Characteristic Analysis for Regional Traffic Data Using Random Matrix Theory

  • Haichun LiuEmail author
  • Changchun Pan
  • Genke Yang
  • Chunxia Zhang
  • Robert C. Qiu
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 340)


Traffic regional feature analysis of city is an important problem in the study of macro transportation system. The analysis of the current traffic regional feature of the city is mainly through the floating car data, such as velocity’s average or variance to characterize the features of different regions. But with the increasing scale of the traffic data in sampling and storage, the average (or variance) velocity as the evaluation of city traffic regional feature such as distinguishing urban and suburb has no significant through our analysis. Therefore, in this paper, we use the Random Matrix Theory (RMT) to analysis traffic data, which is based on floating car position data’s distribution of singular value equivalence matrix eigenvalue to distinguish different region such as urban and suburban. By using 1 week’s floating car data, we verify the data’s eigenvalue of singular value equivalent matrix distribution is better than the velocity’s average or variance in indicating the regional feature of city traffic.


RMT Big data Spectral analysis Data mining 


  1. 1.
    Qi L (2008) Research on intelligent transportation system technologies and applications. Workshop on power electronics and intelligent transportation system, pp 529–531Google Scholar
  2. 2.
    Kriegel HP, Renz M, Schubert M, Zuefle A (2008) Statistical density prediction in traffic networks. In: SIAM intelligent conference, pp 692–703Google Scholar
  3. 3.
    Hiribarren G, Herrea JC (2014) Real time traffic states estimation on arterials based on trajectory data. Transp Res Part B 69:19–30CrossRefGoogle Scholar
  4. 4.
    Jaeyoung L, Mohamed AA, Ximiao J (2014) Development of zone system for macro-level traffic safety analysis. J Transp Geogr 38:13–21CrossRefGoogle Scholar
  5. 5.
    Wigner EP (1958) On the distribution of the roots of certain symmetric matrices. Ann Math 67:325–327CrossRefzbMATHMathSciNetGoogle Scholar
  6. 6.
    Marčeko VA, Pastur LA (1967) Distribution of eigenvalues for some sets of random matrices. Math USSR-Sb 1:457–483CrossRefGoogle Scholar
  7. 7.
    Debbah M, Muller R (2005) MIMO channel modelling and the principle of maximum entropy. J Latex Class Files 1(11):71–77Google Scholar
  8. 8.
    Zhidong B, Silverstein JW (2006) Spectral analysis of large dimensional random matrices. Science Press, BeijingGoogle Scholar
  9. 9.
    Robert Q, Browning JP (2012) Spectrum sensing in cognitive radio with robust principal component analysis. IEEE waveform diversity and design conference, vol 1, p 177Google Scholar
  10. 10.
    Burda Z, Livan G, Swiech A (2013) Commutative law for products of infinitely large isotropic random matrices. arXiv preprint arXiv 1303.5360
  11. 11.
    Plerou V, Gopikrishan P, Rosenow B, Amaral L, Gurh T (2002) Random matrix approach to cross correlations in financial data. Phys Rev E 65:066126CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2015

Authors and Affiliations

  • Haichun Liu
    • 1
    Email author
  • Changchun Pan
    • 1
  • Genke Yang
    • 1
  • Chunxia Zhang
    • 2
  • Robert C. Qiu
    • 3
  1. 1.Department of AutomationShanghai Jiao Tong UniversityShanghaiChina
  2. 2.Shanghai Wainmac Information Technology Co. LtdShanghaiChina
  3. 3.Department of Electrical and Computer EngineeringTennessee Technological UniversityKnoxvilleUSA

Personalised recommendations