, shortterm traffic flow predicting can provide real time information for highway guidance. Also accurate predictions of traffic flow are essential part of efficient operations in advanced traffic management systems. In recent years, the shortterm predicting technologies have gotten the attention of many researchers.
Shortterm traffic flow is a typical nonlinear time series in a time horizon of 15 min or less [2]. During the decades a variety of techniques have been applied to forecast the shortterm traffic flow, such as fuzzy theory [3], neural networks [4], Kalman filter [5], and wavelet analysis [6] et al. But the models generated by these methods could not capture some strongly nonlinear characteristics of shortterm traffic flow data. The prediction results may have big error. Chaos theory is an effective tool to study nonlinear system. This tool has been applied to forecast shortterm traffic flow time series [7, 8]. The forecasting results are better than other methods applied before. However, the accuracy of chaotic forecasting result is not enough for real time highway guidance. There remains room for proposing new prediction technology to make the prediction more accurate.
To use chaos theory we should judge whether the traffic flow data is a chaotic time series or not. The largest Lyapunov exponent is usually used to achieve this goal. Wolf algorithm [9], Jacobin algorithm [10], and small data method [11] are three algorithms often used to calculate Lyapunov exponent. If the traffic flow data is a chaotic time series, the phase space reconstruction theory could be applied to predict the traffic flow.
In this paper, the traffic flow data of YuWu highway in Chongqing, China, is taken as study object. We study the general day traffic flow character of highway and the holiday traffic flow model is not contained in this paper. Firstly, the approved small data method is used to judge whether the traffic flow data is a chaotic time series. Secondly, phase space reconstruction theory is used to reconstruct the traffic flow system. Thirdly, a new predicting model based on model order reduction via quadraticlinear systems is used to predict the chaotic traffic flow. This new prediction model has no truncation error and contains more information of original system. It is more accurate than other prediction models ever used before.
2. Traffic Flow Data Phase Space Reconstruction
2.1. Traffic Flow Data
All the traffic flow data are collected at YuWu highway in Chongqing, China. Sampling time begins at 0:00 November 20, 2013, and ends at 0:00 November 27, 2013. Time interval is 5 minutes. Figure 1 shows the traffic data.
2.2. Mutual Information Method
Takens theorem [12] is the mathematical base of time series phase space reconstruction. Choosing optimal time delay and embedding dimension is the key of phase space reconstruction. Mutual information method is an effective tool to calculate .
Assuming chaotic time series time delay is and embedding dimension is . The reconstruction phase space of this time series iswhere . The average information of variable is called information entropy which can be described aswhere is the probability of event in . Let , if system is a coupling system and is given as the uncertainty of iswhere is the probability that a measurement of yields , given that the measured value of is . Given that has been measured at time the average uncertainty in a measurement of at time iswhere is the uncertainty of in isolation. is the uncertainty of in the measurement of . The amount of a measurement of reduces the uncertainty of . The mutual information isThe mutual information is a function of which is the joint distribution probability of the event and . If a vector is a phase space reconstruction while the mutual information first reaches the minimum the time delay could be the delay time of phase space reconstruction.
Figure 2 shows the traffic flow delay time calculated by the mutual information method. When the system reaches the minimum so the optimal delay time .
2.3. CAO Method
Time series can be reconstructed in phase space as follows: , where is the delay time and is the embedding dimension. The nearest neighbor point of is ; the distance between two points is When the space dimension changes from to , the distance between two points isDefineIf , , is the false nearest neighbor point. Substitute (6) into (8); we can get CAO [13] made some changes to (9) yieldingHere is the Euclidian distance. Let , ; is only dependent on the dimension and delay time . It is found that stops changing when is greater than . is the optimal embedding dimension of reconstruction space.
For time series data from a random set of numbers, will never attain a saturation value as increases in principle. But in practical computations, it is difficult to resolve whether is slowly increasing or has stopped changing if is sufficiently large. In fact, since available observed data samples are limited, it may happen that stops changing at some dimension although the time series is random. Therefore, we define
For random series, . For deterministic series, there must exist making
Figure 3 shows the embedding dimension calculated by CAO method. The optimal embedding dimension .
3. Chaotic Character of Traffic Flow
Correlation dimension and largest Lyapunov exponent are two main indexes to distinguish chaotic system from others.
