Risk from Left Turns Across Multiple Lanes to Bicyclists in a Separated Path: A Case for Protected-Only Left Turns
Ray (Mohammad) Saeidi Razavi, Northeastern UniversityShow Abstract
Peter Furth, Northeastern University
At signalized intersections, permitted left turns (i.e., on a green ball, after yielding) across multiple through lanes and across a separated bike lane or bike path present a substantial threat to bicyclist safety. A conflict study of two such intersections found that when cyclists cross while one or more vehicles is waiting to turn left and there is no opposing through traffic to block it, the chance of a motorist yielding safely (i.e., waiting in the left turn lane) was only 9%, and the chance of their yielding at all – including yielding only after beginning the turn, then stopping in the opposing through lanes – was still only 37%. Non-yielding rates were worse for bikes arriving during green, for bikes riding on the right side of the road and therefore facing a left cross conflict, and for bikes facing a queue with multiple left turning vehicles. Of 112 cyclists who arrived on green when there was at least one left-turning car but no opposing through traffic blocking it, 73 had to slow or stop to avoid a collision. While these conflicts could be essentially eliminated using protected-only left turn phasing (turn on green arrow), existing criteria prefer permitted left turns to reduce vehicular delay. A case study shows how, by considering multiple signalization alternatives, it can be possible to convert left turns to protected-only phasing without imposing a substantial delay burden on vehicles or other road users. Keywords: Bicycle Safety, Left Turn Phasing, Permitted Turns, Signalized Intersections
Spatial Models for High-Accurate Hot Zone Identification for E-bikes
Xuesong Wang (email@example.com), Tongji UniversityShow Abstract
Yongfeng Tang, Tongji University
Mohammed Quddus, Loughborough University
Qingya Zhou, Guangzhou Urban Planning and Design Survey Research Institute
Xinyi Hou, Traffic Police Headquarters of Shanghai Public Security Bureau
The use of electric bicycles (e-bikes) has growing rapidly in recent decades, resulting in a significant rise in e-bike crash rates. Therefore, improving the safety of e-bikes is essential to maintain the rapid growth of this sustainable mode of transport. One policy prescription is to focus on area-wide traffic safety management by identifying hot zones for e-bike crashes and their contributory factors. In macro-level safety modelling, the primary explanatory variable – the area of a zone, vary greatly in sizes. This phenomenon poses realistic challenges to the accuracy of macro-level statistical models. Yet, limited studies examined the influence of discrepancy in zone sizes on macro-level safety analysis related to e-bikes. To fill the research gap, this study aims to examine the influence of the size of an area on macro-level modelling. Spatial data on e-bike crash, road network, land use and socio-economic for 213 administrative units in Shanghai were collected. Then three Poisson log-normal Conditional Autoregressive models were developed with different modelling strategies to address the impact of area scale. In Model 1, area was modelled as a regular independent variable. While in Model 2, area was considered as an exposure variable. Finally, in Model 3, independent variables and dependent variable were divided by area. The results indicated Model 2 outperforms other two models. To identify hot zones, Potential for Safety Improvement estimates of three models were aggregated separately. The findings from this study can provide guidelines in considering the influence of area scale in macro-level modelling and hot-zone identification.
