Kun-Feng Wu

Crashes and crash-surrogate events: Exploratory modeling with naturalistic driving data

There is a need to extend and refine the use of crash surrogates to enhance safety analyses. This is particularly true given opportunities for data collection presented by naturalistic driving studies. This paper connects the original research on traffic conflicts to the contemporary literature concerning crash surrogates using the crash-to-surrogate ratio, π. A conceptual structure is developed in which the ratio can be estimated using either a Logit or Probit formulation which captures context and event variables as predictors in the model specification. This allows the expansion of the crash-to-surrogate concept beyond traffic conflicts to many contexts and crash types. The structure is tested using naturalistic driving data from a study conducted in the United States (Dingus et al., 2005). While the sample size is limited (13 crashes and 38 near crashes), there is reasonable correspondence between predicted and observed crash frequencies using a Logit model formulation. The paper concludes with a summary of empirical results and suggestions for future research.

Analysis of Naturalistic Driving Event Data

Naturalistic driving studies have been conducted over the past 5 years or more and have commonly reviewed video and kinematic data to identify and analyze crash, near-crash, and critical-incident events. But statistical methods that are applicable to these event data are needed. This paper addresses two issues in model development for naturalistic driving event data: the test for omitted-variable bias and the exploration of the advantages of hierarchical model structures in data analysis. With roadway departure event data from the 100-Car Naturalistic Driving Study conducted at Virginia Tech Transportation Institute, Blacksburg, Virginia, logit models were used to estimate the probability that a crash or a near crash would occur, rather than a critical incident. The models indicated a substantial omitted-variable bias for estimation of the effect of context variables but little difference for driver variables. These tests indicated that modeling of naturalistic event data should have included variables that described the attributes of the event, the driver, and the context to reduce the likelihood of bias. Hierarchical model structures offer the advantage of driver-level predictors to parameterize the effects of event attributes and contexts. The models thus reflect how driver decisions are executed: drivers with particular characteristics (one level) find themselves in contexts in which they execute specific driving maneuvers (second level), which lead to certain outcomes. Suggestions for further research include testing with additional data sets and potential applications to analysis of crash surrogates.

Exploring the Association between Traffic Safety and Geometric Design Consistency Based on Vehicle Speed Metrics

AbstractPast design consistency research has demonstrated the relationship between operating speeds and geometric design features on two-lane rural highways. However, little is known about the relationship between geometric design consistency and traffic safety. In this study, design consistency is referred to as the difference between operating speed and inferred design speed, and design consistency density is measured to account for the effect of elements upstream and downstream of the study element. To perform the design consistency–safety evaluation in the present study, geometric design, roadway inventory, crash, and operating speed data were collected along two case-study highways in central Pennsylvania (U.S. 322 and PA 350). Several count regression model formulations were used to explore the statistical association between design consistency and total crash frequency. A statistically significant positive association between geometric design consistency and safety was found. Design consistency sur...

Exploring the application of latent class cluster analysis for investigating pedestrian crash injury severities in Switzerland

One of the major challenges in traffic safety analyses is the heterogeneous nature of safety data, due to the sundry factors involved in it. This heterogeneity often leads to difficulties in interpreting results and conclusions due to unrevealed relationships. Understanding the underlying relationship between injury severities and influential factors is critical for the selection of appropriate safety countermeasures. A method commonly employed to address systematic heterogeneity is to focus on any subgroup of data based on the research purpose. However, this need not ensure homogeneity in the data. In this paper, latent class cluster analysis is applied to identify homogenous subgroups for a specific crash type-pedestrian crashes. The manuscript employs data from police reported pedestrian (2009-2012) crashes in Switzerland. The analyses demonstrate that dividing pedestrian severity data into seven clusters helps in reducing the systematic heterogeneity of the data and to understand the hidden relationships between crash severity levels and socio-demographic, environmental, vehicle, temporal, traffic factors, and main reason for the crash. The pedestrian crash injury severity models were developed for the whole data and individual clusters, and were compared using receiver operating characteristics curve, for which results favored clustering. Overall, the study suggests that latent class clustered regression approach is suitable for reducing heterogeneity and revealing important hidden relationships in traffic safety analyses.

Analysis of Existing Data: Prospective Views on Methodological Paradigms

A large component of the safety research undertaken in the second Strategic Highway Research Program (SHRP 2) is aimed at reducing the injuries and fatalities that result from highway crashes. Through a naturalistic driving study (NDS) involving more than 3,000 volunteer drivers, SHRP 2 expects to learn more about how individual driver behavior interacts with vehicle and roadway characteristics. In anticipation of the large volume of data to be collected during the NDS, several projects were conducted to demonstrate that it is possible to use existing data from previous naturalistic driving studies and data from other sources to further the understanding of the risk factors associated with road crashes. More specifically, the four S01 projects, entitled Development of Analysis Methods Using Recent Data, examined the statistical relationship between surrogate measures of collisions (conflicts, critical incidents, near collisions, and roadside encroachment) and actual collisions. This report presents the results of one of these projects, undertaken by Pennsylvania State University. It documents the second phase of a two-phase project under SHRP 2 Safety Project S01B.

