Hybrid Approach for Clustering Vehicle Classification Data to Support Regional Implementation of the Mechanistic–Empirical Pavement Design Guide

This paper develops a hybrid approach for analyzing vehicle classification data and applies the approach to a fused data set from multiple jurisdictions in the Canadian prairie region. Application of the approach results in a set of regional default truck traffic classification groups for use in the Mechanistic–Empirical Pavement Design Guide. The hybrid approach is a conglomeration of three components: statistical clustering procedures, expert judgment, and industry intelligence. By applying the hybrid approach, analysts receive the joint benefits of analytical rigor and industry-oriented pragmatism. Application of this approach results in eight truck traffic classification groups for the Canadian prairie region that exhibit distinct differences from the default distributions developed for national use in the United States. The benefits of applying the hybrid approach on fused data sets include (a) the statistical strength gained from use of additional classification data, (b) the development of truck traffic classification groups that better reflect the diversity of patterns in a region, and (c) the potential for improved ability to capture future shifts in truck traffic characteristics because of experience gained in other jurisdictions. The paper also identifies limitations to the hybrid approach that should be considered. These limitations include varying data quality between jurisdictions, the sensitivity of low-volume sites to changes in industry patterns and the ability to track these changes, and potential shortages of continuous classification sites. When its benefits and limitations are well understood, the hybrid approach can be applied to truck traffic data analyses in any jurisdiction.