DDPG-Based Optimization for Zero-Forcing Transmission in UAV-Relay Massive MIMO Networks

• Published in: IEEE open journal of the Communications Society Vol. 5; pp. 2319 - 2332
• Main Authors: Le, Mai T. P., Nguyen-Duy-Nhat, Vien, Nguyen, Hieu V., Shin, Oh-Soon
• Format: Journal Article
• Language: English
• Published: New York IEEE 2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; analysis; Array signal processing; Autonomous aerial vehicles; Base stations; Lower bounds; Machine learning; Massive MIMO; Millimeter wave communication; MIMO communication; Optimization; Radio equipment; Relay; Relays; Unmanned aerial vehicles; Wishart matrices; zero-forcing
• ISSN: 2644-125X; 2644-125X
• DOI: 10.1109/OJCOMS.2024.3386595

This study explores the advantages of employing an unmanned aerial vehicle (UAV) in a massive multiple-input multiple-output (MIMO) network with zero-forcing processing at the base station (BS). Considering potential inaccuracies in channel estimation, we derive a closed-form expression for lower bounds on spectral efficiency in the massive MIMO system, utilizing the UAV as an aerial relay. Subsequently, we formulate a comprehensive optimization problem that encompasses UAV placement and user power allocation in the downlink network, aiming to maximize the data rate for terrestrial users. To address the optimization problem, we propose a novel deep learning-based algorithm that jointly optimizes UAV positioning and power allocation. Finally, we present numerical results that not only validate our theoretical framework and but also demonstrates the effectiveness of the proposed approach.