Download or read book Identification of Surrogate Measures of Safety and Means to Assess the Performance of Connected Vehicles written by Elhashemi Mohammed Ali and published by . This book was released on 2018 with total page 167 pages. Available in PDF, EPUB and Kindle. Book excerpt: Sudden changes in weather conditions might have a tremendous impact on traffic operation and safety. Previous studies investigated the impact of adverse weather conditions on traffic safety and to what extent these conditions may increase crash risks on roadways. The increase in weather-related crashes has motivated researchers to study driver behavior and performance under different weather conditions. Adverse weather affects driver decisions and may result in taking improper actions while facing a crash/near-crash event in comparison with clear weather conditions. While driver behavior and performance are considered among the key contributing factors to crashes, little research have been conducted to fully understand the difference between normal driving and safety critical scenarios for developing crash prevention means. Monitoring driver behavior and performance during a safety critical event has been a challenging task for researchers due to the lack of detailed event records. Moreover, the issues associated with traditional police records of crashes have limited a comprehensive analysis of how the deviation from normal driving may lead to a culmination of crashes. In addition, one of the main reasons for the increase number of crashes on roadways is that drivers may not appropriately adapt their behaviors to compensate for adverse weather conditions. The lack of real-time trajectory-level weather information and the sporadic data collected from weather stations have limited researchers from conducting sound safety studies. This study attempts to fulfill some of the research gaps to assist transportation agencies and traffic safety researchers to improve safety and mobility. In general, the research efforts conducted in this dissertation aims to improve traffic safety in adverse weather conditions on freeways. In addition, this dissertation aims to provide practical recommendations to transportation agencies that can efficiently enhance traffic safety in Connected and Automated Vehicle (CAV) environments. The dissertation goal was achieved through utilizing different subsets of the Second Strategic Highway Research Program (SHRP2) – Naturalistic Driving Study (NDS) data. The utilization of the NDS real-time trajectory dataset would open a new horizon in traffic safety research related to connected and automated vehicles. In this study, five main research objectives, each with multiple tasks, were set to enhance traffic safety in adverse weather conditions. The first objective was to provide a better understanding of what happened before and during a near-crash event and comparing it with normal matched trips. This objective would help to develop effective countermeasures that reduce crash risks on freeways. The second objective was to detect Surrogate Measures of Safety (SMoS) on freeways by comparing environmental conditions and vehicle kinematics signatures of near-crash events to their matched normal driving trips. A time-chunking technique was used with different aggregation levels to monitor changes in vehicle kinematics on a timescale. This approach established a comparative study of parametric and non-parametric techniques to estimate near-crashes on freeways. A Binary Logistic Regression model was used as a parametric prediction model, while the Decision Tree (DT), k-Nearest Neighbors (k-NN), and Deep Learning Artificial Neural Network (ANN) were used as non-parametric prediction models. The results showed that the logistic regression model has provided an excellent fit to the input data and can predict near-crashes with an outstanding accuracy. In addition, DT and Deep Learning ANN machine learning algorithms showed higher prediction accuracy of near-crashes compared to the k-NN algorithm. The third objective was to investigate normal and risky driving condition patterns under both rainy and clear weather conditions. The fourth objective was to distinguish between normal driving and risky driving condition patterns in rainy and clear weather conditions using real-time trajectory-level datasets. To achieve the third and fourth objectives, the SHRP2 - NDS data were employed to investigate the behavior of normal and risky driving under both rainy and clear weather conditions. Near-crash events on freeways, which were used as Surrogate Measure of Safety (SMoS) for crash risk, were identified based on the changes in vehicle kinematics, including speed, longitudinal and lateral acceleration and deceleration rates, and yaw rates. Through a trajectory-level data analysis, there were significant differences in driving patterns between rainy and clear weather conditions; factors that affected crash risk mainly included driver reaction and response time, their evasive maneuvers such as changes in acceleration rates and yaw rates, and lane-changing maneuvers. A cluster analysis method was employed to classify driving patterns into two clusters: normal and risky driving condition patterns, respectively. Statistical results showed that risky driving patterns started one second earlier in rainy weather condition than in clear weather condition. Furthermore, risky driving patterns extended three seconds in rainy weather condition, while it was two seconds in clear weather condition.