When2com: Multi-Agent Perception via Communication Graph Grouping

By Yen-Cheng Liu, Junjiao Tian, Nathan Glaser, Zsolt Kira. Computer Vision and Pattern Recognition, 2020

340 citations recorded by the source corpus.

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.

Subfields: Computer Science Applications, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/73c7f1ac1b530dc4ca3f7ef7a26cacc14962f6db

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Science Applications

1. Ali Farhadi
2. Joost van de Weijer

In Computer Vision and Pattern Recognition

1. Aniruddha Kembhavi
2. Luc Van Gool
3. Ali Farhadi
4. Abhinav Kumar Gupta
5. Kristen Grauman
6. Li Fei-Fei
7. Jitendra Malik
8. Xiaogang Wang

Citations in the corpus

• Where2comm: Communication-Efficient Collaborative Perception via Spatial Confidence Maps (2022, methodology): “Method Venue Message packing Communication graph construction Message fusion When2com [12] CVPR 2020 Full feature map Handshake-based sparse graph Attention per-agent V2VNet [1] ECCV 2020 Full feature map Fully connected graph Average per-agent DiscoNet [2] NeurIPS 2021 Full feature map Fully connec”
• A Cordial Sync: Going Beyond Marginal Policies for Multi-Agent Embodied Tasks (2020, background): “Collaborative perception for semantic segmentation and recognition classification have also been investigated recently [55,56].”
• Learning Distilled Collaboration Graph for Multi-Agent Perception (2021, background): “Comparing to the pose-aware When2com, DiscoNet improves by 25.7% in AP@0.5 and 25.9% in Table 3: Comparison of various intermediate collaboration strategies.”