Rapid advancements in automated vehicles (AVs) technology have transformed the measurability, controllability and unpredictability of transportation systems. Evaluating driving risk and effectively managing driving behaviours is critical. AVs should be enabled to identify, analyse, evaluate and devise effective countermeasures for driving risk by autonomously learning from human driving experiences. This learning will enhance the interactive decision-making capabilities and achieve driving behaviours that reflect human-like logic. The primary challenge lies in integrating human-like logic and driving risk constraints based on behavioural decision-making outcomes. This integration is crucial to align the AVs' cognitive levels—movement, comprehension, memory and inference—more closely with human driving necessities, habits and styles. Such alignment holds the potential to improve the prediction and planning of future actions and facilitate the development of motion planning schemes geared towards minimizing driving risk. We performed a comprehensive review of AVs' behavioural decision-making and intelligent motion-planning research from 2000 to 2023 from four key perspectives—driving risk, human-like logic, behavioural decision making and motion planning. Based on the Web of Science and China National Knowledge Internet database, the results of our review indicate that significant progress has been made in AV behavioural decision making and intelligent motion planning over time. When AVs and human-driven vehicles coexist, greater incorporation of human-like logic is required. Guided by these findings, we delineate future development directions and propose a research paradigm for human-like logic and a research framework for human-like logic-driven behavioural decision making and intelligent motion planning of AVs.

Crash risk identification and prediction are expected to play an important role in traffic accident prevention. However, most of the existing studies focus only on highways, not on multi-lane weaving areas. In this paper, a potential collision risk identification and conflict prediction model based on extending Time-to-Collision-Machine Learning (TTC-ML) for multi-lane weaving zone was proposed. The model can accurately learn various features, such as vehicle operation characteristics, risk and conflict distributions, and physical zoning characteristics in the weaving area. Specifically, TTC was used to capture the collision risk severity, and ML extracted vehicle trajectory features. After normalizing and dimensionality reduction of the vehicle trajectory dataset, Naive Bayes, Logistic Regression, and Gradient Boosting Decision Tree (GBDT) models were selected for traffic conflict prediction, and the experiments showed that the GBDT model outperforms two remaining models in terms of prediction accuracy, precision, false-positive rate (FPR) and Area Under Curve (AUC). The research findings of this paper help traffic management departments develop and optimize traffic control schemes, which can be applied to Intelligent Vehicle Infrastructure Cooperative Systems (IVICS) dynamic warning.

1. Driving intention, characterized by ambiguity, dynamism, and unpredictability, necessitates implementing human-like thinking and effective driving intention prediction as pivotal steps toward fu...

Safety study of driving behaviors to reduce traffic accidents is of great significance. The primary objectives of the study were to investigate the relation between driver overtaking behaviors and traffic safety. Twelve drivers of different proficiency were selected for an experiment on a driving simulation system platform. A simulation scene of overtaking on a two-lane highway was set up. The subjects overtook in the virtual scenarios at different speeds. A total of twelve parameters - including speed, acceleration, time of overtaking process, the changes of distance and other parameters - were extracted during the experiments. These data together with the results of a DBQ (Driving Behavior Questionnaire) were analyzed and evaluated by the multiple linear regression method. The results showed that three motion parameters had a strong correlation with drivers' safety. Finally, the study presents a linear model of safety assessment of driver overtaking behavior on two-lane highways. The model may help to identify safe and unsafe drivers and reduce the number of traffic accidents.

Traditional road monitoring methods, including radar, light detection and ranging (LiDAR), and video surveillance, face challenges in accurately and comprehensively identifying unexpected events and potential risks owing to high costs, low accuracy, and environmental complexities. This study proposes an audio-based monitoring method for detecting abnormal road events aimed at uncovering hidden traffic risks. Audio samples collected from various sources, such as FreeSound and Zapslat, are preprocessed using noise reduction, detection, integration, and digital signal conversion via sample quantization. An audio recognition classification model based on the RUSBoost algorithm is constructed and optimized using sequential model-based optimization (SMBO). This optimization addresses issues related to imbalanced samples and features. Optimization increased accuracy from 68.9% to 96.5% and reduced the error rate to 3.5%. The true positive rate (TPR) and true negative rate (TNR) also showed significant improvements, along with notable enhancements in the positive predictive value (PPV), F1 score, Matthews correlation coefficient, and Kappa. For key audio events such as skidding and collisions, TPR and PPV exceeded 93.2%. The SMBO-RUSBoost model distinguished traffic events from noise by extracting multidimensional features from audio data corresponding to various traffic events. Its classification performance provide a solid foundation for traffic risk assessment ad decision-making. The model offers a novel approach for improving road safety.