Download Wisconsin DOT Travel Time Technology Evaluation (T3E) Analysis Plan and more Study notes Artificial Intelligence in PDF only on Docsity!
Wisconsin DOT Travel Time
Technology Evaluation (T3E)
Analysis Plan
Jonathan Riehl, Transportation Systems Engineer
Peter Rafferty, TSM&O Program Manager
Zhe Xu, Research Assistant
Wisconsin Traffic Operations and Safety (TOPS) Laboratory
University of Wisconsin–Madison
Department of Civil and Environmental Engineering
September 2016
Project Title: Travel Time Technology Evaluation (T3E) Project ID Number: 0072-40- Master Contract Number: 0072-39- Work Order Number: 9. DTSD Big Ticket Number: BTO Object Code: 5501 Funding Appropriation: 365 – Highway system management and operations
Table of Contents
1. Task Introduction
This is a detailed analysis plan to determine how best to compare all of the travel time technologies being studied in the Travel Time Technology Evaluation (T3E).
As part of this analysis plan, a detailed literature review was completed. This review looked at previous studies analyzing travel times. This will include looking at related efforts and past efforts including the 2008 AirSage/INRIX evaluation report, the TOPS Bluetooth traffic detector comparison study completed in 2013, and recent Great Lakes Regional Transportation Operations Coalition (GLRTOC) work with Bluetooth and probe data including work completed in Janesville comparing Bluetooth, probe data, and NPMRDS data. The literature review is included in Appendix B.
Next, specific routes/segments are chosen based on data availability and relevancy to the project. Time periods have also been chosen as appropriate for the comparison.
The process for data source retrieval will be determined for all data sets including:
- Purchased TomTom GPS-based probe data and additional interstate TomTom data;
- The free FHWA National Performance Management Research Data Set (NPMRDS);
- Bluetooth detection maintained by WisDOT or GLRTOC;
- Microwave detection;
- Inductive loops, available via WisTransPortal; and
- Automatic Traffic Recorders (ATRs).
Statistics and metrics are chosen based on the literature review and the adaptation of WisDOT travel time quality assurance, quality control (QAQC) process.
This project does not include field data collection such as travel time runs.
See Appendix A for the project management timeline for this project.
2. Background
The overall purpose of the T3E project is to understand the quality of probe data and appropriate use applications. In conjunction with the I-39/90 expansion project and the Verona Road project, a real time data feed has been purchased by WisDOT with expansion and renewal options up to seven years covering Rock and Dane counties. This evaluation will compare the TomTom data with other travel time calculation technologies to determine which technology is most appropriate. It is possible that certain technologies will work better on different types of highways and in rural/urban areas.
2.1. Reasons for Evaluating Technologies
WisDOT has many dynamic message signs (DMS) stating travel times to aid commuters and other travelers throughout the state in typically congested areas. Roadway users expect that these times are accurate, and if the times are not accurate, users will lose faith in the system. In situations where delays are expected, accurate freeway and alternate route travel times are imperative. This allows drivers to divert onto the alternate route when the route offers a faster travel time, thus maximizing the capacity of the built highway network and minimizing user delay cost.
With the onset of connected vehicles, travel time information can be made available in the vehicle as part of the heads-up display. This will result in roadway users expecting the most precise travel times available in all situations.
In order to provide these travel times, WisDOT is performing this evaluation to
- Compare arterial versus freeway travel times
- Compare long term versus short term travel times (cases such as alternative routes for construction projects).
- Compare costs of acquiring and maintaining data
- Compare difficulty of accessing and processing data sources
- Determine other uses of travel time data
- Integrate technologies into the transportation systems management and operations (TSM&O) decision process for detection
The better WisDOT understands the quality of data available now, the better the accuracy of travel times that will be available now for use on installed DMS and in the near future in the roadway users’ vehicles.
2.2. Existing Travel Times
WisDOT travel time information is currently calculated based on speed data collected by a variety of traffic data detection devices located along a road corridor that is then integrated into the Advanced Traffic Management System software (ATMS) used by WisDOT.
