Traffic volume counts are used in transportation planning, design, operation, and safety analyses. A new methodology establishes a statistical basis for comparing traffic volumes generated from different samples. Monte Carlo simulation was used to generate a cumulative probability function (CPF) of traffic volumes based on the fit-of-Weibull probability distribution to a particular traffic sample. A 90% confidence interval of the traffic volumes from a given traffic sample was obtained from the CPF and was used to compare different traffic samples. A case study was performed by using this methodology to determine if shorter time-frame data may be used to represent longer-time traffic. Results from the case study show that traffic at 20-min intervals may be used to represent 1-h traffic when moderate to high traffic volumes are considered. It was observed that subsamples failed to represent the 1-h traffic data for lower traffic volumes of selected vehicle types.
Abstract