Evaluating Robustness of Signal Timings for Conditions of Varying Traffic Flows 2013 Mid-Continent Transportation Research Symposium – August 16, 2013 – 9:30AM – 12:00PM Cameron Kergaye, PhD, PE, PMP Director of Research Utah Department of Transportation
24
Embed
Evaluating Robustness of Signal Timings for Conditions of Varying Traffic Flows
Evaluating Robustness of Signal Timings for Conditions of Varying Traffic Flows. 2013 Mid-Continent Transportation Research Symposium – August 16, 2013 – 9:30AM – 12:00PM Cameron Kergaye , PhD, PE, PMP Director of Research Utah Department of Transportation. Traffic Signal Optimization. - PowerPoint PPT Presentation
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Evaluating Robustness of Signal Timings for Conditions of Varying Traffic Flows2013 Mid-Continent Transportation Research Symposium – August 16, 2013 – 9:30AM – 12:00PM
Cameron Kergaye, PhD, PE, PMPDirector of ResearchUtah Department of Transportation
Traffic Signal Optimization
Should signal timings be optimized for high-than-average traffic counts?
How should signal timing optimization accommodate multiple counts?
Summary
Signal timing plans perform best based on average traffic flows mean, mode, and median when exposed to day-to-day traffic flow variability
Optimizing signal timings for higher traffic demand is better than for lower traffic demand Should be used only when sufficient traffic data are
unavailable.
Introduction Optimization of signal timings is considered to be one
of the most effective tools to improve traffic operations on urban arterials.
However, once traffic signal systems are retimed and implemented, quality of their performance largely depends on day-to-day variability of traffic flows in the field.
As soon as traffic patterns change significantly, performance of the signal timings deteriorates.
Current signal timing practice recommends development of separate signal timing plans for major day-to-day traffic patterns (weekday, weekend, special events, etc.).
Introduction When adaptive traffic control is implemented it is
impractical to develop plans for every traffic pattern that warrants a separate signal timing plan.
Therefore, it is important to develop signal timing plans that will minimize disbenefits of implementing signal timing plans in variable traffic conditions.
Research Background Signal timing plans are based on traffic data that are
usually collected during a short term effort (e.g. 1-week).
The data includes 24-hour weekly volume profiles, turning movement counts, vehicular speeds, and travel time runs.
The data is analyzed with other sources to ensure that they are representative of common field conditions.
Research Background
What is a representative traffic volume pattern? It is the one that generates signal timings that work
best in a variety of traffic conditions.
Finding such a traffic volume pattern and optimizing signal timings in field experiments requires significant resources. Therefore, we use traffic simulation and other methods
that do not require field experiments.
Research Contribution Weekday PM-Peak hour traffic flows from the field are
modeled in microsimulation for a year. Two factors made this modeling approach possible:
Comprehensive set of field traffic flows collected during an entire year
Special tool to validate and balance traffic flows for the model.
Signal timing plans are optimized for each of the representative traffic flows resembling the process that usually occurs in practice.
Each of the signal timing plans is evaluated for the entire set of traffic flows to determine the best representative set of traffic flows.
MethodologyAccurately model day-to-day traffic variations in microsimulation Collect and Process Field Traffic Volumes
Traffic Volumes Recorded by SCATSTraffic Volumes Collected by Automatic Traffic RecordersManually Collected Traffic Volumes
Verify SCATS Traffic Volumes Build, Calibrate, and Validate the VISSIM Model Model Variability of Traffic Flows in VISSIM
Prepare for Modeling Variability of Traffic Flows in MicrosimulationVerify the reasonableness of SCATS traffic volumesBalance traffic flows in the networkVerify VISSIM Traffic Flows
Develop signal timing plans based on traffic flows of representative days Select scenarios of ‘representative-day traffic volumes’ Optimize signal timings in VISGAOST Evaluate Optimized Signal Timings in VISSIM
MIN AVERAGE MODE MEDIAN 75th PERCENTILE 85th PERCENTILE MAX
Cycle Lengths & Offsets
Cycle Lengths, Offsets & Splits
Variations in Performance Indices
4500
5500
6500
7500
8500
9500
10500
11500
12500
200
300
400
500
600
700
800
900
1000
1100
Net
wor
k Tr
affic F
low
[veh
/hou
r]
Perf
orm
ance
Inde
x
DatesMAX MIN EXISTING
AVERAGE MODE 85th PERCENTILE
75th PERCENTILE MEDIAN Total Network Traffic Flow
Performances Based on Various ‘Representative-Days’
2000
4000
6000
8000
10000
12000
14000
0
100
200
300
400
500
600
Net
wor
k Tr
affic T
hrou
ghpu
t [ve
h/ho
ur]
Perf
orm
ance
Inde
x
Optimization Scenarios
Performance Index Traffic Demand
ConclusionsSignal timings optimized for median traffic flows were best. Similar results were found for other average traffic flows (i.e. mean and mode).
Findings show that signal timings developed for traffic flows that most frequently occur in the field bring more benefits than those that are developed for less frequent but higher traffic flows.
Basing signal timings on higher-than-average traffic demand still generates better results than those developed for lower traffic flows.
This is justified where there is a shortage of reliable traffic flow data from the field and when demand is expected to grow significantly.