InfoCom: Kilobyte-Scale Communication-Efficient Collaborative Perception with Information Bottleneck

TL;DR

InfoCom introduces a theoretically grounded, communication-efficient collaborative perception framework that significantly reduces data transmission from megabytes to kilobytes while maintaining near-perfect environmental perception for autonomous driving.

Contribution

It pioneers an information bottleneck-based approach with novel information purification and multi-scale decoding for minimal yet effective communication in collaborative perception.

Findings

Achieves 440-fold reduction in communication compared to Where2comm.

Reduces data transmission from MB to KB scale.

Maintains near-lossless perception accuracy.

Abstract

Precise environmental perception is critical for the reliability of autonomous driving systems. While collaborative perception mitigates the limitations of single-agent perception through information sharing, it encounters a fundamental communication-performance trade-off. Existing communication-efficient approaches typically assume MB-level data transmission per collaboration, which may fail due to practical network constraints. To address these issues, we propose InfoCom, an information-aware framework establishing the pioneering theoretical foundation for communication-efficient collaborative perception via extended Information Bottleneck principles. Departing from mainstream feature manipulation, InfoCom introduces a novel information purification paradigm that theoretically optimizes the extraction of minimal sufficient task-critical information under Information Bottleneck constraints. Its core innovations include: i) An Information-Aware Encoding condensing features into minimal messages while preserving perception-relevant information; ii) A Sparse Mask Generation identifying spatial cues with negligible communication cost; and iii) A Multi-Scale Decoding that progressively recovers perceptual information through mask-guided mechanisms rather than simple feature reconstruction. Comprehensive experiments across multiple datasets demonstrate that InfoCom achieves near-lossless perception while reducing communication overhead from megabyte to kilobyte-scale, representing 440-fold and 90-fold reductions per agent compared to Where2comm and ERMVP, respectively.