When Drones Meet Privately: Secure Coordination with t-PSI

Autonomous multi‑agent systems increasingly operate in environments where privacy, anonymity, and operational security are critical. In such settings, drones must often determine which events or locations have been independently confirmed by at least $t$ participants, while revealing nothing about individual trajectories, identities, or exploration patterns. Threshold Private Set Intersection ($t$‑PSI) naturally captures this requirement, but existing constructions rely on persistent identities, authenticated channels, or linkable communication, making them unsuitable for high‑risk deployments where structural anonymity is essential. We introduce a fully anonymous $t$‑PSI protocol tailored for distributed exploration tasks. Communication occurs in a You‑Only‑Speak‑Once (YOSO) architecture: each participant sends exactly one anonymous, unlinkable message per round, ensuring confidentiality, integrity, and sender anonymity without authenticated channels. The protocol supports multi‑agent exploration scenarios, such as drone‑based reconnaissance, where drones must collaboratively identify zones explored by a minimum subset of units while preserving full operational privacy. We formally prove security against semi‑honest participants and active man‑in‑the‑middle adversaries interacting with the network. Our analysis shows that the protocol achieves threshold privacy, ciphertext unlinkability, sender anonymity, and robustness against message tampering. This work demonstrates that anonymous threshold‑PSI can be realized efficiently in a YOSO framework, enabling privacy‑preserving coordination in adversarial and strategically sensitive environments.