Exploring driving behavioral characteristics in pre-, in-, and post-conflict stages based on car-following trajectory data.

This study investigates driving behaviour in different stages of rear-end conflicts using vehicle trajectory data. Three conflict stages (pre-, in-, and post-conflict) are defined based on time-to-collision (TTC) indicator. Four indexes are selected to capture within-group and between-group characteristics of the stages. Besides, this study also examines the prediction performance of conflict stage identification using specific driving behaviour characteristics associated with each stage. Results reveal variations in dominant driving characteristics and predictive importance across stages. Heterogeneity exists within stages, with differences among clusters. Drivers slow down during in-conflict, with decreasing speed reduction as stages progress. Reaction time increases in post-conflict. Insufficient space gaps contribute to rear-end conflicts in the in-conflict stage. Furthermore, the prediction performance of conflict stage identification, based on the specific driving behaviour characteristics associated with each stage, is commendable. This study enhances understanding and prediction of conflict stage identification in rear-end conflicts.Practitioner summary: This study explores driving behaviour in rear-end conflict stages using trajectory data. It identifies pre-, in-, and post-conflict stages via time-to-collision indicator and assesses within-group and between-group characteristics. Besides, prediction performance for conflict stage identification based on these characteristics is commendable. This research enhances understanding and prediction of rear-end conflicts.

Exploring master scenarios for autonomous driving tests from police-reported historical crash data using an adaptive search sampling framework.

Crash scenario-based testing is crucial for assessing autonomous driving safety. However, existing studies on scenario generation tend to prioritize concrete scenarios for direct testing, neglecting the construction of fundamentally functional scenarios with a broader range. Police-reported historical crash data is a valuable supplement, yet detecting all potential crash scenarios is laborious. In order to address this issue, this study proposes an adaptive search sampling framework based on deep generative model and surrogate model (SM) to extract master scenario samples from police-reported historical crash data. The framework starts with selecting representative samples from the full crash dataset as initial master scenario samples using various sampling techniques. Evaluation indexes are then constructed, and derived scenario samples are synthesized using the deep generative model. To enhance efficiency, an SM is established to replace the generative model's training and data generation process. Based on the SM, an adaptive search sampling method is developed, which iteratively adjusts the sampling strategy using the Similarity Score to achieve comprehensive sampling. Experimental results demonstrate the notable advantage of the adaptive search sampling method over other sampling methods. Furthermore, statistical analysis and visualization assessments confirm the effectiveness and accuracy of the proposed method.

How would autonomous vehicles behave in real-world crash scenarios?

Autonomous Vehicles (AVs) have the potential to revolutionize transportation systems by enhancing traffic safety. Safety testing is undoubtedly a critical step for enabling large-scale deployment of AVs. High-risk scenarios are particularly important as they pose significant challenges and provide valuable insights into the driving capabilities of AVs. This study presents a novel approach to assess the safety of AVs using in-depth crash data, with a particular focus on real-world crash scenarios. First, based on the high-definition video recording of the whole process prior to the crash occurrences, 453 real-world crashes involving 596 passenger cars from China In-depth Mobility Safety Study-Traffic Accident (CIMSS-TA) database were reconstructed. Pertinent static and dynamic elements needed for the construction of the testing scenarios were extracted. Subsequently, 596 testing scenarios were created via each passenger car's perspective within the simulation platform. Following this, each of the crash-involved passenger cars was replaced with Baidu Apollo, a famous automated driving system (ADS), for counterfactual simulation. Lastly, the safety performance of the AV was assessed using the simulation results. A logit model was utilized to identify the fifteen crucial scenario elements that have significant impacts on the test results. The findings demonstrated that the AV could avoid 363 real-world crashes, accounting for approximately 60.91% of the total, and effectively mitigated injuries in the remaining 233 unavoidable scenarios compared to a human driver. Moreover, the AV maintain a smoother speed in most of the scenarios. The common feature of these unavoidable scenarios is that the AV is in a passive state, and the crashes are not caused by the AV violating traffic rules, but rather caused by abnormal behavior exhibited by the human drivers. Additionally, seven specific scenarios have been identified wherein AVs are unable to avoid a crash. These findings demonstrate that, compared to human drivers, AVs can avoid crashes that are difficult for humans to avoid, thereby enhancing traffic safety.

A variable speed limit control approach for freeway tunnels based on the model-based reinforcement learning framework with safety perception.

