Neural Agent Car-Following Models

This paper applies reactive agent techniques based in artificial neural networks (ANN) for the intelligent transportation concept of car-following. The model, which was verified on both microscopic and macroscopic levels, was found to be a better predictor than both the Gipps and psychophysical car-following models. It is also found by local and asymptotic stability analyses that this ANN is robust for such applications despite both mild and severe traffic disturbances. The authors discuss a number of historical car-following models and list their method and weaknesses. These include: GHR model, CA model, the linear model, the psychophysical/AP model mentioned above, the fuzzy logic based model, desired spacing models, capacity drop, and neural network-based models. In future research, the authors intend to explore the studied ANN with a larger data set and in a wider ranger of driving situations.

  • Availability:
  • Authors:
    • Panwai, Sakda
    • Dia, Hussein
  • Publication Date: 2007-3


  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01044832
  • Record Type: Publication
  • Source Agency: UC Berkeley Transportation Library
  • Files: BTRIS, TRIS
  • Created Date: Mar 22 2007 2:08PM