Bandwidth-Efficient Communication Modelling for Autonomous Vehicle Collaborative Perception

Dinghao Jin, Yuan Zeng, Yi Gong; Proceedings of the Winter Conference on Applications of Computer Vision (WACV), 2025, pp. 6146-6155

Accurate perception systems are crucial for intelligent vehicle motion planning and control. Cooperative perception can significantly enhance the perception capabilities of autonomous vehicles by sharing information beyond the field of view of individual vehicles. However most recent studies on cooperative perception have overlooked the bandwidth limitations of information sharing. To address this issue this paper presents a novel multi-agent cooperative 3D object detection framework to improve the detection accuracy of connected autonomous vehicles (CAVs) in bandwidth-constrained wireless communication environments. The framework includes an attention-based communication mechanism to select optimal communication partners and a foreground-background separation strategy to prioritize key information. The effectiveness of the proposed framework is evaluated on three benchmark datasets CARLA-3D V2XSet and OPV2V demonstrating the state-of-the-art performance-bandwidth trade-off.