• DocumentCode
    2313824
  • Title

    Real-life CRONOS evaluation

  • Author

    Boillot, F. ; Midenet, S. ; Pierrelée, J.C.

  • Author_Institution
    INRETS, France
  • fYear
    2000
  • fDate
    2000
  • Firstpage
    182
  • Lastpage
    186
  • Abstract
    CRONOS is a real-time urban traffic control (UTC) algorithm developed at INRETS in France (Boillot et al., 1992). It provides the traffic light states (green, amber, red) of junctions for the next few seconds based on a traffic criterion optimisation and using the rolling horizon concept (Gartner, 1982). Three main features characterise it: its ability to control, in the same optimisation process, a zone of a few junctions (typically 5 or 10 junctions). This feature comes from the good properties of the modified Box algorithm version used for the optimisation purpose; its flexibility in the choice of the traffic light states: CRONOS is acyclic and no stages are defined a priori. Each junction is defined as a set of safety constraints on the traffic light groups which define all the forbidden durations and correlations of the states; and its ability to use video based measurements like the queue lengths along the links or the spatial occupancy in the inner junction storing zones. It was decided to undertake a real-life evaluation of the CRONOS performances. After the building of a complete control chain described in Boillot (1997), an 8 months experience has been undertaken on a single junction. We describe in the paper the main features of this experience and the evaluation results
  • Keywords
    optimisation; CRONOS; forbidden durations; inner junction storing zones; modified Box algorithm; queue lengths; real-time urban traffic control algorithm; rolling horizon concept; safety constraints; spatial occupancy; traffic criterion optimisation; traffic light groups; traffic light states; video based measurements;
  • fLanguage
    English
  • Publisher
    iet
  • Conference_Titel
    Road Transport Information and Control, 2000. Tenth International Conference on (Conf. Publ. No. 472)
  • Conference_Location
    London
  • ISSN
    0537-9989
  • Print_ISBN
    0-85296-725-X
  • Type

    conf

  • DOI
    10.1049/cp:20000129
  • Filename
    861259