NONPARAMETRIC REGRESSION AND SHORT-TERM FREEWAY TRAFFIC FORECASTING

The history of short-term traffic forecasting is briefly reviewed, and it is noted that the existence of two regimes in freeway traffic flow suggests that a control algorithm that anticipates transitions from the uncongested to the congested regime could show immproved performance over simple reactive algorithms. The success of such an anticipatory algorithm is dependent on getting good forecasts of traffic conditions. For traffc forecasting, a non parametric approach, k-NN method, is suggested. The paper then describes an empirical study comparing the accuracy of the k-NN method to simple univariate time-series forecasts.

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 00605244
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Mar 31 1991 12:00AM