Enhancing Traffic Safety with AI and 6G: Latency Requirements and Real-Time Threat Detection

TL;DR

This work tackles real-time traffic safety under a 6G/AI paradigm by decomposing the end-to-end latency $T_{tot}$ into sensor, network, processing, alarm, validation, and actuation stages. It proposes a CNN-based threat-detection pipeline (Darknet with YOLOv4) integrated with edge processing to classify threats and disseminate alerts, while explicitly budgeting $T_S$, $T_P$, and $T_C$ and deriving the remaining network latency budget $T_{eval}+T_{exe}$ for 6G links. Real-time targets are grounded in a $T_{tot}$ window of $150$–$200$ ms, with empirical measurements showing $T_S=22$ ms, $T_P=76.39$ ms, and $T_C=1$ ms; yielding $T_{eval}=T_{exe}=25.20$ ms under certain conditions. The results demonstrate predictable AI processing times (AI inter-frame times around $72$–$79$ ms with occasional outliers) and ultra-fast thread classification (<1 ms), supporting practical deployment of 6G-enabled emergency-vehicle warnings and broad traffic-safety alerts in smart cities.

Abstract

The rapid digitalization of urban infrastructure opens the path to smart cities, where IoT-enabled infrastructure enhances public safety and efficiency. This paper presents a 6G and AI-enabled framework for traffic safety enhancement, focusing on real-time detection and classification of emergency vehicles and leveraging 6G as the latest global communication standard. The system integrates sensor data acquisition, convolutional neural network-based threat detection, and user alert dissemination through various software modules of the use case. We define the latency requirements for such a system, segmenting the end-to-end latency into computational and networking components. Our empirical evaluation demonstrates the impact of vehicle speed and user trajectory on system reliability. The results provide insights for network operators and smart city service providers, emphasizing the critical role of low-latency communication and how networks can enable relevant services for traffic safety.

Figures (7)

• Figure 1: Topological overview of AI-based traffic control system.
• Figure 2: Illustration of an intra-cell intersection. The user and sensor trajectories intersect within the boundaries of the cell.
• Figure 3: Example of a $3 \times 3$ grid. The user's cell is at the centre while neighbouring cells represent potential inter-cell intersections.
• Figure 4: Traffic scenario of extended view using 6G and AI enabling technology.
• Figure 5: Selection of reference images used. The labelled boxes indicate vehicle classification.