 |
|
 |
|
| Artificial Neural Network Bus Arrival Time Prediction Tool for Transit Signal Priority with Nearside Bus Stops |
 |
| Investigators:
|
Dr. Francois
Dion, Department of Civil & Environmental Engineering
Dr. Ghassan
Abu-Lebdeh, Department of Civil & Environmental Engineering
|
| Research
Assistant |
Mohammad
Ghanim, Ph.D. Student, Department of Civil and Environmental Engineering
|
| Dates: |
January
2006-December 2006 |
Abstract
Many transit
agencies are currently considering implementing priority systems providing
buses with temporary green signal extensions and early green recalls.
While many studies have demonstrated a potential for bus delay reduction
and negative traffic impacts, very few have addressed the problems
posed by variable dwell times at nearside bus stops. At these intersections,
the high uncertainty of accurately predicting the arrival time of
buses at the intersection may significant impact the performance of
transit signal priority systems.
This project
aims to develop an ANN-based prediction tool for providing reasonably
reliable information about the expected arrival time of buses at intersections
where dwell activities downstream of a detection point may create
uncertainties in the predictions. To obtain a tool that can be applied
in real-world settings, predictions are to be based on information
that is typically available to traffic signals controllers, traffic
information systems, or that can be obtained from current sensing
technologies. At a minimum, the tool should be able to make predictions
based only on information provided by inductive loops or other point
detectors and a general knowledge of traffic conditions.
Development
of the prediction tool is conducted with the help of the VISSIM microscopic
traffic simulation model. This model is used to generate synthetic
data for training and testing the ANN model, as well as an evaluation
tool to assess the performance of the resulting prediction tool.
Evaluation results
show that use of the prediction tool with standard signal priority
logic can provide additional bus delay reductions when compared to
the current practice of predicting bus arrivals using average values
for travel and dwell times.
Resulting Publications / Presentations
- Ghanim, M., Dion,
F., and Abu-Lebdeh, G. Artificial Neural Network Bus Arrival Time Prediction
Tool for Transit Signal Priority with Nearside Bus Stops, Proceedings
of the 86th Transportation Research Board Meeting, DC, January 2007
- Dion, F. and Ghanim, M. Impact of Dwell Time Variability on Transit Signal Priority
Performance at Intersection with Nearside Bus Stop, Proceedings of
the 86th Transportation Research Board Meeting, DC, January 2007
|
