Comparative Performance Analysis of Channel Estimation Techniques for Massive MIMO System

Abstract

The need for an internet connection is growing universally, so people need much higher data rate connection to meet their need but every physical resource in communication like frequency band, transmit signal strength are finite. Within the given limited resource, higher data speed is accomplished by a new technology called massive Multiple Input Multiple Output (massive MIMO) system. Massive MIMO fulfills the high data rate requirement through antenna diversity gain. It is one of 5G wireless network technology with array antennas at both transmitter and receiver sides to providing high spectral and energy efficiencies. In massive MIMO, the signal obtained by the receiver is in different phase and amplitude from the transmission signal. Therefore, the system quality is highly depending on the accuracy of the channel estimation. Channel estimation plays a significant role in the performances of massive MIMO system because the dimension of the channels matrix is large, phase changes and noises are added when the signals pass through channel, these reduces the efficiencies of the whole system. To solve these problems, this Thesis focuses on the comparative analysis of pilot-based channel estimation schemes for massive MIMO system, which includes: Minimum Mean square Error (MMSE), Element Wise Minimum Mean Square Error (EW-MMSE), Maximum Likelihood (ML) and Least Square (LS) estimator with respect to Normalized Mean square Error (NMSE), Signal to Noise Ratio (SNR), Spectral Efficiency (SE), number of BS antennas, and number of computational complexities. Based on the simulation results, MMSE channel estimator has the best performance, followed by EW-MMSE, ML and LS channel estimators in terms of NMSE and SNR. And depending on the number of computational complexities of each coherent block as a function of BS antennas M with constant UE, the complexity of LS is less than that of MMSE, and it has almost the same complexity as ML and EW-MMSE. And also, according to SE with respect to BS antenna M, the MMSE provides the highest SE using the highest complexity, EW-MMSE and ML achieves a good balance between SE and complexity, and LS has the lowest complexity, but also provides the lowest SE. Finally, the result confirms that MMSE channel estimation technique has the best performance compared to EW-MMSE, ML and LS channel estimators with the cost of high complexity