Dilemma zone at signalized intersection has been recognized as a major potential causing rearend

and right-angle crashes, and has been widely studied by researches since it was initially

proposed as the GHM model in 1960. However, concepts conventionally defined to represent the

yellow phase dilemma lack integrity. This research conducts a comprehensive literature review

with attempt to clarify the interrelationship among dilemma zone, option zone, and indecision

(decision) zone, and to develop a heuristic framework to present the contributing factors in

dilemma zone modeling. A new method for modelling the locations and lengths of the dilemma

zone using video-capture techniques and vehicle trajectory data is presented in this report. First,

dilemma zone is mathematically modeled based on the GHM model. Then, field-observed

trajectory data extracted by the video-capturing-based approach are used to calibrate the

contributing factors involved in the dilemma zone model. The high accuracy of the time-based

trajectory data has significantly enhanced the accuracy of the calibrated dilemma zone models.

Two sets of trajectories are explored for calibrating the dilemma zone contributing factors. One

is concerned with maximum yellow-onset safe passing distance and minimum yellow-onset

stopping distance. The other is concerned with Xth percentile yellow-onset passing distance and

(100-X)th percentile stopping distance for the prevailing travel behaviors. The latter alternative

actually precludes “too conservative” and “too aggressive” maneuvers in response to yellow

indications.

One critically important result is the dilemma zone look-up charts that are developed

based on the calibrated dilemma zone models. Such charts provide a convenient tool to identify

the locations and lengths of dilemma zones for any speed and yellow duration conditions.

Additionally, impact of arrival type and vehicle types are also explored. Results reveal that

traffic in a good progression (Arrival Type ≥ 4) has a further option zone. It is also discovered

that the length of option zone decreases as the vehicle size increases, while the downstream

boundary of option zone is further from the stop line as the vehicle size increases. In overall, this

project aims to conduct a preliminary research for providing a proof of concept about the

modeling of dynamic dilemma zones, and validating the feasibility of the methodology for

calibrating the dilemma zone model using trajectory data. The methodology used in this study

establishes a solid basis for future research of the optimum signal detection placement and

related dilemma zone protection problems with consideration of multi-speed protection.

This project is granted by the Ohio Transportation Consortium (OTC). The research team

is thankful to Dr. Ping Yi at University of Akron and staffs at OTC for their strong support to the

project. Gratitude goes to Ph.D. students, Mr. Zhixia Li and Mr. Qingyi Ai, and M.S. students,

Mr. Vijay Krishna Nemalapuri and Mr. Sudhir Reddy Itekyala for their effective assistances in

field data collection. In particular, Mr. Zhixia Li took the lead in data collection and analysis and

participated in drafting the report. Finally, the research team also expresses our thanks to Ms.

Brenda Slaughter, senior Grant Administrator at UC Sponsored Research Services and Mr. Tom

Davis, senior Grant Administrator at UC Department of Civil and Environmental Engineering

for their administrative support. This research could be not successfully finished without all their

active participations, critical contributions, and strong supports.