Comparative Analysis of LS vs MMSE Estimators for SISO OFDM Systems

This research conducts a comparative study between LS and MMSE estimators to enhance the performance of SISO OFDM systems. The LS estimator employs the least squares method for signal estimation, while the MMSE estimator minimizes the mean square error to achieve optimal signal estimation. In code implementation, the LS estimator typically involves simple matrix operations: Ĥ_LS = Y/X where Y represents the received signal and X denotes the transmitted pilot symbols. This approach offers computational simplicity but exhibits sensitivity to noise. The MMSE estimator requires more complex calculations incorporating channel statistics: Ĥ_MMSE = R_hy * inv(R_yy) * Y where R_hy is the cross-correlation matrix between channel and received signal, and R_yy is the autocorrelation matrix of the received signal. This method demonstrates superior noise robustness but demands prior knowledge of channel characteristics. This study comprehensively compares the advantages and disadvantages of both estimators, analyzing their bit error rate performance under varying signal-to-noise ratio conditions. The research aims to identify the optimal estimator for specific system requirements, considering implementation complexity, computational resources, and performance trade-offs in practical OFDM deployments.