Bayesian Methods for Estimating Average Vehicle Classification Volumes

Principal Investigator(s):

Gary Davis, Professor, Civil, Environmental and Geo-Engineering

Project summary:

This project will (1) investigate a new method for clustering ATR data and (2) extend methods developed to make efficient use of short vehicle counts to the problem of estimating mean daily traffic for different vehicle classifications. The univariate time-series model developed in earlier work will be replaced by a multivariate time-series model, with the different classification counts being the components of the multivariate observation. Model fitting and validation will be done, factor groups developed, and then (a) a method for assigning short count sites to factor groups, and (b) a method for estimating mean daily traffic by vehicle classification will be tested.

Project details:

  • Project number: 1996014
  • Start date: 03/1997
  • Project status: Completed
  • Research area: Transportation Safety and Traffic Flow
  • Topics: Data and modeling