V2X Collective Perception Boosts FOV by 260% in New Study

A team led by Markos Antonopoulos and including researchers from the Institute of Communication and Computer Systems (ICCS) in Greece has published a paper detailing a hybrid validation methodology for V2X-enabled Collective Perception (CP) in complex traffic scenarios. The core innovation is a probabilistic Bayesian fusion algorithm that aggregates heterogeneous sensor observations—lidar, camera, radar—from multiple connected and autonomous vehicles into a shared probabilistic occupancy grid. Each cell in this grid encodes both occupancy likelihood and uncertainty, enabling explainable situational awareness beyond any single vehicle's field of view. The researchers bridge the gap between pure simulation and real-world evaluation using a hybrid testing framework that combines CARLA-based virtual environments with actual vehicle-in-the-loop experimentation.

In experimental results from a challenging roundabout scenario, the system achieved a 260% increase in field-of-view coverage compared to ego-only perception. Specifically, occupied-cell recall—a metric for detecting actual obstacles—rose from 0.82 (single vehicle) to 0.94 when six agents collaborated under nominal localization conditions. This work directly addresses the critical need for safe and certifiable deployment of cooperative autonomous vehicles, providing a reproducible, interpretable foundation for validating CP systems. The paper is set to be presented at ITS World 2026 and is available on arXiv under the subjects of Robotics and Multiagent Systems.