\date \textsuperscript1Department of Astronomy, University X, City, Country\\ \textsuperscript2Institute of Data Science, University Y, City, Country\\ \textsuperscript3Observatory Z, City, Country\\ *Corresponding author: email@domain.com

: At each discrete timestep, the current token holder ( a_i ) selects a neighbor ( a_j ) (randomly or via a utility‑based heuristic) and transmits ( \tau ).

I can help you search for alternative software solutions or technical products if you provide more details about what the software is supposed to do. Share public link

Swarm‑based Dynamic Surveillance Systems (SDSS) have emerged as a promising paradigm for large‑scale, resilient, and adaptive monitoring of complex environments. However, the integration of heterogeneous sensor modalities across dozens to hundreds of autonomous agents remains a bottleneck, particularly when operating under stringent bandwidth, power, and latency constraints. This paper introduces , a lightweight, hierarchical sensor‑fusion architecture that leverages probabilistic graphical models, edge‑computing primitives, and a novel “fusion‑token” protocol to achieve near‑optimal situational awareness while respecting real‑time constraints. We present a detailed system model, formal proofs of convergence, a suite of simulation experiments, and a hardware‑in‑the‑loop (HIL) validation on a fleet of 48 quadrotor platforms equipped with visual, infrared, acoustic, and LiDAR sensors. Results demonstrate a 43 % reduction in communication overhead , a 27 % improvement in detection latency , and robustness to up to 35 % node failures , outperforming state‑of‑the‑art decentralized fusion baselines. We conclude with a discussion of open research directions, including adaptive token routing, privacy‑preserving fusion, and cross‑domain transfer learning.

\titleA Comprehensive Study of FSDSS‑548: [Brief descriptive subtitle] \author First Author\textsuperscript1,*, Second Author\textsuperscript2, Third Author\textsuperscript3

A2: The director is not publicly credited in the work's metadata.