3.1. Correlation Dimension
GP algorithm is the most popular method to calculate correlation dimension for its simplicity. Figure 4 shows the  curves. These curves got through increasing embedding dimension until the slope of the curve’s linear part is almost constant. Then, the correlation dimension estimation can be obtained using least square regression.
Through Figure 4 we can get the correlation dimension ; its fractal reveals that the system is a chaotic system.
3.2. The Largest Lyapunov Exponent of Traffic Flow
In the reconstructed phase space the distance between each point and its nearest neighbor point is expressed aswhere is the mean period. Based on the study in [14] the largest Lyapunov exponent is defined as where is sampling period and is the distance between the pair of nearest neighbors after discretetime steps. Sato also gives an alternate form of (13)where is constant. Combining (13) and (14) yields
By taking the logarithm of both sides of above equation, we obtain
The largest Lyapunov exponent equals the slope of (16).
We can obtain the largest Lyapunov exponent of traffic flow data . Its a positive value which means the system exist chaos phenomena.
4. The Volterra Prediction Model
For singleinputmultipleoutput system in the form ofwhere are the state variables, , is the input signal.
Assume an autonomous system as follows:Let , . The system matrix could be decomposition is unitary; the close form solution to (18) isAlso the input can be taken asCombining (17)–(21) yieldsIf the eigenvalue , , the Sylvester equationhas unique solution. Then let denote the order identity matrix; we getfrom which the solution to (22) is Equation (25) shows the system solution dependent on initial condition and steady state . If the system is driven into period steady state regardless of the fact that is stable or not.
The study [14] reveals that nonlinear ordinary differential equations can be rewritten equivalently in a special representation, quadraticlinear differential algebraic equations (QLDAEs). The QLDAEs take the form asSimilarly, any generalized polynomial system can be converted into generalized QLDAEs as follows:The system (27) is equivalent representation of the original system; this step makes no approximation of original system.
Based on the Volterra theory that the response of a nonlinear system can be decomposed into responses of a series of homogeneous nonlinear systems, the solution of system can be written as follows: where , and . is order Volterra kernel. If the input of system (17) is the response of system is Setting and substituting (29) into (26) we can getand so on.
Substituting (21) into (30) the steady state solution of system is , and solves the Sylvester equationA unique always exists since the eigenvalues of and the imaginary eigenvalues in never add to zero. When the initial condition and input the steady state solution of system is . Therefore, the time argument in the steady state solution can be omitted without ambiguity; namely, .
Set , because and is a scalar so . Equation (31) can be rewritten asMoreover, the derivative of isEquations (34) and (35) could construct another Sylvester equation like (23) with unknown . The new Sylvester equation is
The steady state solution of (36) is . Also , and so on. Setting , , the third Sylvester equation isThe steady state solution of (37) is .
Based on the analysis we could get the pattern of the steady state response , , , and the steady state solution of system isThe initial condition . Equation (38) is exactly the same desired solution sought by SN algorithm.
In this Volterra algorithm because , , the number of Sylvester equations has limitation. This algorithm could converge below any preset tolerance.
In order to simplify the algorithm programing, the traffic flow data is normalized as follows:where is the normalization time series, is the original time series, is average value of original time series, and and are the maximum and minimum value of original time series.
5. Simulation Results
The effectiveness of the Volterra prediction algorithm is verified by the traffic flow shown in Figure 1. The 1500 data of 2500 traffic flow data are used to train the Volterra algorithm. The last 2 hours and 4 hours traffic data of November 27, 2013, are used to verify the effectiveness of the prediction algorithm mentioned in this paper. Also comparing with RBFNN prediction algorithm and old Volterra algorithm, the results show that the accuracy of new Volterra algorithm is better than RBFNN algorithm and old Volterra algorithm. The prediction error is shown in Table 1. The new Volterra algorithm prediction results are shown in Figure 5, and the RBFNN prediction results are shown in Figure 6; the old Volterra algorithm results are shown in Figure 7.

(a)
(b)
(a)
(b)
(a)
(b)
Through Figures 5, 6, and 7 it can be found that the trend of prediction is similar with real traffic flow. But the new Volterra algorithm is more accurate than RBFNN and old Volterra algorithm.
