Decentralized Semantic Communication and Cooperative Tracking Control for a UAV Swarm Over Wireless MIMO Fading Channels

Conventional communication strategies in UAV swarms often lead to excessive bandwidth consumption and energy overhead due to frequent exchange of control signals. Inspired by the semantic communication paradigm—which emphasizes transmitting only task-relevant information—we propose a cooperative semantic communication-control framework that selectively transmits the most informative control data. This approach significantly reduces communication burden and power consumption while maintaining accurate swarm coordination. Specifically, we consider a UAV swarm composed of one leader and multiple followers, interconnected through unreliable MIMO wireless channels. We first develop a dynamic model that captures both inter-UAV interactions and MIMO channel imperfections. Incorporating power costs, we formulate the joint communication and cooperative tracking control problem as a drift-plus-penalty optimization. A closed-form decentralized solution is then derived, adapting to tracking errors and local channel conditions. Using Lyapunov drift analysis, we establish sufficient conditions for swarm stability. Numerical simulations demonstrate that the proposed scheme substantially outperforms existing methods in both tracking accuracy and communication efficiency.