When2com: Multi-Agent Perception via Communication Graph Grouping

TL;DR

This paper introduces When2com, a framework for multi-agent perception that learns to form communication groups and decide when to communicate, reducing bandwidth while maintaining high perception performance.

Contribution

It presents a novel distributed communication framework that dynamically groups agents and optimizes communication timing for multi-agent perception tasks.

Findings

Significantly reduces communication bandwidth.

Maintains superior perception performance.

Generalizes across different perception tasks.

Abstract

While significant advances have been made for single-agent perception, many applications require multiple sensing agents and cross-agent communication due to benefits such as coverage and robustness. It is therefore critical to develop frameworks which support multi-agent collaborative perception in a distributed and bandwidth-efficient manner. In this paper, we address the collaborative perception problem, where one agent is required to perform a perception task and can communicate and share information with other agents on the same task. Specifically, we propose a communication framework by learning both to construct communication groups and decide when to communicate. We demonstrate the generalizability of our framework on two different perception tasks and show that it significantly reduces communication bandwidth while maintaining superior performance.