Transcript Presentation - 15th TRB National Transportation Planning

Southeast Michigan Council of Governments
Development of a Traffic
Count Database for Travel
Model Calibration
Brian D. Mohr, Li-yang Feng, and Tom Bruff
Southeast Michigan Council of Governments
11th TRB Applications Conference
Daytona Beach, FL
May 7, 2007
*Detroit
SEMCOG Region
Oakland
Livingston
St. Clair
Macomb
Communities: 234
Washtenaw
Wayne
Monroe
Population:
4.9 million
Licensed drivers:
3.4 million
Annual VMT:
49 billion
Miles of road:
23,000
Presentation Overview
Count
source
Count
source
Count
source
Regional
traffic count
database
Travel model
interface
Model
validation
files
Challenges
• No consistency in count data format
• No central database for storing counts
• No efficient procedure for linking counts
with model
• No way to determine if counts sufficient
for validation
Solutions
• Create regional traffic count database
(RTCD)
– Central repository for all counts
– Source of counts for model, other apps
• Create automated procedure for linking
counts with model
• Use power of GIS (linear referencing)
RTCD Development Issues
• Transferring data from multiple sources
into common format
– Michigan Department of Transportation
(MDOT)
– Local agencies (counties, cities)
• Developing RTCD database structure
• Coding counts with GIS attributes
Traffic Count Database
System (TCDS)
• Software features
– Web-based, friendly user interface
– Direct uploading from traffic counters
– Logical relational database structure
– Numerous quality control measures
• Provides SEMCOG easy access to
counts from all seven counties, Detroit
Importing MDOT’s Traffic
Count Data
• Challenges
– Understanding MDOT’s count program
– Three types of data: PTR, ITS, short counts
– Different data format for each type
• Solution: MDOT data conversion tool
– Converts all data into common format
– Flags “representative” PTR, ITS counts
MDOT Data Conversion Tool
SEMCOG’s Regional Traffic
Count Database (RTCD)
• Oracle database with three primary tables:
– Count location descriptions
– Counts
– GIS (linear referencing) coding
• Number of records in database
– Locations: 28,080
– Counts available for validation: 58,750
– Total counts available: 342,049
GIS (Linear Referencing)
Coding
• Counts coded with LRS attributes from
base map (route/milepoint system)
• All SEMCOG transportation data coded
• Counts coded as segments
Every point ( ) on road network represented
by two coordinates: route number, milepoint
Every count location (
) represented by
three coordinates: route number, beginning
milepoint, ending milepoint
Local agency
counts
(via TCDS)
Other
counts
Regional
Traffic Count
Database
(RTCD)
MDOT counts
(via conversion
Tool)
Local agency
counts
(via TCDS)
Other
counts
Regional
Traffic Count
Database
(RTCD)
MDOT counts
(via conversion
Tool)
Travel model
interface
Model
validation
files
Model Validation File
Creator
• Required input
– RTCD tables (locations, counts, GIS info)
– TransCAD link definition table
– Count alignment table
– Desired statistical parameters
• Output: data-based or model-based
validation file
Two Types of Model
Validation Files
• Data-based
– Links assigned counts wherever possible
– Not a statistically sound validation file
• Model-based (statistical sample)
– Links stratified by geographic area and
functional class, sample randomly selected
– Only some links will have counts assigned
Model Validation File
Creator
Model Validation File
Creator
Matching Counts and Links
Using Linear Referencing
• Model links also coded with LRS
attributes from GIS base map
• Allows geographic comparison of
count locations and model links
Multiple Counts at Same
Location
• Count attributes used to match counts
with appropriate model links
• Hierarchy of preferred counts developed
• Examples of preferred counts
– Counts taken in same year as model
validation year
– Directional counts
– Counts with shorter intervals
Results
• Entire process works successfully
– Already used on 2002, 2005 highway
networks in current model
– Will be used soon with 2006 counts
• RTCD data used in many transportation
applications, including web data tool
www.semcog.org
Future Initiatives
• Continue to improve count data quality
• Append data from additional sources
into RTCD
• Fully automate importing of data into
RTCD
• Begin performing model validation with
statistically generated sample of links
Keys That Made This
Project Successful
• Establishing data formatting consistency
• Effective communication, coordination
with local agencies and MDOT
• Establishing data quality controls
throughout process
• Use of GIS (linear referencing)
• Automating tasks wherever possible
Local agency
counts
(via TCDS)
SEMCOG
website
Data-based
validation
files
Other
counts
Regional
Traffic Count
Database
(RTCD)
MDOT counts
(via conversion
tool)
Model
Validation
File Creator
Model-based
validation
files
Other
transportation
applications
Development of a Traffic
Count Database for Travel
Model Calibration
Brian D. Mohr, Li-yang Feng, and Tom Bruff
Southeast Michigan Council of Governments
11th TRB Applications Conference
Daytona Beach, FL
May 7, 2007