JOINT USER CLUSTERING AND PSEUDO RANDOM BASED PILOT DECONTAMINATION IN MASSIVE MIMO TDD SYSTEM

Abstract

Wireless communication technology is now progressing to 5G and beyond. The number of wireless connections is growing in line with the growing number of subscribers and demands. One of the promising technology to overcome this is the usage of massive Multiple Input Multiple Output systems (M-MIMO. However, M-MIMO system’s performance suffers as a result of the random nature of the wireless channels. In Pilot based channel estimation there is a problem of pilot contamination due to the reuse of pilot sequences, which is the bottleneck for M MIMO systems. This problem arises when the number of user terminals grows, they compute for the limited pilot sequence with minimum coherence interval in channel estimation. Therefore, users’ channels to the BS need to be effectively estimated by applying proper pilot assignment method to achieve the performance of M-MIMO. In this thesis, joint user clustering and pseudo random sequence based pilot de contamination for M-MIMO TDD system is proposed. The K-means clustering algorithm is used to cluster users in the cell depending on the severity of the problem affecting each user: cell center users and cell edge users. The same pilot sequences are reused to the cell center users and orthogonal pilot sequences are assigned to the cell edge users. Pseudo random sequences are used to improve the orthogonality of pilot sequence of cell center users of different cells. At the BS, the MMSE channel estimation has been used. Then, using the estimated channels, multi cell MMSE detectors have been used to detect data from users in the uplink. The performance of the proposed system is analyzed based on the following met rics: NMSE, SINR, and target cell’s SE. To simulate the results Matlab software has been used. The simulation results of the proposed system when compared to random pilot assignment, it shows that for cell edge users and cell center users the normalized mean square error(NMSE) of the estimated channel has improved by 7.76% and 19.021%, respectively. Furthermore, the proposed technique improves the target cell’s SE by 8.714%.