6. Conclusion
This paper analyzes the nonlinear character of YuWu highway traffic flow time series. Using the largest Lyapunov exponent and correlation dimension to show that the traffic flow data are a chaotic time series, the traffic flow is reconstructed by phase space theory and the optimal time delay and embedding dimension are confirmed. Finally a new Volterra algorithm is used to predict the shortterm traffic flow. The prediction results show that this algorithm could predict the shortterm traffic flow effectively. The accuracy could satisfy the demand of highway guidance.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Acknowledgments
Project is supported by the National Nature Science Foundation of China (51407012) and Fundamental Research Funds for the Central Universities (2013G1321037).
References
 D. Chunjiao, S. Chunfu, Z. Chengxiang, and M. Meng, “A shortterm statespace model for free flow prediction based on spatialtemporal characteristics,” Tumu Gongcheng Xuebao, vol. 46, no. 8, pp. 111–118, 2013. View at: Google Scholar
 H. Yin B, J. Xu M, and S. J. Huang, “Traffic flow prediction of signalized intersections using fuzzy_neural network approach,” Journal of South China University of Technology, vol. 28, no. 6, pp. 11–15, 2000. View at: Google Scholar
 X.X. Weng and G.L. Du, “Hybrid elman neural network model for shortterm traffic prediction,” in Proceedings of the 8th IASTED International Conference on Control and Applications, pp. 281–284, Montreal, Canada, May 2006. View at: Google Scholar
 Z. Ye, Y. Zhang, and D. R. Middleton, “Unscented Kalman filter method for speed estimation using single loop detector data,” Transportation Research Record, no. 1968, pp. 117–125, 2006. View at: Google Scholar
 C. L. Yang, L. Jia, Q. L. He, and Q. J. Kong, “Study of traffic flow forecast algorithms based on chaotic wavelet networks,” Journal of Shandong University, vol. 35, no. 2, pp. 46–49, 2005. View at: Google Scholar
 Z. S. Yao, C. F. Shao, and Y. L. Gao, “Research on methods of shortterm traffic forecasting based on support vector regression,” Journal of Beijing Jiaotong University, vol. 30, no. 3, pp. 19–22, 2006. View at: Google Scholar
 Y.M. Zhang, X.J. Wu, and S.L. Bai, “Chaotic characteristic analysis for traffic flow series and DFPSOVF prediction model,” Acta Physica Sinica, vol. 62, no. 19, Article ID 190509, 2013. View at: Publisher Site  Google Scholar
 S.X. Liu, H.Z. Guan, and H. Yan, “Chaotic behavior in the dynamical evolution of network traffic flow and its control,” Acta Physica Sinica, vol. 61, no. 9, Article ID 090506, 2012. View at: Google Scholar
 A. Wolf, J. B. Swift, H. L. Swinney, and J. A. Vastano, “Determining Lyapunov exponents from a time series,” Physica D: Nonlinear Phenomena, vol. 16, no. 3, pp. 285–317, 1985. View at: Publisher Site  Google Scholar  MathSciNet
 G. Barna and I. Tsuda, “A new method for computing Lyapunov exponents,” Physics Letters A, vol. 175, no. 6, pp. 421–427, 1993. View at: Publisher Site  Google Scholar  MathSciNet
 M. T. Rosenstein, J. J. Collins, and C. J. de Luca, “A practical method for calculating largest Lyapunov exponents from small data sets,” Physica D, vol. 65, no. 12, pp. 117–134, 1993. View at: Publisher Site  Google Scholar  MathSciNet
 B. H. Zhang, D. X. Sun, and Y. L. He, “Analysis and prediction of complex dynamical characteristics of shortterm traffic flow,” Acta Physica Sinica, vol. 63, no. 4, Article ID 040505, 2014. View at: Publisher Site  Google Scholar
 S.Q. Zhang, J. Jia, M. Gao, and X. Han, “Study on the parameters determination for reconstructing phasespace in chaos time series,” Acta Physica Sinica, vol. 59, no. 3, pp. 1576–1582, 2010. View at: Google Scholar
 C. Gu, “QLMOR: a projectionbased nonlinear model order reduction approach using quadraticlinear representation of nonlinear systems,” IEEE Transactions on ComputerAided Design of Integrated Circuits and Systems, vol. 30, no. 9, pp. 1307–1320, 2011. View at: Publisher Site  Google Scholar
Copyright
Copyright © 2015 Meng Hui et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.