2.4. Other Technologies
Many technologies exist to calculate route travel times. Although some of these are used in this study, there are many that will not be compared. For completion purposes, all major methods are listed here. These are detailed in Section 2 of the Literature Review and summarized here.
2.4.1. Point Sensors
A point sensor measures the presence and speed of vehicles that travel by the location point where the sensor device is deployed. These include loop detectors, microwave detectors, and ATRs. These devices are generally used for volume, speed, and occupancy measurements. However, travel times can be measured between two devices using either the half-distance approach or the minimum speed approach as outlined in the literature review.
2.4.2. Video and License Plate Readers
Travel time can be measured by automatic plate recognition systems (APRs). The measurement requires at least two fixed APR systems on the road. When a vehicle passes by the first APR system, the video recorder of the APR will read its plate number. Then when the same vehicle passes through the second APR system, its plate number will be recorded again. Finally, the server will match the plate numbers and their time stamp tags. By matching the time stamp and measuring the distances between the set of APR systems, the travel time and travel speed of the vehicles could be measured.
2.4.3. Radar
Radar detectors can collect velocity, flows, and occupancy data when they are deployed along the roadside. Since the radar detection is strongly impacted by the road environment, radar is more widely implemented on rural highways rather than in urban areas. Although radar is suitable with massive data collection, the collected data has low accuracy.
2.4.4. Bluetooth
Bluetooth detectors scan the area range and check if any Bluetooth enabled device are detected. Once the vehicle equipped with Bluetooth devices drive into the detection range of a Bluetooth reader, enter and exit time stamps of the devices are recorded. Therefore, travel time and travel speed can be determined between points on the roadway.
The Bluetooth data gives a straight measurement of travel time between pairs of scanners. The data includes the “duration” of time required for the vehicle to pass the range detection limits of the Bluetooth scanner. Thus, Bluetooth data can give the entry and exit timestamp for each of the detectors which provides the duration of each Bluetooth device.
2.4.5. Wi-Fi Technology
Wi-Fi Technology can be used to measure the travel time of vehicles when the location of the probe vehicle and its distance to the next Wi-Fi spot is known. However, the measurement is affected by the noise impacting the localization of the car. Therefore, this technology is accurate enough for route planning, but it does not work well for individual road section estimation.
2.4.6. High-Frequency GPS Data
High-frequency GPS is a method where the probe vehicle can send GPS information every few second or each second (no more than 10 seconds). This aspect makes the data the most accurate for travel time estimation. However, the number of GPS enabled probes may limit its application. There are also some map matching problems for the complex environment such as roundabouts or intersections. This is the general strategy used by providers such as TomTom, Inrix, HERE, Google, and Waze; although they do use a variety of other probe data sources that are proprietary and thus not fully disclosed.
2.5. Current Wisconsin Travel Time Information Sharing and Users
Travel times in Wisconsin are currently available through 511 Wisconsin online and through an XML feed. Access to the 511 site is open to the public. The XML feed is available by subscription with subscribers including media outlets, researchers, and construction project teams. In particular, the Zoo Interchange team in Milwaukee is using travel time records for performance evaluation.
With the onset of connected vehicle technologies, the same travel times disseminated through 511 could eventually be displayed real-time on vehicle’s heads-up display units, which will vastly expand the routes in which travel times are made available.
The Madison Area Transportation Planning Board, Madison’s Metropolitan Planning Organization (MPO), currently is working with WisDOT to obtain Bluetooth travel time information. Research has been conducted at the University of Wisconsin-Madison and is in preliminary phases at the University of Wisconsin-Milwaukee using a combination of WisDOT Bluetooth detectors and detectors used by GLRTOC on DMSs throughout the state on major corridors.
Figure 1. Travel Time Technology Evaluation (T3E) Route Overview Map
3.3 Study Time Periods
To make sure that statistical comparisons are as consistent as possible, specific dates and times have been chosen for the analysis. These dates are limited to the intersection of data availability and thus are different depending on the corridor. Time periods chosen for the study are shown in Table 3.
