Empirical Observation of the Impact of Traffic Oscillations on Freeway Safety

Traffic oscillations (also known as stop-and-go driving) are a typical feature of congested traffic flow. They are known to increase fuel consumption and emissions, and decrease driving comfort. It is also speculated that larger amplitudes of oscillations (i.e. larger changes in flow or speed) increase the probability of certain crash types (e.g. rear-end crashes). However, no current study exists that irrefutably confirms or disproves this speculation. The objective of this research is to find empirical evidence to substantiate this hypothesis and to quantify the relationship between the amplitude of oscillations and probability of crash event. This proposed research will be…

Traffic oscillations (also known as stop-and-go driving) are a typical feature of congested traffic flow.  They are known to increase fuel consumption and emissions, and decrease driving comfort. It is also speculated that larger amplitudes of oscillations (i.e. larger changes in flow or speed) increase the probability of certain crash types (e.g. rear-end crashes). However, no current study exists that irrefutably confirms or disproves this speculation. The objective of this research is to find empirical evidence to substantiate this hypothesis and to quantify the relationship between the amplitude of oscillations and probability of crash event.

This proposed research will be conducted using freeway traffic and incident data. It will be supplemented by a statewide database of reported motor vehicle crashes. Various features of oscillations (e.g. amplitude, period, etc.) will be measured from traffic data collected from inductive loop detectors. Existing databases for crashes and incidents will be used to analyze incidents in correlation with oscillations. All the data necessary for this research are available via Portland Oregon Regional Transportation Archiving List (PORTAL), which provides several years of archived data and a wealth of supporting data and statistics. Use of this archive will minimize the research team’s data collection and processing efforts. The large archive will increase the chances of resulting in findings that are meaningful and statistically significant.

This study will consist of two primary stages: (1) general analysis to identify which crash types are particularly affected by traffic oscillations and (2) detailed analysis via econometric modeling to quantify the probability of each crash type as a function of various characteristics of oscillations and relevant factors such as freeway geometry, congestion level, and others. These analyses will be conducted for several freeway locations in order to confirm reproducibility and to examine any site-specific features.