Estimation of Travel Time by Vehicle Type on Unobserved and Observed Links with Limited Detector Data from Different Sources

Estimating travel time by vehicle type is challenging, particularly for an expressway corridor with limited data on different vehicle types. In practice, path travel time is collected by automatic vehicle identification (AVI) detectors, which can differentiate vehicle types and is installed on expressways with comparative longer spacing. Additionally, while link travel times for all vehicles may be available for links equipped with point detectors, in reality, the number of observed links is limited. Most current methods focus on estimating the travel times of site-specific links or paths without classifying vehicle type. Few attempts have been made to estimate travel times by vehicle type on observed links in a long study corridor or path with very limited detector data on observed links from different sources. Therefore, this paper proposes a quadratic optimization model that makes full use of both AVI and point detectors and estimates the travel time by vehicle type on both the observed and unobserved links along the expressway concerned. A case study using real-world data from Hong Kong is presented to show the merits of the proposed modeling approach. The estimation results are satisfactory and robust in different scenarios.