Comparison of queue estimation accuracy by shockwave-based and input-output-based models

Abstract

Accurate queue estimation of traffic congestion at recurrent freeway bottlenecks is an important part of developing effective traffic control strategies. Two traffic flow analytical techniques are always employed in studying queuing processes: shockwave analysis and input-output analysis. These two techniques have been taken as separate tools for estimating queue length (size) in a number of studies. Shockwave analysis is particularly well-suited to evaluating the space occupied by the queuing process in units of length, e.g. metre, kilometre etc. Input-output analysis is always employed for vertical queue size in terms of the number of queued vehicles; however, some studies argued that this method underestimated queue size. This paper highlights the reason in the literature for underestimation of queue length with input-out analysis. The objective of this study was to compare real-time queue estimation accuracy using two proposed models based on the two analytical techniques, after investigating the dynamic characteristic of traffic flow at the bottleneck in the freeway. The parameters were determined using discrete field data collected along an urban freeway corridor, Whitemud Drive (WMD), in Edmonton, Canada. This study illustrated that although the two proposed models estimate queue length in different ways, they are compatible and capable of yielding accurate estimates of real-time queue length upstream of the studied bottleneck.