Implicit Virtual Leader: Decentralized Vision-Only Relative Pose Estimation for Multi-Robot Formations
Robotics
Summary
The authors present a new way for robot groups to know their positions relative to each other without relying on GPS or fixed leaders. They use a special kind of neural network called a Graph Neural Network to create a shared, invisible reference point called an implicit virtual leader. This system only uses camera images and communication between robots and can handle different robot types and group sizes. They also include methods to estimate uncertainty in their position predictions. Their approach works well in both simulations and real-world tests.
Authors
Shiyuan Yang, Zelin Wang, Zhijia Tao, Yilin Wang, Zhengyu Hou, Xiaosong Kong, Borong Zhang, Yip Fun Yeung, Yuankai Luo, Sharon Lee, Qingbiao Li
Abstract
Classical leader-follower formation control suffers from single points of failure and error propagation, and relies on absolute localization sensors that are ill-suited for GPS-denied environments. We address these limitations by introducing a fully decentralized, vision-only relative pose estimation framework based on Graph Neural Networks (GNNs). The key idea is the implicit virtual leader (IVL): a non-physical formation reference frame that is not tied to any individual robot but is implicitly learned within the GNN using only monocular images and inter-robot communication. We attach a heteroscedastic GNLL head for aleatoric uncertainty and MC~Dropout for epistemic uncertainty, and conduct a systematic comparison across simulation and real-world test sets. Our framework achieves competitive pose estimation accuracy and generalizes naturally to heterogeneous robot platforms and varying formation sizes.