Effects of Hours of Service and Driving Patterns on Motor Carrier Crashes

There is a need to explore the relationship, if any, between the probability of a crash and the hours worked by truck drivers. The need arises from the continued adjustment of federal hours of service regulations for truck drivers. This research used data logs from less-than-truckload carrier operations in 2004 to 2005 and in 2010 to estimate the probability of a crash after a certain amount of time spent driving, given no crashes until that time. Driver logs for 7 days before each crash were used and compared with a random sample (two drivers) of drivers who did not crash and were selected from the same company, terminal, and month. This study involved 686 subjects, including 224 crash-involved drivers. Discrete-time survival analysis models indicated a consistent increase in crash odds as driving time increased beyond the fourth hour. Breaks from driving reduced crash odds by as much as 50% compared with situations of drivers with no breaks. Crash odds were lowest when drivers returned to work during the day without an immediately preceding extended recovery period (but with at least minimum required off-duty time). Drivers returning to work immediately after a 34-h recovery period had crash odds 50% to 150% higher than those for drivers without the recovery immediately before a trip. Drivers had the highest crash odds immediately after returning from the extended time off; the effect then diminished with time.

Examining fatal crash reductions by first harmful events since the introduction of the Federal Highway Safety Improvement Program

The federal Highway Safety Improvement Program (HSIP) has been associated with the reduction in fatal crashes since 2006, but the reasons for the reduction remain largely unknown. This paper examines the reduction in fatal crashes in terms of different types of first harmful events that can provide insight into crash causes and prevention strategies. In this study, fatal crashes were categorized into four types: overturn, collision with motor vehicle in transport, collision with fixed object, and collision with nonmotorist. Fixed-effects and mixed-effects Poisson models were used to estimate the magnitudes of fatal crash reduction by first harmful events for each state. Fatal crashes due to collisions with nonmotorists and motor vehicles in transport have been reduced by 10% and 5.3%, respectively, compared with the 2001 to 2005 period. Fatal crashes due to overturn and collision with a fixed object decreased in some states but remained unchanged or increased in other states. Nevertheless, the numbers of national fixed-object and overturn fatal crashes have been reduced by 3% and 0.7%, respectively, as a whole. This study also investigated possibilities that could be associated with the magnitudes of the reductions, for example, the different traffic laws among states. It was found that although different safety improvement projects were implemented to target the various types of crashes, the improvements were also likely to be beneficial to other crash types. These are referred to as spillover effects. Nationally, fatal crashes have decreased since the introduction of the HSIP partly because of the reduction in fatal crashes due to collisions with nonmotorists and motor vehicles in transport and partly because of spillover effects.

An evaluation scheme for assessing the effectiveness of intersection movement assist (IMA) on improving traffic safety

ABSTRACT Objective: Intersection movement assist (IMA) has been recognized as one of the prominent countermeasures to reduce angle crashes at intersections, which constitute 22% of total crashes in the United States. Utilizing vehicle-based sensors, vehicle-to-vehicle (V2V), and vehicle-to-infrastructure (V2I) communications, IMA offers extended vision to provide early warning for an imminent crash. However, most of IMA-related research implements their methods and strategies only in simulations, test tracks, or driving simulator studies that have quite a few assumptions and limitations and hence the effectiveness evaluations reported may not be transferable or comparable. Methods: This study seeks to develop a generalized evaluation scheme that can be used not only to assess the effectiveness of IMA on improving traffic safety at intersections but to facilitate comparisons across similar studies. The proposed evaluation scheme utilizes the concepts of traffic conflict in terms of time-to-collision (TTC) as a crash surrogate. This approach avoids the issue of having insufficient crash frequency data for system evaluation. To measure the effectiveness of IMA on reducing traffic conflicts, a relative risk is calculated for comparing the risk of with/without using the IMA. As a proof-of-concept study, this study applied the proposed evaluation scheme and reported the effectiveness of IMA on improving traffic safety in a field operation test (FOT). Seven test scenarios were conducted at 4 intersections, and a total of 40 participants were recruited to use the IMA for 6 months. Results: It was estimated that IMA users have 26% fewer conflicts with TTC less than 5 s and have 15% fewer conflicts with TTC less than 4 s. However, the results vary across different sites and different definitions of conflicts in terms of TTC. Conclusions: Overall, IMA is promising to effectively reduce angle crashes related to sight obstruction and has potential to reduce not only crash frequency but crash severity.

Evaluation of Effectiveness of the Federal Highway Safety Improvement Program

The Safe, Accountable, Flexible, Efficient Transportation Equity Act: A Legacy for Users established the Highway Safety Improvement Program (HSIP), which authorized about

Method for the Use of Naturalistic Driving Study Data to Analyze Rear-End Crash Sequences

1.3 billion/year from 2006 to 2009 for highway safety projects. The HSIP aims “to achieve a significant reduction in fatalities and serious injuries on all public roads,” and the number of national traffic fatalities seems to have decreased at about the same time. This study sought to evaluate the effectiveness of the HSIP in reducing fatal crashes in the United States. The study adopted fixed-effect panel models and multilevel mixed-effect models to deal with random fluctuations both before and after introduction of the HSIP and state-specific effects. The results show a drop of about 7.5% in national traffic fatalities since introduction of the HSIP compared with the average for 2001 to 2005, but the magnitude of reduction varied by state. States’ safety-related spending did not increase after introduction of the HSIP. Increased federal safety funding was offset by reduced state funding (crowd-out effect). The magnitude of states’ fatal-crash reduction was highly associated with years of available crash data, prioritizing method, and use of roadway inventory data. Moreover, states that prioritized hazardous sites by using more detailed roadway inventory data and the empirical Bayes method had the greatest reductions; all of those states relied heavily on the quality of their crash data systems. This study found that effectiveness of the HSIP in reducing national fatal crashes is very likely attributable to mandated reporting requirements, which helped states allocate safety spending more effectively and efficiently. It also suggests that more consistent and reliable crash data will allow states to employ more sophisticated prioritization methods and make better highway safety investment decisions.