to Platoon Dispersion Prediction
Saroja Devarakonda, Graduate Student (sarojadsr@hotmail.com)
and
Ping Yi, Advisor (pyi@uakron.edu)
Department of Civil Engineering
The University of Akron
Akron, OH 44325-3905
Platoons, which form at an intersection signal, do not remain intact as they progress along an arterial link towards next signal. The calculation of delays and stops of coordinated traffic signals relies on the ability to accurately predict the platoon dispersion behavior from one signal to another. Very limited field studies have been done for analyzing platoon behaviors so far because the three parameters affecting platoon dispersion — platoon dispersion factor, travel time factor, and smoothing factor — are site specific due to road grades, intersection spacing, curvatures, parking activities, etc. To date, all traffic models, including TRANSYT and SCOOT, are using default values (obtained from a field study more than 20 years ago) for the three parameters. Clearly, this may risk the implementation of virtually ineffective signal timing plans.
In this research, a model was tested which calibrates the aforementioned three platoon dispersion parameters based on links travel time data acquired in real time. Data from four sites in the City of Akron were collected for use in the test. First, Monte Carlo simulation was used to generate platoon patterns for each of the four sites based on the calibrated parameters. Then, these patterns were compared with the field platoon dispersion data. It was found that the platoon patterns predicted by the model had followed the field data very closely. Graphical comparison and statistical testing were used to demonstrate the effectiveness of the model being tested. The advantages of the model and its limitations in practical applications were also discussed.