Analysis of Component Errors in the Highway Capacity Manual Travel Time Reliability Estimations for Urban Streets

Abstract

The Highway Capacity Manual 6th edition (HCM6) includes a new methodology to estimate and predict the distribution of average travel times (TTD) for urban streets. The TTD can then be used to estimate travel time reliability (TTR) metrics. Previous research on a 0.5-mi testbed showed statistically significant differences between the HCM6 estimated TTD and the corresponding empirical TTD. The difference in average travel time was 4 s that, while statistically significant, is not important from a practical perspective. More importantly, the TTD variance was underestimated by 70%. In other words, the HCM6 results reflected a more reliable testbed than field measurement. This paper expands the analysis on a longer testbed. It identifies the sources and magnitude of travel time variability that contribute to the HCM6 error. Understanding the potential sources of error, and their quantitative values, are the first steps in improving the HCM6 model to better reflect actual conditions. Empirical Bluetooth travel times were collected on a 1.16-mi testbed in Lincoln, Nebraska. The HCM6 methodology was used to model the testbed, and the estimated TTD by source of travel time variability was compared statistically to the corresponding empirical TTD. It was found that the HCM6 underestimated the TTD variability on the longer testbed by 67%. The demand component, missing variable(s), or both, which were not explicitly considered in the HCM6, were found to be the main source of the error in the HCM6 TTD. A focus on the demand estimators as the first step in improving the HCM6 TTR model was recommended.

Validating and Calibrating the Highway Capacity Manual Arterial Travel Time Reliability Methodology

“…Because they are all positive it indicates that the TTD for each scenario is positively skewed. Note that most arterial TTDs are positively skewed (6,(15)(16)(17).…”

“…Note that there are several HCM6 input parameters that may be adjusted during the calibration process. It was decided to focus on the hour-of-day, day-of-week, and month-of-year demand parameters because it was found that the demand component of the HCM6 TTR model had the greatest impact on the error associated with the HCM6-estimated TTD (6). The goal of the calibration is to adjust the demand parameters such that the resulting estimated HCM6 TTD matches the corresponding empirical TTD for a given case study.…”

“…While informative, these metrics are not robust enough for all planning or operations decision-making. Consequently, in recent years, TTR which attempts to capture the day-to-day variability in TTs, has received considerable research attention (3)(4)(5)(6). For example, a study in Orlando, Florida, examined the relationship between TTR and crash frequency on an urban expressway (7).…”

“…Recently, it was shown that the HCM6 approach severely underestimated the TTD and associated TTR metrics on a 1.16 mi testbed in Lincoln, NE (5,6). Subsequently, it was shown that this underestimation issue could be rectified using a calibration methodology designed specifically for the HCM6 TTR methodology (9).…”

New Travel Time Reliability Methodology for Urban Arterial Corridors

“…With the recent increase in urban traffic congestion, there has been an increasing need for reliable travel time information. Consequently, travel time reliability (TTR), which attempts to capture the day-to-day variability in travel times, has recently received considerable research interest ( 3 , 4 ).…”