·To determine the performance of detectors on volume, speed, occupancy, and classification under various traffic and weather conditions
·To determine the best use of detectors for real-time traffic management
Highway
Traffic in the United States has increased 30 percent in the past ten years, and the number of cars on the road is projected to increase by 50 percent in the next decade. As vehicular traffic increases, it becomes increasingly more difficult to collect traffic data accurately in the congestion using traditional sensors such as piezos, road tube or standard inductive loops. As our freeways and urban areas become more congested data quality deteriorates. Traditional sensors do not count accurately at low speeds or with bumper-to-bumper traffic. Traditional inductive loop technology can join vehicles resulting in lower counts on highly congested, high volume roadways. Standard axle sensors can miss axles or miscalculate spacing at low or irregular vehicle speeds resulting in inaccurate vehicle classifications. As a result of all this, States have not been able to provide accurate data to the FHWA for major arterials and urban freeways, especially at peak times because the traditional sensors are unable to cope with recurring congestion due to sensor saturation. In an effort to find a better alternative for traffic monitoring, the Rhode Island Department of Transportation (RIDOT) and the University of Rhode Island (URI) have decided to combine effort and resources to install, test, research and evaluate several different vehicle detector technologies. The sensors shall be installed on the same stretch of roadway and evaluated over time to see which technologies perform better for their designated purpose.
TASK 1. STUDY DESIGN
The first task of the project consists of developing a study plan that will identify the specific questions that the research aims to answer. The type of data collected in the study will consist of vehicle speed,volume, occupancy, and classification along with “exogenous” data such as geometric features of the roadway, roadway classification, and the speed limit. At the outset of the project, a candidate list of locations will be determined. The ideal situation is to place one set of detection devices in South County and one set in the Providence metropolitan region. Data will be collected and reviewed on candidate locations to determine suitability for inclusion in the study.
As a part of the study design, a literature review is to be conducted to review testing procedures that have been done previously and to investigate the best performance measures and statistical procedures.
TASK 2.. Data Collection
At the designated locations, RIDOT will deploy equipment consisting of the detection device as well as the necessary communications connections. It is expected that the data will be gathered via a telephone line connection via a modem. This information will either go directly to RIDOT or come to the University, depending on what is most feasible. For the ground-truthing portion of the research, video data will be captured. This will provide the basis for comparison for the detection devices. Determining the logistics for changing video-tape or streaming it will be determined explicitly in the study design portion of the research.
TASK 3. DATA REDUCTION
In a “data reduction” step, the output from the various devices will be segmented by its specific ability, whether it is vehicle classification, vehicle volumes, or vehicle speeds. Data will be characterized by
· roadway/direction of travel
· weekday/weekend-day,
· hourly volume (or smaller designation if needed)
· vehicle classification (per vehicle type)
· speeds (mph or kph, depending on the device)
. occupancy
Additionally, as the data is reduced, there will be room for comments that can be to contain any special notes by the reviewer.
TASK 4. DATA ANALYSIS
The data analysis will come down to making comparisons versus the ground-truth data as to the accuracy of the data. Most devices tend to perform well under average traffic conditions, but it is the peak hour or peak periods, when data is most important. Because of this, statistical analysis will indicate the accuracy by time of day or by non peak-period versus the peak hour or peak period.
Distributions of vehicle counts, speeds, or classification will be tabulated and plotted versus various factors can be easily accomplished through the use of Excel’s built-in data analysis and charting functions or some other appropriate statistical package.
TASK 5. Draft & Final Report
A final report will be provided to the URITC and RI DOT documenting the results of the data collection and analysis for the sites specified, using the data to address the questions laid out in the study design phase of the project. The report will include a review of data detection and studies that deal with investigation various traffic detection technologies. Additionally, the report will include a set of conclusions and recommendations that address issues of data reliability and suggested uses of certain detection technologies.
Tasks and Proposed End Dates
Task 1. Study Design 9/30/2002
Task 2. Data Collection 4/31/2003
Task 3. Data Reduction 5/31/2003
Task 4. Data Analysis 6/15/2003
Task 5. Draft Final Report 6/30/2003
Task 6. Final Report 9/30/2003
$241,872.00
1 Graduate Research Assistant and possibly students from Transportation Engineering courses
This project relates most closely to the overall development of the Rhode Island Intelligent (RIIR) and the Transportation System Analysis Laboratory (TSAL). Also, there would be a linkage to the work done with the group headed by Dr. Peckham, which is investigating the ability to diagnose traffic anomalies through real-time vehicle detection. Ideally, the discovery of the effective detector types would lead to what would be most beneficially for deployment in capturing data on the RIIR with its linkage to the TSAL. Also, the URI TC is working to build its knowledge base about detectors.
We would plan to submit an article to a transportation journal, as well as submit a paper to the Transportation Research Board. A final report will also be submitted to the URITC, which will be accessible through its web-site or via the center in hard-copy form.
The application of the research relates to more effective use of technology with regard to data collection for transportation agencies and municipalities. The information learned from this also could also be applied to the Rhode Island Intelligent Road as it moves from a logical architecture / physical architecture development to one specifying certain devices. The knowledge gained from this research can immediately be transferred into the transportation engineering classes, such as Transportation Engineering (CVE 346), Highway Engineering (CVE 447), Intelligent Transportation Systems (CVE 442), and Traffic Systems Operations (CVE 542). This research effort should provide good information for technology transfer in general.
Vehicle detection; ITS; data-collection; video detection; acoustic sensors; micro-loops; radar detection; fiber optics