Geographically Weighted Poisson Regression under Linear Model of Coregionalization Assistance: Application to a bicycle crashes study
Shujuan Ji, Chang'an UniversityShow Abstract
Yuanqing Wang (firstname.lastname@example.org), Chang'an University
Yao Wang, Chang'an University
ABSTRACT Cycling benefits individuals and society. However, cyclists are vulnerable road users and its safety concern arises macro-level spatial crash studies. This study intends to investigate the spatial effects of population, land use and bicycle lane infrastructures on the bicycle crashes. This was done by dealing with the issue of spatial correlation and spatial non-stationary simultaneously by developing semi-parametric Geographically Weighted Poisson Regression (sGWPR) model. It is a model combined both constant and geographically varying parameters. For determining which parameters are fixed and non-stationary, previous study suggested monitoring Akaike Information Criterion (AICc) to make decision whether a parameter should vary geographically or not. Yet, only relying on AICc might bury some spatial associations. In this study, we propose Linear Model of Coregionalization (LMC) to assist the decision. Here, we use bicycle crash data across the metropolitan area of Greater Melbourne to establish sGWPR models suggested by AICc and LMC. Comparing the two sGWPR models, we found the sGWPR model under LMC results has better performance, and 30% improvement in the mean squared prediction error (MSPE). It also uncovers more details about spatial relationship between bicycle crashes and bicycle lane intersection density (BLID), which is not suggested under AICc results. The parameters of BLID, a new measurement of bicycle lane facilities proposed by us, are positive and vary over space in majority analysis zones in Greater Melbourne. Keywords: semi-parametric GWPR, spatial non-stationary, spatial correlation, macro-level
Grouped random parameter multinomial logit for studying the influence of traffic, geometric and context variables on urban crash types
Paolo Intini (email@example.com), Politecnico di BariShow Abstract
Nicola Berloco, Politecnico di Bari
Achille Fonzone, Edinburgh Napier University
Grigorios Fountas, Edinburgh Napier University
Vittorio Ranieri, Polytechnic University of Bari
Pasquale Colonna, Politecnico di Bari
Previous research has shown that different factors may influence the occurrence of crashes of different types. In this study, a dataset including information from crashes occurred at segments and intersections of urban roads in Bari, Italy was used to estimate the likelihood of occurrence of various crash types. The crash types considered are: single-vehicle, angle/head-on, rear-end and sideswipe. Models were estimated through a mixed logit structure considering various crash types as outcomes of the dependent variable and several traffic, geometric and context-related factors as explanatory variables (both site- and crash-specific). To account for systematic, unobserved variations among the crashes occurred on the same segment or intersection, the grouped random parameters approach (estimating segment- or intersection-specific parameters) was employed. This approach allows assessing the variability of results across the observations for individual segments/intersections. Segment type, bus lanes were included as explanatory variables in the model of crash types for segments. Traffic volume per entering lane, total entering lanes, total number of zebra crossings, balance between major and minor traffic volumes at intersections were included in the model of crash types for intersections. Area type was included in both segment and intersection models. Typical traffic at the moment of the crash and the period of the day were included in both segment and intersection models. Significant variations in the effect of several predictors across different segments or intersections were identified. The applicability of the study framework is demonstrated, in terms of identifying high-risk or anomalous sites with respect to specific crash types.
The Role of Lighting Condition Across the Pedestrian Injury Severity Spectrum
Nick Ferenchak (firstname.lastname@example.org), University of New MexicoShow Abstract
Risa Gutierrez, WSP
Patrick Singleton, Utah State University
Pedestrian fatalities in the United States increased 53% from 2009 to 2018. 89% of those additional fatalities occurred in the dark. Have similar increases occurred across the pedestrian injury severity spectrum? Has lighting condition had a similarly strong relationship with those outcomes? We analyzed pedestrian fatalities, serious injuries, and minor injuries that occurred in Pennsylvania from 1999 to 2018 using linear regressions and t-tests to explore the strength and statistical significance of trends. Findings suggest that all pedestrian injury severities have experienced increases over the last decade, although increases were sharper for less severe injuries. For each severity level, pedestrian crashes increased faster at night than during the day. However, this difference in trend is most noticeable for pedestrian fatalities – deaths are increasing in the dark but not in daylight conditions – whereas dark and daylight pedestrian injuries are increasing at more similar (and larger) rates. A pedestrian injured in the dark is 2.5 times more likely to be killed than a pedestrian injured during the day. Morbidity and mortality outcomes have gotten worse, with injured pedestrians now less likely to be minorly injured and more likely to be seriously or – in the dark – fatally injured. A lack of street lighting does not seem to be the cause of the disproportionate increase in pedestrian injuries at night, since collisions of all severity levels with a street light increased at higher rates than those without a street light.
A Comparison of Pedestrian Injury Severity Crash Factors at Intersections and Non-Intersection Locations
Zade Koch, Massachusetts Institute of Technology (MIT)Show Abstract
Jinhua Zhao, Massachusetts Institute of Technology (MIT)
Sirisha Rangavajhala, Charles Stark Draper Laboratory Inc
Peter Miraglia, Charles Stark Draper Laboratory Inc
Despite the frequent application of data-mining techniques to pedestrian injury severity analysis, few studies have compared the magnitude of injury severity factors between crash location types. In this study, the magnitude of pedestrian injury factors are compared between intersections and non-intersection locations. Data was taken from NHTSA's 2016-2018 Crash Report Sampling System (CRSS) database, containing a nationwide sample of vehicle-pedestrian crashes. Two logit models were created using 7 independent variables: weather, lighting condition, speed limit, speeding violation, vehicle body type, driver impairment, and pedestrian age. One model was fit to 4,828 injured pedestrians at intersections. The other model was fit to 4,663 injured pedestrians at non-intersection locations. The average marginal effect of each crash factor on the probability of severe injury was calculated for both models. The difference of marginal effects was examined to determine if the factors influencing injury severity vary in magnitude at intersection or non-intersection locations. The difference between the two location types for nearly all factors was not statistically significant. This suggests that the 7 independent variables influence injury severity similarly at both intersections and non-intersection locations. The lack of a statistically significant difference may be due to limitations in the data source and warrants further investigation.
