New protocol lets drone swarms reshape dynamically without hierarchy

Researchers Tao He and Gangshan Jing have introduced a new distributed control method for multi-agent systems—like drone swarms or robot teams—that allows formations to dynamically reshape by compressing or stretching along different axes. The core challenge in such "open" systems is that agents can join or leave, and communication links can change, which disrupts the Laplacian matrix that encodes the desired formation shape. Their approach, called Laplacian spectral shaping, strategically adjusts partial weights of this matrix to preserve critical spectral properties: positive semidefiniteness, the correct kernel, and positive definiteness of the follower block.

The proposed protocols work in arbitrary dimensional space and handle agent joining, edge addition, agent leaving, and edge removal without requiring a predefined parent-child hierarchy. Any two agents can serve as leaders, and the sensing graph is sparser than in existing affine formation control methods. Both theoretical analysis and numerical simulations validate the method's effectiveness. This research, submitted to arXiv on July 18, 2026, and available under arXiv:2607.16709, opens the door to more flexible, resilient control of autonomous formations in complex environments.