Deep Learning Channel Estimation for OFDM 5G Systems with Different Channel Models

Abstract At cellular wireless communication systems, channel estimation (CE) is one of the key techniques that are used in Orthogonal Frequency Division Multiplexing modulation (OFDM). The most common methods Decision‐Directed Channel Estimation, Pilot-Assisted Estimation (PACE) and blind estimation. Among them, PACE commonly has a steadier performance. Applying deep learning (DL) CE getting increasing interest researchers during past 3 years. main objective this paper to assess efficiency DL-based compared conventional techniques, such as least-square (LS) minimum mean-square error (MMSE) estimators. A simulation environment evaluate OFDM performance at different models been used. DL process estimates from training data also employed get estimated impulse response channel. Two have comparison: Tapped Delay Line Clustered models. evaluated under parameters including number pilots (64 or 8 pilots), subcarriers (64), length cyclic prefix (16 0 samples) carrier frequency (4 GHz) through computer using MATLAB. From results, trained estimator provides better results estimating detecting transmitted symbols LS MMSE estimators although, complexity proposed LSTM exceeds equivalent estimator. Furthermore, demonstrates its effectiveness with various pilot densities periods.