Using Archived Data to Measure Operational Benefits of a Systemwide Adaptive Ramp Metering SWARM Sys

1
Evaluation the Data Accuracy of Loop Detectors in
Adaptive Signal Control Systems
Meead Saberi K., Portland State University Vahab
Esmaeli, Ferdowsi University of Mashhad, Iran Ali
Hatami, Ferdowsi University of Mashhad, Iran
2
Objective

• How accurate are the loop detectors data (count
  data)?
• Is there any relationship between traffic volume
  and error rate of count data?
• What are the reasons for error?

3
Study Area
IRAN

• Khayyam Malek Abad intersection
• Modarres Emam Khomeini intersection
• Sanabad Kolahdooz intersection
• Rah Ahan intersection
• Sajjad Bozorgmehr intersection
• Daraie Intersection

6 main urban intersections in Mashhad city (white
square points on the map)
4
Data Quality
Data quality is the fitness of data for all
purposes that require it. Measuring data quality
requires an understanding of all intended
purposes for that data Turner, S., Defining and
Measuring Traffic Data Quality Texas
Transportation Institute Cambridge Systematics
Inc, 2002.
Accuracy Completeness Validity Timeliness
Coverage Accessibility
5
Methodology
Study approaches were categorized into approaches
with two, three and four lanes in which one loop
detector installed in each lane. Archived
traffic volume data, aggregated at 15 minutes
intervals, of each approach during the study
period were used. Recorded videos during the
same periods were processed manually by two
persons to count the actual number of passenger
cars, heavy vehicles, motorcycles, and bicycles
pass the detectors.
6
Error Ratio
Actual traffic volume
Detected traffic volume
Error ratio is positive which means the actual
traffic volume is smaller than the volume which
loops are detected .
Error ratio is negative which means the actual
traffic volume is greater than the detected
number by loop detectors.
7
Reasons of Error
8
Reasons of Error
9
Reasons of Error
10
Reasons of Error
11
Reasons of Error
12
Volume vs. Error (4 lanes)
13
Volume vs. Error (4 lanes)
14
Volume vs. Error (4 lanes)
15
Volume vs. Error (4 lanes)
16
Volume vs. Error (3 lanes)
17
Volume vs. Error (3 lanes)
18
Volume vs. Error (2 lanes)
19
Hypothetical Relationship
Volume vs. Error
20
Next Step
Expanding the sample size to prove the
relationship between volume and error
21
Acknowledgement

• Traffic Control Center of Mashhad
• Ferdowsi University of Mashhad, Iran
• STEP, Portland State University

22
Questions?
Thanks