To improve the traffic safety and efficiency of freeway tunnels, this study proposes a novel variable speed limit (VSL) control strategy based on the model-based reinforcement learning framework (MBRL) with safety perception. The MBRL framework is designed by developing a multi-lane cell transmission model for freeway tunnels as an environment model, which is built so that agents can interact with the environment model while interacting with the real environment to improve the sampling efficiency of reinforcement learning. Based on a real-time crash risk prediction model for freeway tunnels that uses random deep and cross networks, the safety perception function inside the MBRL framework is developed. The reinforcement learning components fully account for most current tunnels' application conditions, and the VSL control agent is trained using a deep dyna-Q method. The control process uses a safety trigger mechanism to reduce the likelihood of crashes caused by frequent changes in speed. The efficacy of the proposed VSL strategies is validated through simulation experiments. The results show that the proposed VSL strategies significantly increase traffic safety performance by between 16.00% and 20.00% and traffic efficiency by between 3.00% and 6.50% compared to a fixed speed limit approach. Notably, the proposed strategies outperform traditional VSL strategy based on the traffic flow prediction model in terms of traffic safety and efficiency improvement, and they also outperform the VSL strategy based on model-free reinforcement learning framework when sampling efficiency is considered together. In addition, the proposed strategies with safety triggers are safer than those without safety triggers. These findings demonstrate the potential for MBRL-based VSL strategies to improve traffic safety and efficiency within freeway tunnels.

Multi-objective deep reinforcement learning approach for adaptive traffic signal control system with concurrent optimization of safety, efficiency, and decarbonization at intersections.

This study introduces a novel approach to adaptive traffic signal control (ATSC) by leveraging multi-objective deep reinforcement learning (DRL) techniques. The proposed scheme aims to optimize control strategies at intersections while concurrently addressing the objectives of safety, efficiency, and decarbonization. Traditional ATSC schemes primarily emphasize traffic efficiency and often lack the ability to adapt to real-time dynamic traffic conditions. To overcome these limitations, the study proposes a DRL-based ATSC algorithm that integrates the Dueling Double Deep Q Network (D3QN) framework. The performance of the proposed algorithm is evaluated through a simulated intersection in Changsha, China. Specifically, the proposed ATSC algorithm outperforms both traditional ATSC and ATSC with efficiency optimization only algorithms by achieving more than a 16% reduction in traffic conflicts and a 4% reduction in carbon emissions. In terms of traffic efficiency, waiting time reduces by 18% compared to traditional ATSC, but slightly increases (0.64%) compared to DRL-based ATSC algorithm that integrates D3QN framework. This small increase indicates a trade-off between efficiency and other objectives such as safety and decarbonization. Moreover, the proposed scheme demonstrates superior performance specifically in highly traffic-demand scenarios in terms of all three objectives. The findings of this study contribute to the advancement of traffic control systems by providing a practical and effective solution for optimizing signal control strategies in real-world traffic scenarios.

Analysis of dynamic determinants of vehicles involved in crash affecting severity based on in-depth crash data.

OBJECTIVE: The dynamic characteristics of vehicles involved in crashes may be an important factor affecting the crash severity. This study investigates the relationship between the dynamic characteristics of vehicles involved in crashes in the five seconds before the occurrence and the crash severity. The findings aim to offer insights for preventing severe crashes and advancing autonomous vehicle technology. METHODS: This study aims to investigate the impact of dynamic features, such as speed, acceleration, and relative distance of vehicles involved in the crash in the five seconds before the crash, on the crash severity. Five hundred ninety-six crash samples from the China In-depth Mobility Safety Study-Traffic Accident database were selected for crash reconstruction. A random parameters logit model was used to extract and analyze the effect of dynamic features of the vehicles involved in the crash on the crash severity. RESULTS: The random parameters logit model demonstrated a satisfactory fit. Analysis of the parameter estimation results of the model showed that the variables of speed, acceleration, and relative distance between vehicles involved in the crash at some time points during the five seconds before the crash significantly affected the crash severity. Notably, the coefficient of variation of relative distance over 5 s emerged as the most influential positive determinant of the crash severity. CONCLUSIONS: Certain dynamic characteristics of vehicles involved in a crash in the five seconds before a crash significantly impact the crash severity. The study's findings can serve as a reference for preventing severe crashes and advancing the development of autonomous vehicles.

Exploring influences on the intention of motorcyclists to drink and ride: An investigation in a fourth-tier city of China.

OBJECTIVE: The analysis of motorcyclists' intention to drink and ride can help reduce the possibility of accidents caused by the relevant behavior of motorcyclists. The main objectives of this study are to identify important factors in motorcyclists' intention to drink and ride and to make some recommendations that could effectively reduce their riding intention after drinking. METHODS: To explore the effects of demographic and psychological variables on motorcyclists' behavioral intention to drink and ride, a questionnaire based on the extended theory of planned behavior was designed. Two hundred and five fully completed questionnaires were collected through a survey in Shaoguan, China. A hierarchical regression model was used to analyze observed data. RESULTS: The final hierarchical regression model explained 37.5% of the variance in intention to drink and ride. While initial tests showed that gender, marital status and age influenced some TPB variables, gender was the only demographic variable found to be significant on the riding intention after drinking alcohol in the hierarchical regression analysis. Among the psychological factors quantified by the extended theory of planned behavior, all factors had significant effects on intention except for risk perception, and subjective norms were the most influential factor. CONCLUSIONS: The extended theory of planned behavior can be well used to examine the factors that influence motorcyclists' intention to drink and ride. A more nuanced understanding of these factors can be found in the results.