Modeling distracted driving behavior considering cognitive processes.

The modeling of distracted driving behavior has been studied for many years, however, there remain many distraction phenomena that can not be fully modeled. This study proposes a new method that establishes the model using the queuing network model human processor (QN-MHP) framework. Unlike previous models that only consider distracted-driving-related human factors from a mathematical perspective, the proposed method reflects the information processing in the human brain, and simulates the distracted driver's cognitive processes based on a model structure supported by physiological and cognitive research evidence. Firstly, a cumulative activation effect model for external stimuli is adopted to mimic the phenomenon that a driver responds only to stimuli above a certain threshold. Then, dual-task queuing and switching mechanisms are modeled to reflect the cognitive resource allocation under distraction. Finally, the driver's action is modeled by the Intelligent Driver Model (IDM). The model is developed for visual distraction auditory distraction separately. 773 distracted car-following events from the Shanghai Naturalistic Driving Study data were used to calibrate and verify the model. Results show that the model parameters are more uniform and reasonable. Meanwhile, the model accuracy has improved by 57% and 66% compared to the two baseline models respectively. Moreover, the model demonstrates its ability to generate critical pre-crash scenarios and estimate the crash rate of distracted driving. The proposed model is expected to contribute to safety research regarding new vehicle technologies and traffic safety analysis.

Analyzing relationships between latent topics in autonomous vehicle crash narratives and crash severity using natural language processing techniques and explainable XGBoost.

Safety is one of the most essential considerations when evaluating the performance of autonomous vehicles (AVs). Real-world AV data, including trajectory, detection, and crash data, are becoming increasingly popular as they provide possibilities for a realistic evaluation of AVs' performance. While substantial research was conducted to estimate general crash patterns utilizing structured AV crash data, a comprehensive exploration of AV crash narratives remains limited. These narratives contain latent information about AV crashes that can further the understanding of AV safety. Therefore, this study utilizes the Structural Topic Model (STM), a natural language processing technique, to extract latent topics from unstructured AV crash narratives while incorporating crash metadata (i.e., the severity and year of crashes). In total, 15 topics are identified and are further divided into behavior-related, party-related, location-related, and general topics. Using these topics, AV crashes can be systematically described and clustered. Results from the STM suggest that AVs' abilities to interact with vulnerable road users (VRUs) and react to lane-change behavior need to be further improved. Moreover, an XGBoost model is developed to investigate the relationships between the topics and crash severity. The model significantly outperforms existing studies in terms of accuracy, suggesting that the extracted topics are closely related to crash severity. Results from interpreting the model indicate that topics containing information about crash severity and VRUs have significant impacts on the model's output, which are suggested to be included in future AV crash reporting.

Developing safety performance functions for freeways at different aggregation levels using multi-state microscopic traffic detector data.

Safety Performance Functions (SPFs) have been widely used by researchers and practitioners to conduct roadway safety evaluation. Traditional SPFs are usually developed by using annual average daily traffic (AADT) along with geometric characteristics. However, the high level of aggregation may lead to a failure to capture the temporal variation in traffic characteristics (e.g., traffic volume and speed) and crash frequencies. In this study, SPFs at different aggregation levels were developed based on microscopic traffic detector data from California, Florida, and Virginia. More specifically, five aggregation levels were considered: (1) annual average weekday hourly traffic (AAWDHT), (2) annual average weekend hourly traffic (AAWEHT), (3) annual average weekday peak/off-peak traffic (AAWDPT), (4) annual average day of the week traffic (AADOWT), and (5) annual average daily traffic (AADT). Model estimation results showed that the segment length and volume, as exposure variables, are significant across all the aggregation levels. Average speed is significant with a negative coefficient, and the standard deviation of speed was found to be positively associated with the crash frequency. It is noteworthy that the operation of the high occupancy vehicle (HOV) lanes was found to have a positive effect on crash frequency across all the aggregation levels. The model results also showed that the AAWDPT and AADOWT models consistently performed better (the improvements range from 3.14%-16.20%) than the AADT-based SPF, which implies that the differences between the day of the week and peak/off-peak periods should be considered in the development of crash prediction models. The model transferability results indicated that the SPFs between Florida and Virginia are transferrable, while the models between California and the other two states are not transferrable.

In-depth approach for identifying crash causation patterns and its implications for pedestrian crash prevention.

INTRODUCTION: A pedestrian crash occurs due to a series of contributing factors taking effect in an antecedent-consequent order. One specific type of antecedent-consequent order is called a crash causation pattern. Understanding crash causation patterns is important for clarifying the complicated growth of a pedestrian crash, which ultimately helps recommend corresponding countermeasures. However, previous studies lack an in-depth investigation of pedestrian crash cases, and are insufficient to propose a representative picture of causation patterns. METHOD: In this study, pedestrian crash causation patterns were discerned by using the Driving Reliability and Error Analysis Method (DREAM). One hundred and forty-two pedestrian crashes were investigated, and five pedestrian pre-crash scenarios were extracted. Then, the crash causation patterns in each pre-crash scenario were analyzed; and finally, six distinct patterns were identified. Accordingly, 17 typical situations corresponding to these causation patterns were specified as well. RESULTS: Among these patterns, the pattern related to distracted driving and the pattern related to an unexpected change of pedestrian trajectory contributed to a large portion of the total crashes (i.e., 27% and 24%, respectively). Other patterns also played an important role in inducing a pedestrian crash; these patterns include the pattern related to an obstructed line of sight caused by outside objects (9%), the pattern that involves reduced visibility (13%), and the pattern related to an improper estimation of the gap distance between the vehicle and the pedestrian (10%). The results further demonstrated the inter-heterogeneity of a crash causation pattern, as well as the intra-heterogeneity of pattern features between different pedestrian pre-crash scenarios. Conclusions and practical applications: Essentially, a crash causation pattern might involve different contributing factors by nature or dependent on specific scenarios. Finally, this study proposed suggestions for roadway facility design, roadway safety education and pedestrian crash prevention system development.

Assessment of the safety benefits of vehicles' advanced driver assistance, connectivity and low level automation systems.

The Connected Vehicle (CV) technologies together with other Driving Assistance (DA) technologies are believed to have great effects on traffic operation and safety, and they are expected to impact the future of our cities. However, few research has estimated the exact safety benefits when all vehicles are equipped with these technologies. This paper seeks to fill the gap by using a general crash avoidance effectiveness framework for major CV&DA technologies to make a comprehensive crash reduction estimation. Twenty technologies that were tested in recent studies are summarized and sensitivity analysis is used for estimating their total crash avoidance effectiveness. The results show that crash avoidance effectiveness of CV&DA technology is significantly affected by the vehicle type and the safety estimation methodology. A 70% crash avoidance rate seems to be the highest effectiveness for the CV&DA technologies operating in the real-world environment. Based on the 2005-2008 U.S. GES Crash Records, this research found that the CV&DA technologies could lead to the reduction of light vehicles' crashes and heavy trucks' crashes by at least 32.99% and 40.88%, respectively. The rear-end crashes for both light vehicles and heavy trucks have the most expected crash benefits from the technologies. The paper also studies the effectiveness of Forward Collision Warning technology (FCW) under fog conditions, and the results show that FCW could reduce 35% of the near-crash events under fog conditions.