How Built Environment Characteristics of Pedestrian-Vehicle Collision Locations Affect Pedestrian Injury Severity Involving Distracted Driving?
Nazmul Arefin Khan, Dalhousie UniversityShow Abstract
Muhammad Habib, Dalhousie University
This study develops an injury severity model that demonstrates level of pedestrians’ injury severity during pedestrian-vehicle collisions, specifically those involving distracted driving. It uses data from a police-reported collision database that contains pedestrian-vehicle collision information from 2007 to 2011 in Nova Scotia, Canada. A latent segmentation-based ordered logit (LSOL) model is developed in this paper that comprehensively examines the influence of built environment characteristics on pedestrian injury severity. It estimates a latent segment allocation model within LSOL modeling framework to capture unobserved heterogeneity across pedestrians. Two latent segments, high- and low-risk segments, are identified probabilistically based on pedestrian characteristics and action, driver action, and collision attributes. Results suggest that higher mixed land-use, transit stop density, length of sidewalk in the collision locations, and collisions occurring near schools yield lower pedestrian injury severity. In contrast, pedestrian-vehicle collisions in arterial roads, curved roads, sloped roads, and roundabouts tend to result in severe injuries. Interactions between distracted driving type and built environment characteristics are examined in this study. For example, using a communication device while driving on straight roads increases likelihood of higher pedestrian injury severity. This study also confirms the existence of heterogeneity across latent segments. For instance, higher percentage of people commuting by walking in the collision areas yield severe pedestrian injury in high-risk segments and lower injury severity in low-risk segments. The findings of this study will assist transportation planners and road safety stakeholders in developing effective and prioritized policies to reduce pedestrian injury severity involving distracted driving incidents.
Investigating Factors Affecting Pedestrian Crash Severity at High Speed Urban Arterial Roadways: A Case Study of Louisiana
Raju Thapa, Louisiana Department of Transportation and DevelopmentShow Abstract
Julius Codjoe (Julius.Codjoe@la.gov), Louisiana Department of Transportation and Development
Paul Kornyoh, Jacobs
Pedestrian fatalities on roadways is a key public safety concern, especially since overall trends point to higher fatalities for the period 2013 to 2017. In 2016, the state of Louisiana was ranked as the sixth worst state in pedestrian fatalities. Though previous studies have identified several factors contributing to severity of pedestrian crashes, none were found purposely modeled to address contributing factors on high-speed urban roadways. This study focused on identifying factors contributing to the severity of pedestrian crashes on high-speed urban arterials, using the state of Louisiana as a test case. A total of 1,337 crashes of different severity levels occurring on the defined roadways between 2013 to 2017 were extracted from the database. A generalized logistic regression model was developed to model a binary dependent variable, comprising combined severity levels A and B, against combined Levels C to E, as a factor of various independent variables. The result showed younger pedestrians less vulnerable to severe crashes compared to older pedestrians. The opposite was the case for younger drivers. Pedestrian actions, such as working on vehicle in road, playing, getting off from the vehicle, and standing alongside the roadway were found to be more dangerous. Distracted conditions of both pedestrians and drivers were found to result in mostly severe crashes. Presence of shoulder also affected the severity of crashes. Intersection crashes were found less fatal when compared to those at non-intersections. Crashes at dark were found to be more severe compared to day-time crashes.
