Task Order 5307
Traffic Operations Research

Click here to view: TO5307_Nov_28_06.pps


Implementation of a Tool for Measuring
ITS Impacts on Freeway Safety Performance

Thomas F. Golob
Institute of Transportation Studies
University of California, Irvine

Wilfred W. Recker
Civil and Environmental Engineering, Institute of Transportation Studies
University of California, Irvine

Objective

Implement a real-time tool for safety analysis. We intend to calibrate and verify a tool that translates traffic flow, as measured by ubiquitous single loop detectors, into safety performance in terms of expected numbers of crashes by type of crash per exposed vehicle mile of travel. A primary goal of this effort is to provide an easily accessible tool for use in assessing the safety performance of freeway operations and to evaluate and document improvements to safety arising from such ITS deployment as system-wide ramp metering (SWARM), freeway service patrol (FSP) and other incident response measures, and driver information. By quantifying the safety benefits accrued from smooth and efficient traffic operations, Caltrans will be able to incorporate safety measures in assessment of performance gains resulting from ITS deployment. Another application will be to forecast the safety implications of proposed projects by evaluating the levels of safety implied by traffic simulation model outputs. The tool can also be used to forecast the safety consequences of doing nothing.

The traffic flow data we need have been received directly from front-end processors (FEP) using the UCI ATMIS Testbed Intertie with Caltrans District 12. These data are currently being used in two ongoing PATH projects: Development of a Path Flow Estimator for Deriving Steady-State and Time-Dependent Origin-Destination Table Trips, and A Tool for the Incorporation of Non-Recurrent Congestion Costs of Freeway Accidents in Performance Management. Accident data from TASAS will be linked with these traffic flow data. In assessing the levels of safety associated with traffic flow conditions on segments of the major Orange County freeways in 2001-2002, we will be investigating the roles of factors that highway managers can manage or engineers can build into roadways. We will be searching for evidence of how traffic flow can be affected in order to reduce freeway crashes.

We will validate the model using 2002-2003 data for the same region for which the model is calibrated. Using TASAS data from the same year, the model validation will be conducted by comparing the temporal/spatial distribution of the safety performance measures against the actual temporal/spatial clustering of accidents. The goal will be to test the model's ability to distinguish locations and conditions with high accident rates from those with low accident rates. The model transferability will be tested by applying the calibrated model to PeMS loop data for at least one other area in California. Because the methodology does not depend on specific geometric characteristics (directly), but rather is based on the traffic conditions arising from both roadway layout and demand, the goal is to demonstrate that the tool can be readily transferred to any urban freeway that is fully instrumented with loop detectors without the need for extensive calibration. Once validated, code will be developed to deploy the model, first as a stand-alone on the Testbed website using data from the Caltrans District 12 FEP as input, and eventually as a component of PeMS.

The effectiveness of the tool will be tested via both traffic simulation models and field data collected before and after some ITS improvement. For evaluation using field data, we will work with Caltrans District 12 to identify any specific traffic improvement projects that are scheduled to come on line during the course of the project; we will then target those specific areas for before and after studies. We will be able to track the "after" case using the ongoing Testbed facility at UCI. We also intend to demonstrated how the tool can be adapted for use with enhanced traffic flow data from probes, radar, and vehicle re-identification.