Specific study time ranges within the chosen time periods will be used and comparisons will be made within the corridor and cross-corridor depending on highway classification. The time ranges used are:
- AM Rush, 7:00am-9:00am (weekdays)
- AM Peak, 7:30am-8:30am (weekdays)
- PM Rush, 3:00pm-6:00pm (weekdays)
- PM Peak, 4:30pm-5:30pm (weekdays)
- Weekday Daytime, 6:00am-6:00pm
- Weekend Daytime, 7:00am-7:00pm
- Nighttime, 10:00pm-4:00am
- Holiday Travel (Memorial Day or Independence Day)
Table 3. Selected Time Periods for Study by Corridor
Corridor Corridor Start/End
Available Period Chosen Periods
US 12/
I-39/90 to WIS 73
04/14/2015 to Present
05/01/2015 to 05/31/2015 and 05/01/2016 to 05/31/ US 14 M (Madison)
US 12/18 to County MM
04/14/2015 to Present
05/01/2015 to 05/31/2015 and 05/01/2016 to 05/31/
County M US 18/151 to County MM
04/14/2015 to Present
05/01/2015 to 05/31/2015 and 05/01/2016 to 05/31/ US 14 J (Janesville)
I-39/90 to WIS 140
04/14/2015 to 11/02/ 05/01/2015 to 05/31/
WIS 73
I-39/90 to WIS 106
04/14/2015 to Present
05/01/2015 to 05/31/2015 and 05/01/2016 to 05/31/ E Washington (US 151)
Blair St to Portage Rd
06/10/2016 to Present 07/01/2016 to 07/31/
I-39/
IL Border to I-
06/05/2015 to Present
07/01/2015 to 07/31/2015 and 07/01/2016 to 07/31/
US 12 I-39/90 to Parmenter St
04/15/2015 to 05/04/ 04/15/2015 to 05/04/
4.1.2. TomTom CTT (Custom Travel Times)
Access Point: TomTom, http://trafficstats.tomtom.com/ Access Settings: Routes, dates, and time sets Interval Size: 15 minutes Dates Available: January 1, 2008, (0:00) - Present Routes Available: Most freeways and arterials as well as some major collectors Link Type: TomTom Segment Identifiers Data Format: Google KML, ArcGIS Shapefile, and Excel Spreadsheet Information Provided: Average/Percentile Speeds, Average/Median Travel Time Data Access Screen: See Figures 3, 4, and 5
Figure 3. Data Access Screen (Routes) for TomTom Custom Travel Time Tool
Figure 4. Data Access Screen (Dates) for TomTom Custom Travel Time Tool
Figure 5. Data Access Screen (Times) for TomTom Custom Travel Time Tool
4.1.4. Bluetooth
Access Point: Drakewell, https://drakewell06.drakewell.com/ Access Settings: Bluetooth units, dates, times Interval Size: 1 minute Dates Available: Route Dependent as shown below
Table 4. Bluetooth Data Availability by Route
Corridor Begin Date End Date Bluetooth Units On Route
US 12/18 05/13/2014 Present
WDS-0029, WDS-0030 2 , WDS-0031, WDS-0032,
WDS-0033 1 , WDS-0130, WDS-0034 1 , WDS-0035,
WDS-0131 2 , WDS-0041, WDS-0044, WDS-0046,
WDS-0047, WDS-0050, WDS-0051, WDS-0052,
WDS-0132 2 , WDS-0053, WDS-0133, WDS-0134,
WDS-0054, WDS-
US 14 M
(Madison) 05/16/2014 Present WDS-0048, WDS-0049, WDS-0078, WDS-
US 14 J
(Janesville) 10/23/2014^ 11/02/2015^ GL-004, GL-017 (old)^
(^3) , GL-014 (old)
E Washington (US 151) 06/10/2016 Present GL-021, GL-014, GL-
I-39/90 06/05/2015 Present
GL-005, GL-019, GL-023, WDS-0001,
WDS-0136 4 , WDS-0135 4 , WDS-0002, WDS-0003,
WDS-0004, WDS-0005, WDS-0006, WDS-0007,
WDS-0008, WDS-0009, WDS-0010, WDS-0012,
WDS-0013, WDS-0014, WDS-0016, WDS-0017,
WDS-0019, WDS-0020, WDS-0021, WDS-0022,
WDS-0023, WDS-0025, WDS-0026, WDS-
US 12 11/19/2014 05/04/2015 GL-021 (old), GL-018 (old) 5 , GL-001 (old)
(^1) Data from these units only available from 11/17/ (^2) Data from these units only available from 05/22/ (^3) Data from this unit only available until 04/03/ (^4) Data from these units only available from 10/22/ (^5) Data from this unit only available from 04/15/