Predicting Pedestrian Crash Occurrence And Injury Severity In Texas
Mashrur Rahman, University of Texas, AustinShow Abstract
Kara M. Kockelman (email@example.com), University of Texas, Austin
Kenneth Perrine, University of Texas, Austin
This study investigates pedestrian-involved crashes across Texas from 2010 through 2019. Crashes were mapped to over 708,738 road segments , along with road design, land use, transit, hospital, rainfall and other location features. Negative binomial model results show how total and fatal pedestrian-crash rates and counts rise with a segment’s number of lanes, transit stops, population and job densities, as well as proximity to schools and hospitals, while greater median and shoulder widths provide some protection. Higher speed limits are associated with lower crash frequencies but more fatalities. A heteroskedastic ordered probit (HOP) model for injury severity demonstrates how pedestrian crashes are more likely to be severe and fatal at night (8 PM – 5 AM), without overhead lighting, and when the pedestrians or drivers are intoxicated. Use of light-duty trucks (including SUVs, pickup trucks, CUVs, and vans) also significantly increases the risk of pedestrians being severely injured or killed. While newer vehicle safety features may be argued to lower crash severity, newer crash-involved vehicles in Texas are not found to deliver less pedestrian injury. However, being a younger or female pedestrian, on a straight segment, off the state (and interstate) highway system, in the presence of a traffic control device (stop sign or signal) lowers the likelihood of pedestrian injury, when one does become involved in such a crash.
Automated Video Processing for Pedestrian-Vehicle Conflict Analysis
Spencer Maddox, Kittelson & Associates, Inc. (KAI)Show Abstract
Colin Usher, Georgia Institute of Technology (Georgia Tech)
Kari Watkins, Georgia Institute of Technology (Georgia Tech)
Pedestrian fatalities have risen in the United States over the past decade. On an individual corridor, however, it is difficult to determine whether crashes and fatalities are statistically significant or random occurrences. When considering mitigation efforts, transportation planners and engineers therefore need to accurately categorize pedestrian exposure and risk. Traditionally, risk and exposure were calculated by performing manual counts. Advancements in automated video processing, where objects are tracked from a recorded video, can categorize conflicts automatically. Using outputs from a developed tracking system, this paper defines a successful methodology to identify conflicts and calculate the post-encroachment time. This methodology can be applied to both intersection and non-intersection locations. Results from four sample sites support previous research that mid-block crossings occur more often when crosswalks are not nearby and the relationship between pedestrians and conflicts are not necessarily linear. Using this conflict identification methodology and automated video processing provides transportation planners and engineers with a better understanding of pedestrian risk at key locations.
Factors related to road safety of couriers: A comparison of three regions in China
Changxin Sun (firstname.lastname@example.org), Wuhan UniversityShow Abstract
Yi HE, Wuhan University
Chaozhong Wu, Wuhan University
The couriers have become a popular mode for urban cargo transportation in China. Riders' risky behaviors of couriers lead to many accidents. We surveyed population attributes, working conditions, seven types of selected risky driving behaviors, and road accident involvement of 824 participants engaged in couriers from the BTUA, YRDUA ,and the PRDUA. Statistical analysis and regional comparison of data collected identified high-risk population groups and high-risk behaviors by regions, calculated the road accident risk degree (RARD) and the combined workload index (WLX) and discussed in traffic laws and management. A well-fitting path model is set up to explore the relationship between working conditions and road risk. The results show that couriers have a high workload and road accident risk, and BTUA is relatively better in three regions. The young people with intermediate working experience are the high-risk group. Among the seven risk behaviors, distracted driving and aggressive driving are of particular concern that have the high incidence and risk and national and local regulations also lack management of specific risky behaviors (especially in Guangzhou) and punitive measures. Finally, the path analysis model shows that workload promotes the generation of anxiety and anger, and promotes the occurrence of risk behaviors and road accidents to a certain extent. The road safety of couriers can be effectively improved by improving different working conditions according to the path differences in three regions.
