Clustering-based Pilot Assignment for User-Centric Cell-Free Mmwave Massive MIMO Systems

Bowen Zhong,Xu Zhu,Eng Gee Lim

DOI: https://doi.org/10.1109/vtc2022-fall57202.2022.10012778

2022-01-01

Abstract:The cell-free (CF) massive multiple-input multiple-output (MIMO) system, combined with millimeter wave (mmWave), is considered as one of the most promising enablers for future wireless networks. However, with the increase of user equipments (UEs) in the network, the pilot sequence reuse will bring interference during the channel estimation period, which is called pilot contamination. In this paper, an efficient pilot assignment scheme called the k-means clustering-based pilot assignment (KCPA) is proposed to mitigate the severe pilot contamination in user-centric (UC) CF massive MIMO systems at mmWave frequencies. By exploiting the knowledge of UE locations, the pilot contamination could be reduced through structured pilot assignment based on the UE clustering result. In addition, a pilot assignment scheme based on the Tabu search, called the k-means clustering-based Tabu search pilot assignment (KCTSPA) is devised to further improve the performance. Numerical results verify that the proposed pilot assignment schemes can greatly reduce the uplink channel estimation error of mmWave UC systems with a low complexity compared to the conventional pilot assignment schemes.