Travel time prediction algorithm based on neural network and data fusion

Current travel time prediction algorithms often need large numbers of travel time data, which are difficult to gain and highly cost, to identify the algorithms' parameters. In this paper, we propose new travel time prediction algorithms, which use neural network to predict future speed dynamically, and use data fusion to integrate the speed data of different detectors and, to calculate the travel time. Vehicle plate recognition technology is used to collect the real travel time of the test section on Beijing Third-Ring freeway to evaluate the algorithms. From the obtained results, the average prediction error is less than 10%. For the covering abstract see ITRD E140665.

  • Authors:
    • ZHANG, K
    • CHEN, D
    • Guan, Jizheng
  • Publication Date: 2007


  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01151471
  • Record Type: Publication
  • Source Agency: Transport Research Laboratory
  • Files: ITRD
  • Created Date: Mar 1 2010 8:32AM