An exploratory study of vehicle class headway ratios as passenger car equivalence values using 3SLS estimation

Daniel P Van Boxel, Purdue University

Abstract

The current Highway Capacity Manual (HCM) method for accounting for trucks on basic freeway sections uses a microscopic simulation of equivalent delay. Although this method can provide passenger car equivalent (PCE) values separately for single unit and combination trucks, the HCM combines them into one generic truck class. An alternative method such as headway ratios, however, may provide a strong theoretical density basis as well as use real data to determine more robustly PCEs separately for each truck class. The present study implements such a method, using real-time vehicle data from I-65 outside Indianapolis. Three stage least squares estimates specific class headways from which one calculates the ratio as the PCE value. Using this method produces some variation in Level of Service (LOS) from the traditional method. Further, extrapolation of traffic conditions may drastically change LOS determinations. ^ This model, while exploratory, also examines regional stability of PCE values. A likelihood ratio test determines that two data sets, one from an Indiana Microloop and one from an Ohio Weigh-in-Motion station, spawn statistically distinct models. Future studies will have to address this disparity by considering regional indicators or preferred data types.^ While the headway ratio method explored in this study is not novel, the implementation of using predicted headways is. Using a limited data set, however, the model is not sufficient for design purposes. Additional data, geometric variables, and standard data types must predate a final model adoption. Still, however, this study develops headway ratio as a feasible and rational method for computing passenger car equivalents and hence revised LOS.^

Degree

M.S.C.E.

Advisors

Kumares C. Sinha, Purdue University, Darcy M. Bullock, Purdue University.

Subject Area

Engineering, Civil|Transportation

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS