Speed Estimation Model for Signalised Arterials based on Aggregated QV Relationships

Abstract

This study aims to develop an average travel speed estimation model for signalised arterials based on the relationship between flow (Q) and speed (V). Such relationship, or the QV relationship, describes the travel speed at given flow levels, which plays crucial role in designing traffic management schemes to evaluate the traffic performance after devising the policies. On the other hand, in the state of the practice, the QV relationship of signalised arterials only relates the daily traffic volume and daily average travel speed, which is unable to capture the dynamic changes in traffic, and therefore, hinders the proper evaluation of traffic performance after the policies. This study developed the methodology to estimate the QV relationship of signalised arterials using traffic data aggregated in smaller time intervals (e.g., 15 minutes). The proposed QV relationship relates the space-mean flow and speed along a corridor by utilising multiple detectors. The travel speed estimation model was developed based on a multiple linear regression model, which defined the average travel speed using the space-mean flow, road geometry and time of day. The model exhibited sound accuracy with R² of 0.77, and revealed the impact of road geometry and temporal difference on travel speed.