INVESTIGATION OF CHANNEL ESTIMATION METHODS FOR MASSIVE MIMO SYSTEM WITH PILOT DECONTAMINATION IN RICIAN FADING

Abstract

Massive multiple input multiple output (MIMO) systems are essential to meet the growing need for enhanced capacity and faster data rates in modern wireless communication networks, particularly in the fifth generation (5G). These systems serve many users at once on the same frequency band by using a lot of antennas at the base station (BS). The computational complexity of many antenna numbers and pilot contamination are the main obstacles to estimating the channel of the massive MIMO system. To overcome pilot contamination challenges, this thesis aims to investigate pilot-based channel estimate methods with pilot decontamination in Rician fading. In this thesis, soft pilot reuse is combined with the weighted graph coloring pilot decontamination mechanism to overcome the impact of pilot contamination in Rician fading scenarios, which is the main weakness in massive MIMO systems. In both Rayleigh and Rician fading, the accuracy of the channel estimate is increased by mitigating the effects of pilot contamination. MATLAB simulations are used to examine how well the minimum mean squared error (MMSE), element-wise minimum mean squared error (EW-MMSE), and least squares (LS) techniques work with the Rayleigh and Rician fading channels model. The simulation result shows that the better method for channel estimation in both uplink and downlink situations is MMSE with Rician fading. In the uplink system, spectral efficiency (SE) of the channel estimate is increased from 20.9 b/s/Hz/cell to 24.7 b/s/Hz/cell with pilot decontamination at BS antenna M = 90. For downlink, it increased from 21.8 b/s/Hz/cell to 27.5 b/s/Hz/cell. In Rayleigh fading, for the uplink system, SE increased from 14.0 b/s/Hz/cell to 16.5 b/s/Hz/cell at M = 90 with pilot decontamination,whereas in the downlink it improved from 14.2 b/s/Hz/cell to 17.5 b/s/Hz/cell. The results show that the addition of decontamination techniques increases the accuracy of channel estimates, hence improving system performance for massive MIMO networks. So, MMSE are useful for utilizing channel characteristics, while the proposed method (soft pilot reuse combined with weighted graph coloring) is a reliable way to reduce the impact of pilot contamination on Rayleigh and Rician fading.