Integrated Sensory Platform: Driving Information Sharing in Multi-Level Autonomous Driving Mixed Traffic Mode

This article presents an innovative integrated solution called the Integrated Sensory Platform which combines driver and passenger behavior monitoring, vehicle blind spot detection, and V2X communication technology, all aimed at reducing traffic accidents in mixed-traffic scenarios involving multiple levels of automated driving. In this approach, deep learning visual AI technology is employed to identify the types of behaviors of occupants within the cabin, as well as the types of targets detected in blind spots. A three-branch algorithm framework is proposed, which enhances posture and temporal feature extraction, reduces the model's dependence on detailed features, and strengthens its ability to resist occlusion. Additionally, an expanded group channel reordering mechanism is designed to optimize the classification and recognition performance of the posture and temporal feature branches. The identified information is then broadcasted to surrounding vehicles via V2X communication, enabling deep information sharing among multiple vehicles in a spatial traffic environment. This information not only provides decision-making data for high-level automated driving vehicles but also offers operational guidance for fully human-driven vehicles, achieving the goal of collaborative decision-making among vehicles of varying automation levels and improving road traffic safety in complex mixed-traffic conditions.