Description

This report describes the development and evaluation of real-time crash risk-assessment models for four freeway corridors: U.S. Route 101 NB (northbound) and SB (southbound) and Interstate 880 NB and SB. Crash data for these freeway segments for the 16-month period from January 2010 through April 2011 are used to link historical crash occurrences with real-time traffic patterns observed through loop-detector data. 'The crash risk-assessment models are based on a binary classification approach (crash and non-crash outcomes), with traffic parameters measured at surrounding vehicle detection station (VDS) locations as the independent variables. The analysis techniques used in this study are logistic regression and classification trees. Prior to developing the models, some data-related issues such as data cleaning and aggregation were addressed. The modeling efforts revealed that the turbulence resulting from speed variation is significantly associated with crash risk on the U.S. 101 NB corridor. The models estimated with data from U.S. 101 NB were evaluated on the basis of their classification performance, not only on U.S. 101 NB, but also on the other three freeway segments for transferability assessment. It was found that the predictive model derived from one freeway can be readily applied to other freeways, although the classification performance decreases. The models that transfer best to other roadways were determined to be those that use the least number of VDSs–that is, those that use one upstream or downstream station rather than two or three.\ The classification accuracy of the models is discussed in terms of how the models can be used for real-time crash risk assessment. The models can be applied to developing and testing variable speed limits (VSLs) and ramp-metering strategies that proactively attempt to reduce crash risk.

Publication Date

5-1-2012

Publication Type

Report

Topic

Miscellaneous Transportation Topics

MTI Project

1006

Keywords

Real-time crash risk, Data mining, Classification tree, Proactive traffic management, Loop detector data

Disciplines

Transportation

Share

COinS