An Improved Data Collection Method for

Intersection Delay Studies

 

Anthony Ciccolini, Graduate Students (tciccolini@uakron.edu)

Julius Szahlender, Graduate Student (julius@uakron.edu)

and

Ping Yi, Advisor (pyi@uakron.edu)

 

Department of Civil Engineering

The University of Akron

Akron, OH 44325-3905

Abstract

 

     Estimating an intersection’s operational delay remains to be one of the most difficult tasks in traffic analysis today.  Intersection delay data are needed for effective intersection signal control and efficient network-wide traffic flow optimization.  However, delay data cannot be easily obtained because of the complexity involved in the data collection process.  The 1997 Highway Capacity Manual (HCM), the most current version released by the Transportation Research Board, suggested an approximation method of delay data collection for most engineering applications. This method often requires an enormous amount of data recording effort of each of a large data collection crew, especially when the intervals for data summary are small for data accuracy.  In addition, the data collection personnel has to make a subjective judgment at each interval if a vehicle is stopped, slow moving, or is passing through the intersection with the normal speed.

 

     This research proposes an improved method for delay data collection based on the input/output principle of traffic flows at an intersection. With the use of an advanced image sensing system, the input and output flows at an intersection can be automatically and accurately measured at any selected size of time interval.  Such information is then used in an algorithm to obtain delay data. The advantage of the improved method is twofold.  First, it is no longer labor intensive. By using a live video or prerecorded videotape recorded at the intersection, the intersection’s operational delay can be estimated in a lab by a single person. Second, because the image process system is able to continuously extract data from the video image, there is no need to determine by the data collection personnel if a vehicle is in motion or stopped at each ending point of the selected time interval. The effectiveness of the proposed method was demonstrated in a comparison with the HCM method.  Field data from an intersection were used in the study, and the comparison included several commonly used time interval sizes.