Edge-Assisted ML-Aided Uncertainty-Aware Vehicle Collision Avoidance at Urban Intersections

TL;DR

This paper addresses the critical challenge of preventing collisions at urban intersections by deploying an Intersection Manager at the MEC edge to fuse V2I data and predict vehicle trajectories with quantified uncertainty. It introduces an ML-aided, uncertainty-aware collision detection pipeline that combines light-weight LSTM-Encoder-Decoder models for trajectory and interval predictions with a Random Forest classifier to forecast collisions and trigger DENMs. By leveraging prediction intervals, the framework reduces false alarms and increases available reaction time, outperforming state-of-the-art baselines in SUMO-derived Luxembourg and Monaco scenarios. The results demonstrate sub-meter trajectory accuracy, robust uncertainty estimation, and effective preemptive collision avoidance, highlighting the practicality of edge-assisted AI for safer urban mobility. The work points toward federated learning and multi-maneuver extensions to further enhance privacy, scalability, and resilience in real-world deployments.

Abstract

Intersection crossing represents one of the most dangerous sections of the road infrastructure and Connected Vehicles (CVs) can serve as a revolutionary solution to the problem. In this work, we present a novel framework that detects preemptively collisions at urban crossroads, exploiting the Multi-access Edge Computing (MEC) platform of 5G networks. At the MEC, an Intersection Manager (IM) collects information from both vehicles and the road infrastructure to create a holistic view of the area of interest. Based on the historical data collected, the IM leverages the capabilities of an encoder-decoder recurrent neural network to predict, with high accuracy, the future vehicles' trajectories. As, however, accuracy is not a sufficient measure of how much we can trust a model, trajectory predictions are additionally associated with a measure of uncertainty towards confident collision forecasting and avoidance. Hence, contrary to any other approach in the state of the art, an uncertainty-aware collision prediction framework is developed that is shown to detect well in advance (and with high reliability) if two vehicles are on a collision course. Subsequently, collision detection triggers a number of alarms that signal the colliding vehicles to brake. Under real-world settings, thanks to the preemptive capabilities of the proposed approach, all the simulated imminent dangers are averted.

Figures (13)

• Figure 1: An edge computing-based scenario, including an IM collecting data generated by the vehicles and the road infrastructure and transmitting alarms in case of imminent collisions.
• Figure 2: An overview of the proposed network-assisted collision detection methodology.
• Figure 3: The encoder-decoder LSTM architecture used for vehicle trajectory estimation and uncertainty estimation.
• Figure 4: Collision avoidance - An overview of the latency components involved.
• Figure 5: Physical topology of the Luxembourg (left) and Monaco intersections (right). The layout highlights the distinct nature of the presented four-way intersections.