Graph Coloring-based Pilot Assignment Technique for Scalable Cell-Free Massive MIMO Systems

Abstract

Considering the rapidly increasing number of cellular users, ensuring a high data rate and delivering quality services in a 5G network would be challenging. Cell-free massive multiple-inputs multiple-outputs (mMIMO) emerges as a prominent solution to address this challenge. In cell-free mMIMO, multiple wireless access points work together to serve multiple users through coordination. Implementing cell-free mMIMO in real-world situations comes with challenges, and one notable concern is the ability to scale up in scenarios with high demand for accessing cellular network resources. Scala bility can be addressed by considering user-centric cell-free mMIMO and employing an effective pilot assignment technique for channel estimation. Various pilot assignment techniques have been proposed in the literature, such as random, interference-based K-means, and user-group pilot assignments. However, the practical implementation and complexity of mentioned pilot assignment techniques pose significant challenges in scalable cell-free mMIMO systems. This thesis proposes a pilot assignment tech nique that uses graph coloring to address these challenges. Simulation results show that the proposed graph coloring-based pilot assignment technique notably enhances the 90% likely spectral efficiency compared to conventional pilot assignment methods by reducing pilot contamination for uplink and downlink transmission by reducing pilot contamination and interference between users. Sensitivity studies are carried out to an alyze the performance of the proposed graph coloring-based technique with variations in the number of users and the number of pilots. The future implementation of a graph coloring-based pilot assignment technique can be employed for channel estimation in beyond 5G networks.