Traffic Incident Detection Based on Wavelet and Neural Network

Traffic incident is the main cause of nonrecurring congestion. To detect incident and to verify incident as soon as possible will reduce their adverse effects. Several traditional algorithms used in automatic incident detection are reviewed. The parameters of traffic flow detection were analyzed and processed by means of wavelet and wavelet packet and then pattern recognition of the parameters' character was done by neural network. By comparing the performances of new algorithm with traditional algorithms, it shows its advantages in many aspects, such as higher diction rate, lower false alarm rate and shorter mean detection time.

Language

  • English

Media Info

  • Media Type: Web
  • Features: Figures; References; Tables;
  • Pagination: pp 2869-2875
  • Monograph Title: International Conference on Transportation Engineering 2009

Subject/Index Terms

Filing Info

  • Accession Number: 01531060
  • Record Type: Publication
  • ISBN: 9780784410394
  • Files: TRIS, ASCE
  • Created Date: Nov 12 2013 1:42PM