Partially Distributed Hybrid Precoding Design for mmWave Downlink Cell-Free mMIMO Systems

Cell-free massive MIMO (CF mMIMO) systems have become an attractive technology for enhancing the spectral efficiency of next-generation wireless communication systems due to their ability to effectively eliminate inter-cell interference. In the millimeter wave (mmWave) band, hybrid precoding technology can improve spectral efficiency while limiting the number of radio frequency (RF) chains per access point (AP), thereby reducing system power consumption. This paper focuses on the design of hybrid precoding for the downlink in mmWave CF mMIMO systems. We aim to maximize the system's weighted sum-rate and propose a partially distributed hybrid precoding framework. This framework solves each AP's hybrid precoding matrix by deploying a weighted minimum mean square error (WMMSE) algorithm at the central processing unit (CPU) and a matrix decomposition algorithm at each AP. Simulation results show that our proposed method achieves system weighted sumrate performance comparable to that of centralized algorithms, while offering lower computational complexity and reduced signaling overhead. Additionally, compared with distributed noncooperative precoding methods, our approach improves the sumrate by more than 50 %.