Title: Enhancing Channel Estimation Performance for MIMO-OFDM in 5G Network Using Linear Minimum Mean Square Error
Authors: Chukwuka James Obagha, C.A. Nwabueze, C. N. Muoghalu
Volume: 8
Issue: 8
Pages: 100-111
Publication Date: 2024/08/28
Abstract:
In wireless networks such as the 5G, channel estimation can be used to improve reliability of communication. Channel estimation is a technique for determining the characteristics of channel and how transmitted signals are affected by it. This work presents enhanced channel estimation performance for MIMO-OFDM in 5G network using linear minimum mean square error (LMMSE) technique. It is desired to meet high quality performance at the receiver which is usually expressed in terms of bit error rate (BER). In order to achieve this, a MIMO-OFDM system with quadrature phase shift keying (QPSK) modulation for transmitting symbols over wireless channel was formulated in MATLAB. An LMMSE channel estimation algorithm was developed and implemented considering three different channel models such as additive white Gaussian noise (AWGN) channel model, MIMO simple channel model and sparse channel model, to evaluate its performance effectiveness. Simulations were conducted and the results obtained showed that the LMMSE provided BER of 0.00288 to 0.0 as the signal power, expressed in terms of SNR in dB, increases from 0 to 18 dB in AWGN channel. In simple channel and sparse channel, it yielded BER of 0.00351 to 0.000014 as the SNR increases from 0 to 18 dB and BER of 0.0036 to 0.000019 as the SNR increases from 0 to 18 dB respectively. Further performance evaluation to ascertain the superiority of LMMSE over other channel estimation schemes such as least square (LS) and minimum mean square error (MMSE) was conducted with respect to the three different channel models. Hence in AWGN channel, LMMSE yielded BER 0.00288 at 0 dB whereas LS and MMSE yielded 0.0280 and 0.0273 respectively. In simple channel, at 0 dB, LMMSE offered BER of 0.00351 while LS and MMSE provided 0.0341 and 0.0320 respectively. Also, in sparse channel, the LMMSE, LS and MMSE algorithms provided BER of 0.0036, 0.0346 and 0.0340 at 0 dB respectively. Generally, the best performance of the LMMSE was achieved in AWGN channel model but at the expense of complexity. In addition when compared with other estimators, the simulation results showed that LMMSE algorithm outperformed LS and MMSE algorithms in all the channel models for all values of SNR. Therefore, the LMMSE algorithm will offer significant gain in SNR by providing reduced bit error in wireless 5G compared with LS and MMSE.