2006•Unpublished venueRequires access

Automated estimation of incident duration on dutch highways

Willem Jan Knibbe, Tom P. Alkim, J.F.W. Otten, Mansa Aidoo

Open publisher page 24 citations

Abstract

Incident management is one of the corner stones of reducing congestion on the national Dutch highways. The consequences of incidents on highways become more and more serious if, as is the case now, the traffic load continues to grow. For that reason, a national incident management program has been implemented in the Netherlands. Part of this program is to continue searching for improvements and new measures to reduce the impact of incidents. One of such possible improvements is not only to inform drivers of incidents, but also to provide them with an estimate of the time it will take until road clearance. This paper reports investigations of the practical feasibility of estimating this time, using a classification tree method. If feasible, such estimates can also be applied in traffic control, e.g. to assess if rerouting traffic is worthwhile for a particular incident. Our research included practical application of a prototype application in a real-time environment. We used information that is currently available at the time of incident occurrence and found that it is possible to give a reliable estimate of incident duration for 2 out of 3 incidents. Limited availability of historical data reduced the number of accurate estimates obtained in practice to 1 out of 3 incidents. Further, most of the traffic operators interviewed would welcome a tool to support them with an estimated incident duration

About this research paper

What this paper is about

Incident management is one of the corner stones of reducing congestion on the national Dutch highways. The consequences of incidents on highways become more and more serious if, as is the case now, the traffic load continues to grow. For that reason, a national incident management program has been implemented in the Netherlands. Part of this program is to continue searching for improvements and new measures to reduce the impact of incidents. One of such possible improvements is not only to inform drivers of incidents, but also to provide them with an estimate of the time it will take until road clearance. This paper reports investigations of the practical feasibility of estimating this time, using a classification tree method. If feasible, such estimates can also be applied in traffic control, e.g. to assess if rerouting traffic is worthwhile for a particular incident. Our research included practical application of a prototype application in a real-time environment. We used information that is currently available at the time of incident occurrence and found that it is possible to give a reliable estimate of incident duration for 2 out of 3 incidents. Limited availability of historical data reduced the number of accurate estimates obtained in practice to 1 out of 3 incidents. Further, most of the traffic operators interviewed would welcome a tool to support them with an estimated incident duration

Why it matters

OpenAlex reports 24 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Incident management is one of the corner stones of reducing congestion on the national Dutch highways. The consequences of incidents on highways become more and more serious if, as is the case now, the traffic load continues to grow. For that reason, a national incident management program has been implemented in the Netherlands. Part of this program is to continue searching for improvements and new measures to reduce the impact of incidents. One of such possible improvements is not only to inform drivers of incidents, but also to provide them with an estimate of the time it will take until road clearance. This paper reports investigations of the practical feasibility of estimating this time, using a classification tree method. If feasible, such estimates can also be applied in traffic control, e.g. to assess if rerouting traffic is worthwhile for a particular incident. Our research included practical application of a prototype application in a real-time environment. We used information that is currently available at the time of incident occurrence and found that it is possible to give a reliable estimate of incident duration for 2 out of 3 incidents. Limited availability of historical data reduced the number of accurate estimates obtained in practice to 1 out of 3 incidents. Further, most of the traffic operators interviewed would welcome a tool to support them with an estimated incident duration

Key concepts: Incident management, Duration (music), Estimation, Incident report, Computer science, Transport engineering, Control (management), Operations research

Related papers

Back to paper searchBrowse research topicsOriginal source
Automated estimation of incident duration on dutch highways — Research Paper | ScholarLens