Pedestrian Safety Hazard Due to Jaywalking and Cell Phone Induced Distractions: A Synopsis from Highway Intersections in Bangladesh
Moinul Hossain, Islamic University of Technology (IUT)Show Abstract
Afia Jahin Prema, Islamic University of Technology
Dewan Tanvir Ahammed, Islamic University of Technology
Niaz Mahmud, Islamic University of Technology
Md Asif Raihan, Bangladesh University of Engineering and Technology
Md. Abdullah Al Mamun, Universita degli Studi di Roma La Sapienza
Pedestrian fatalities account for 22% of all road traffic fatalities around the world. The statistics are even grimmer for the developing countries where jaywalking is predominant. There, along with jaywalking, the use of cell phones while crossing the road is acerbating pedestrian casualties. This delves into thought processing of jaywalkers and pedestrians using cell phones while crossing roads to devise countermeasures for improving pedestrian safety. The study observes pedestrian behavior at 32 intersections on national and regional highways of Bangladesh through video data and subsequently interviews 2,016 pedestrians found jaywalking and/or using cell phones while crossing the road. Data on their socio-economic and demographic characteristics, various risk perceptions, physical obstructions forcing jaywalking, distracting cell phone use, road crossing behavior and their knowledge about basic rules of the road were collected. Next, Bayesian Networks (BBN) were constructed to answer ‘who’, ‘why’ and ‘how’ related questions regarding jaywalkers and pedestrians who use a cell phone while road crossing. The findings suggest that jaywalking is more predominant among males, aged between 26-40 years who have received secondary education despite having decent knowledge regarding basic rules of the road. The most influential factors concerning risky jaywalking and using cell phone while road crossing are ‘Gender’, ‘Jaywalker Activities’, ‘Driving experiences’, ‘Purposes of Journey’, and ‘Frequency of visit that area’. The identified high impact variables associated with jaywalking, and also the triggering factors of cell phone-induced jaywalking are expected to assist decision-makers to develop pragmatic pedestrian safety policies in the context of developing countries.
Categorical Principal Component Analysis (CATPCA) of Pedestrian Crashes in Central Florida
Hatem Abou-Senna (email@example.com), University of Central FloridaShow Abstract
Essam Radwan, University of Central Florida
Hassan Tahsin, Cairo University
This research investigates the characteristics and contributing causes of pedestrian crashes that occurred in Central Florida over a 5 year-period at intersections and mid-blocks along roadway segments. The factors affecting pedestrian crashes were classified into five main categories; location characteristics, pedestrian factors, driver/vehicle characteristics, environmental-related factors and crash characteristics. Categorical Principal Components Analysis (CATPCA) was applied to understand the structure of a set of variables and to reduce the dimensionality of the dataset to a predefined number of dimensions and components. The analysis showed that majority of the intersection crashes were during night time with pedestrians under influence and failing to yield to the right of way (ROW). They included mainly left-turn and right-turn crashes. In addition, drivers were also found at fault due to vision issues resulting from absence of lighting at intersections and categorized as failure to yield to the ROW. At midblock locations, major crash types were through moving vehicles hitting pedestrians crossing and walking along the roadway especially during night time conditions. However, majority of the crashes were at locations away from the designated crossings likely due to the long distances between legal locations and pedestrian’s failure to utilize them. The findings of this research and examining the factors affecting pedestrians’ crash likelihood and injury severity can lead to better crash mitigation strategies, countermeasures and policies that would alleviate this growing problem in Central Florida.
Effects of Various Speed Management Countermeasures on Bicycle Crashes for Urban Roads in Central Florida
Jorge Ugan, University of Central FloridaShow Abstract
Mohamed Abdel-Aty, University of Central Florida
Qing Cai (firstname.lastname@example.org), University of Central Florida
Nada Mahmoud, University of Central Florida
Ma'en Al-Omari, University of Central Florida
Brenda Young, Florida Department of Transportation
In recent years, cycling has become an increasingly popular mode of transportation around the world. In contrast to other popular modes of transportation, cycling is more economic and energy efficient. Many studies that have focused on bicycle safety, were limited in terms of bicycle exposure data This study tries to improve the current safety performance functions for bicycle crashes at urban corridors by utilizing crowdsource data from STRAVA and on-street speed management countermeasures data. Since there is a disproportion in the representation of cyclists from the STRAVA data, adjustments, using a Tobit model, were done to more accurately represent the cyclists based on video detection data. This study aims to (1) identify a method to get more accurate bicycle exposure data, (2) analyze the effect that speed management countermeasures have on bicycle safety, and (3) incorporate other contributing factors to bicycle safety. To achieve these objectives, a Bayesian hierarchical model was used to predict the frequencies of bicycle crashes and adjust STRAVA data at the same time. Traffic, roadway attributes, on-street speed management countermeasures data, and land use data were considered in the model. The results revealed several key components for bicycle safety at urban intersections. The study concluded that crowdsourced data adjusted based on video detection and on-street speed management countermeasures data are significant when analyzing bicycle safety.
DISCLAIMER: All information shared in the TRB Annual Meeting Online Program is subject to change without notice. Changes, if necessary, will be updated in the Online Program and this page is the final authority on schedule information.