Performance Analysis of LS and LMMSE Channel Estimation Techniques for LTE Downlink Systems

TL;DR

This paper compares the performance of LS and LMMSE channel estimators in LTE downlink systems, analyzing their effectiveness under different channel lengths and SNR conditions through MATLAB simulations.

Contribution

It provides a detailed performance comparison of LS and LMMSE estimators considering varying cyclic prefix lengths and channel conditions in LTE systems.

Findings

LMMSE outperforms LSE when cyclic prefix is longer or equal to channel length at high SNR.

LMMSE performs better at low SNR when cyclic prefix is shorter than channel length.

LSE shows better performance at high SNR with shorter cyclic prefix.

Abstract

The main purpose of this paper is to study the performance of two linear channel estimators for LTE Downlink systems, the Least Square Error (LSE) and the Linear Minimum Mean Square Error (LMMSE). As LTE is a MIMO-OFDM based system, a cyclic prefix is inserted at the beginning of each transmitted OFDM symbol in order to completely suppress both inter-carrier interference (ICI) and inter-symbol interference (ISI). Usually, the cyclic prefix is equal to or longer than the channel length but in some cases and because of some unforeseen channel behaviour, the cyclic prefix can be shorter. Therefore, we propose to study the performance of the two linear estimators under the effect of the channel length. Computer simulations show that, in the case where the cyclic prefix is equal to or longer than the channel length,LMMSE performs better than LSE but at the cost of computational complexity.In the other case, LMMSE continue to improve its performance only for low SNR values but it degrades for high SNR values in which LS shows better performance for LTE Downlink systems. MATLAB Monte-Carlo simulations are used to evaluate the performance of the studied estimators in terms of Mean Square Error (MSE) and Bit Error Rate (BER) for 2x2 LTE Downlink systems.