Abstract:Real-time assessment and short-term warning of driving risks are critical for AI-assisted vehicles to significantly improve the safety and reliability of mobility. However, existing methods do not comprehensively consider these factors, making it difficult to achieve more accurate risk assessments. Aiming at this problem, this paper proposes a new driving risk assessment framework by integrating multimodal data. First, based on naturalistic driving experiments, we collected multimodal data encompassing human-vehicle-road factors. Then, using the Latent Dirichlet Allocation (LDA) model, we identified three risk levels based on driving behavior features: normal driving, longitudinal risky driving, and lateral risky driving. To better understand the spatiotemporal importance of multiple factors, a spatiotemporal dual-channel neural network based on a multi-layer attention mechanism (MLA-DCNN) is developed. This model has a spatiotemporal dual-channel structure, which can integrate "low-level" historical sequences and "high-level" extract statistical features of multiple features. In addition, it adopts three layers of attention mechanism, respectively used to capture the differences of features in temporal, spatial, and extracted-level dimensions. Results reveal that the LDA model is more effective than traditional clustering methods in uncovering latent patterns of driving risk. The proposed model achieved an impressive accuracy of 91.04%, demonstrating higher risk assessment capabilities than the other alternative models. In addition, the multilayer attention enhances the interpretability of the model and is able to capture the spatiotemporal importance of different factors across various road environments. This method can be applied to connected and automated vehicles (CAVs) using multimodal natural driving data collected by in-vehicle sensors. It enhances the risk warning capabilities of driving assistance systems, and the multidimensional importance analysis also supports decision-making for traffic management authorities.

A quantitative blind area risks assessment method for safe driving assistance

Drive with Quantified Risk: A Probabilistic Approach for Occlusion-Aware Trajectory Planning

Occlusion-Aware Risk Assessment for Autonomous Driving in Urban Environments

A Multimodal Data-Driven Approach for Driving Risk Assessment

Unified Quantitative Method of Driving Risk by Comprehensive Considering Driver-Vehicle-Road Factors

Occlusion-aware Risk Assessment and Driving Strategy for Autonomous Vehicles Using Simplified Reachability Quantification

An Emergency Driving Intervention System Designed for Driver Disability Scenarios Based on Emergency Risk Field

Shadow Based Non-Line-of-Sight Pedestrian Rushing Detection for Automated Driving

Risk Quantification and Prediction of Non-Driving-related Tasks on Driver's Critical Intervention Behavior in Autonomous Driving Scenarios

Spatial Risk Distribution: A Quantitative Approach for Assessing Hazards from Space

A Method for Vehicle Collision Risk Assessment through Inferring Driver's Braking Actions in Near-Crash Situations

Overcoming Blind Spots: Occlusion Considerations for Improved Autonomous Driving Safety

Risk Assessment Method for Driving Scenarios of Autonomous Vehicles Based on Drivable Area.

Identification of dangerous area within vehicles operation for driver assistance

Risk Occupancy: A New and Efficient Paradigm through Vehicle-Road-Cloud Collaboration

A probabilistic risk assessment framework considering lane-changing behavior interaction

Risk perception and the warning strategy based on safety potential field theory

To Turn or Not to Turn, SafeCross is the Answer

Online Safety Verification of Autonomous Driving Decision-Making Based on Dynamic Reachability Analysis

Autonomous Driving Risk Assessment with Boundary-based Environment Model

A Driving Risk Surrogate and Its Application in Car-Following Scenario at Expressway