Modelling Travel Time After Incidents on Freeway Segments in China

Han Ru,Zhihao Duan,Xingliang Liu,Jinliang Xu,Chao Gao

doi:10.1109/access.2019.2951792

Han Ru, Zhihao Duan + Show 3 more

Open Access

https://doi.org/10.1109/access.2019.2951792

Copy DOI

Abstract

The reduction in incident-induced delays on freeways is a main objective of transportation management. The use of travel time estimation model for freeway segments is an important method for estimating delays resulting from incidents on freeways. In this study, freeways with temporary partial lane closures were considered to simulate traffic accidents occupying lanes. Travel time, traffic volumes, and speeds under various traffic conditions on a few typical Chinese freeway segments under regular and simulated accident conditions were investigated through field experiments. The collected traffic data collected were used to establish travel time models based on the Bureau of Public Roads (BPR) function for basic freeway segments under both regular and accident conditions, and to obtain the model parameters. The results demonstrate that the calibrated BPR models established in this study fit the data well. In addition, this study proposes an application method for the established travel time models by which variations in travel time can be estimated rapidly and easily. The results of this study can be used to reduce travel time for road users and contribute to decision making of transportation management systems to improve traffic efficiency after incidents.

Highlights

Travel time is widely recognized as an important performance measure of highway operating conditions and is one of the key factors for road users in arranging travel plans and choosing travel routes [1]
If the number of available lanes and v/C ratio can be determined after incidents, the model developed in this research can be used to calculate accurate travel times
Vehicle characteristics and driver behavior have changed considerably in the years since the standard Bureau of Public Roads (BPR) function was introduced; the standard BPR function is based on data that do not reflect today’s traffic operating conditions in China

Summary

Introduction

Travel time is widely recognized as an important performance measure of highway operating conditions and is one of the key factors for road users in arranging travel plans and choosing travel routes [1]. The estimation and forecasting of travel time is even more important for traffic operators and emergency response services. Tang developed a new travel speed prediction method based on an evolving fuzzy neural network in 2017 [2]. In 2019, he proposed an improved traffic flow prediction model based on artificial neural networks and a traffic flow prediction method combining denoising schemes and the support vector machine model [3], [4]. Many researchers have focused on predicting travel time using various methodologies, such as theoretical model analysis, regression model analysis, machine learning methods and software simulation.

Objectives

Methods

Findings

Conclusion