HydraCollab: Adaptive Collaborative-Perception for Distributed Autonomous Systems

2026-06-30

AI Summary

This work addresses the inherent trade-off between communication bandwidth and perception accuracy in distributed autonomous systems. To overcome this challenge, the authors propose an adaptive collaborative perception framework that employs a spatial confidence map–driven feature selection mechanism to transmit only high-information features and dynamically switches between intermediate and late fusion strategies to optimize collaboration efficiency. The approach significantly reduces communication overhead while simultaneously improving perception performance. Evaluated on the V2X-R and V2X-Radar datasets, the method achieves state-of-the-art accuracy–communication trade-offs, delivering performance gains of 0.78% and 0.75%, respectively, while using only 41% and 26% of the bandwidth required by the Where2comm baseline.

Abstract

Collaborative-perception enables multi-robot systems to enhance situational awareness by sharing perceptual information. Existing collaborative-perception systems face an inherent trade-off between communication bandwidth requirements and perception accuracy, where methods that exchange more information achieve better perception results at the cost of increased communication overhead. However, real-world communication networks impose bandwidth constraints that require minimizing communication overhead without sacrificing perception performance. To address this challenge, we propose HydraCollab, an adaptive collaborative-perception framework that (i) selectively transmits the most informative sensor features and (ii) dynamically employs collaboration strategies (intermediate or late) based on spatial confidence maps. Extensive evaluations on the V2X-R, V2X-Radar and UAV3D-mini datasets demonstrate that HydraCollab achieves the best overall trade-off between accuracy and communication cost among existing collaborative-perception methods. Relative to SOTA Where2comm, HydraCollab uses only 41% of the bandwidth on V2X-R and 26% on V2X-Radar while improving performance by 0.78% and 0.75% respectively. Our code and models are available at https://github.com/AICPS/HydraCollab.

Problem

• collaborative perception
• communication bandwidth
• perception accuracy
• distributed autonomous systems
• bandwidth constraints

Innovation

• adaptive collaborative perception
• feature selection
• dynamic fusion strategy
• bandwidth efficiency
• spatial confidence map