Distributed Fairness-Aware Joint Sensing and Communication in UAV Swarms via Hybrid MARL-ADMM

Abstract

Integrated sensing and communication (ISAC) is a key enabler of sixth-generation (6G) wireless systems, allowing communication and sensing to share hardware and spectrum resources. Unmanned aerial vehicle (UAV) swarms provide a flexible ISAC platform through the use of adaptive threedimensional mobility and distributed coordination. However, jointly optimizing communication throughput, sensing quality, and energy efficiency under realistic hardware impairments will lead to a non-convex and highly-coupled problem. This paper proposes an AI-enhanced, fairness-aware joint sensingcommunication (JSC) framework for UAV swarms operating under phase noise, timing jitter, and power amplifier nonlinearity via an impairment-aware SINR model. A hybrid multi-agent reinforcement learning (MARL)-ADMM architecture enables scalable, distributed optimization under dynamic channel conditions.