Florida experienced serious pedestrian safety problems and had the highest pedestrian fatality rate in the U.S. from 2008—2011 based on National Highway Traffic Safety Administration (NHTSA) Traffic Safety Facts annual reports. The 2014 edition of Dangerous by Design ranked four metropolitan areas in Florida as the top four most dangerous areas to walk in the U.S.–Metro Orlando, Tampa—St. Petersburg, Jacksonville, and Miami. The potential reason was primarily the rapid spread of low-density neighbourhoods that rely on wider streets with higher speeds to connect homes, shops, and schools–roads that tend to be more dangerous for people walking. One of Florida’s highest priorities is investigating major contributing causes for pedestrian fatalities and developing effective countermeasures to significantly improve pedestrian safety in the state. This project intended to research and understand the interactions between drivers and pedestrian features (e.g., pedestrian signs, pedestrian signals, traffic signals, crosswalks, and pavement markings) at signalized intersections. Understanding these interactions can help the Florida Department of Transportation (FDOT) develop effective countermeasures to improve pedestrian safety. The Strategic Highway Research Program 2 (SHRP2) Naturalistic Driving Study (NDS) recorded the driving behaviour of a large sample of drivers in their personal vehicles, offering project researchers comprehensive naturalistic driving behavioural data for researching the interactions between drivers and various pedestrian features at selected signalized intersections through which they drove. This database provides great and unique opportunities to research the proposed research question in this project. The major proposed research question for this study was, “Based on information from the SHRP2 NDS and Roadway Information Database (RID) datasets, how do drivers interact with pedestrian features at signalized intersections?” These pedestrian features include pedestrian signs, pedestrian signals, traffic signals, crosswalks, and pavement markings. Driver interactions with pedestrian features may be observed from driver responses to different pedestrian features, yielding behaviours to pedestrians, driver speeds, braking patterns, and attention and/or distraction. The proposed research question covered four broad areas: Vulnerable Road Users, Roadway Features and Driver Performance, Intersections, and Driver Speed. Gender and age group were included in the research. The findings of the research can help FDOT to better understand (1) driver interactions with various pedestrian features at signalized intersections, (2) the effectiveness of pedestrian features, (3) the impact of gender and age group on driver interactions, (4) specific interactions between drivers and pedestrians, and (5) the impact of driver attention and/or distraction on driver interactions. These understandings can assist FDOT in developing effective engineering, education, and enforcement strategies and countermeasures to reduce pedestrian fatalities and enhance pedestrian safety in Florida. SHRP2 NDS data consist of two large datasets. The main dataset includes naturalistic driving data from instrumented vehicles and supplemental driver information managed by the Virginia Tech Transportation Institute (VTTI); the second dataset includes the RID managed by the Center for Transportation Research and Education (CTRE) at the University of Iowa. Since the major research question targets pedestrian safety, the research team acquired the NDS data stream of vehicles passing selected signalized intersections in Tampa Bay with specific pedestrian features, high pedestrian activity, and/or crash frequency. This provided an excellent opportunity for researching the impact of pedestrian features and having a higher possibility for observing interactions between drivers and pedestrians. The intersection locations were identified in coordination with FDOT. To minimize the data required, a driver sample size that is sufficient for performing the analysis was used (This can be expanded to the whole database in the future research). Coordination with VTTI and CTRE was required for proper data acquisition. For the NDS data, specific data streams were requested and filtered to include only a few seconds of video before and after drivers negotiate through the specified signalized intersections. The data requested included but were not limited to forward movement video; vehicle forward and lateral acceleration; braking; turn indications; headlights; GPS path for the sections; and driver characteristics. Understanding the interactions between drivers and pedestrian features is important for developing effective countermeasures to improve pedestrian safety at signalized intersections. Driver interactions with pedestrian features include driver speeds, braking patterns, and yielding and stopping behaviours. Most commonly, drivers who fail to comply with pedestrian-related features near the crosswalks of signalized intersections can increase accident risk for pedestrians. By using traditional data collection methods, several studies were conducted to evaluate the effectiveness of pedestrian signs. Herman (2000) designed a treatment-and-control study to evaluate the effects of “NO TURN ON RED/YIELD TO PEDESTRIANS” variable message signs. A video camera was set up on a sidewalk along a main road to record pedestrian and motorist behaviour at selected intersections. The sites with the signs were found to have a lower ratio of motorists who illegally turned right on right, but there were no significant differences in terms of the number of right-turn-on-green motorists who yielded to pedestrians. Karkee et al. (2006) conducted a “before-and-after” study to test differences in several measures for “Turning Traffic Must Yield to Pedestrians” signs. Data were collected during AM and PM peak hours during each study period, and results showed increases in yielding behaviour and both pedestrian and vehicle delay after the signs were installed. A study performed by Pulugurtha et al. (2010) evaluated the effects of traffic signs based on field observations of pedestrian and driver behaviours. Results showed a general improvement in driver yielding behaviour due to installation of “YIELD TO PEDESTRIANS” signs. Similarly, using field observation or fixed video recording methods, Pecheux et al. (2009) and Fitzpatrick et al. (2014) included some general findings for pedestrian features such as “YIELD TO PEDESTRIANS” signs in their pedestrian safety studies. Although these previous studies attempted to evaluate the effectiveness of traffic signs, the traditional methods failed to collect important safety-related factors such as driver speed profiles, braking patterns, driver characteristics (e.g., gender, age, frequency of risk-taking), vehicle factors, and various roadway/environmental conditions. The SHRP2 NDS datasets and RID databases provide rich and unique information for understanding the interactions between drivers and pedestrian features at signalized intersections to increase pedestrian safety. The first major goal of this project was to understand the interactions between drivers and pedestrian features at signalized intersections using the SHRP2 NDS and RID datasets and to obtain initial results and findings. The second major goal was to demonstrate that the research team effectively used the NDS and RID databases to conduct research and analysis, leading to development of effective countermeasures in future studies to improve pedestrian safety at signalized intersections. The specific objectives of this project were to: 1. Acquire knowledge and request data from the SHRP2 NDS and RID datasets to conduct initial analysis of driver interactions with pedestrian features at signalized intersections. 2. Develop effective data extraction and analysis tools and identify specific parameters and factors that will aid in initial analysis in this project and full analysis in future studies pertaining to the research question: “Based on information from the SHRP2 NDS and RID datasets, how do drivers interact with pedestrian features at signalized intersections?” 3. Conduct initial analysis using a more manageable dataset of acquired NDS data, provide initial findings on the research question, and offer recommendations that can be implemented by FDOT in future studies. 4. Demonstrate the effective and successful use of the SHRP2 NDS and RID datasets via this project to provide recommendations and guidance for future studies. (Author/publisher)
Abstract