Routes Available: Limited – based on where units are placed Link Type: Latitude/Longitude Points Data Format: Excel Spreadsheet Information Provided: Speed, Travel Time, Match Count Data Access Screen: See Figure 7
Figure 7. Data Access Screen for Bluetooth Data (using Drakewell Online)
4.1.5. ATR (Automated Traffic Recorder)
Access Point: TOPS Lab TRAffic DAta System (TRADAS), http://transportal.cee.wisc.edu/products/hourly-traffic-data/ (Data downloaded and then stored in Oracle database) Access Settings: Traffic site ID, dates, epochs (times) Interval Size: 60 minutes Dates Available: January 1, 2014, (0:00) - Present Routes Available: Limited – based on where units are placed; statewide coverage Link Type: Latitude/Longitude Points Data Format: Comma Separated Value Information Provided: Volume, Speed, Classification Data Access Screen: See Figure 8
Figure 9. Data Access Screen 1 for Microwave/Loop Data (using V-SPOC online)
Figure 10. Data Access Screen 2 for Microwave/Loop Data (using V-SPOC online)
4.2. Travel Time Computation
Travel time computation varies by type of data. The computation steps are described briefly below:
4.2.1. TomTom LTA (Live Traffic Archive)
The most difficult data to access is data from the TomTom Live Traffic Archive tool. This data is served in a protocol buffer format from TomTom. Data is accessed using a Protobuf reader and a .proto decoder file. The software used for accessing this data is a modified version of Record Editor (https://sourceforge.net/projects/protobufeditor/) which is a free, open-source software.
Data from the LTA tool is served for the entire state with limited spatial definitions. Therefore, once data is decoded using Record Editor, data must be extracted to a mappable format. Links are represented in OpenLR format which provides the start and end coordinates. This information must be matched to a roadway segments (preferable on the State Trunk Network (STN) used by MetaManager) to create actual highway links. This process is difficult due to varying lengths of segments by route and a disconnect between these segments and the STN and NPMRDS TMC links. Once this is done once, data can be extracted and matched to these links, assuming no changes in the OpenLR codes. If these codes change, the links would have to be reprocessed.
Data is obtained in one-minute intervals and is not filtered for outliers or confidence. Historic data is available for all routes. Full computation steps will be included in the final report as part of the description of Task 3.
4.2.2. TomTom CTT (Custom Travel Times)
TomTom data from the Custom Travel Times tool is much easier to work with than the LTA data, as the output format provided includes an ArcGIS shapefile and an Excel spreadsheet. Excel data can be joined to the routes provided in the shapefile. For reference of this project, the links provided in the shapefile are adequate, however it is preferable to match these segments to the STN.
Data is obtained in 15-minute intervals and is not filtered for outliers or confidence. Historic data is available for all routes. Full computation steps will be included in the final report as part of the description of Task 3.
4.2.3. NPMRDS (National Performance Management Research Data Set)
The National Performance Management Research Data Set is provided as a CSV file which can be joined to the NPMRDS route map which offers segments geo-referenced to traffic message channels (TMCs) and HERE link IDs. Again, for reference of this project, these links are adequate, however it is preferable to match these segments to the STN.
