Free space optical turbulent channel estimation based on the deep combined CNN and BiLSTM network

Abstract Free space optical (FSO) communication systems are highly susceptible to atmospheric turbulence, which significantly impacts channel state estimation and system performance. Traditional channel estimation methods, such as Least Squares (LS), Linear Minimum Mean Square Error (LMMSE), and Extended Kalman Filter (EKF), often fail to handle the nonlinear characteristics of channels under varying turbulence conditions. This paper introduces a novel deep learning–based channel estimation method combining Convolutional Neural Networks (CNN) and Bidirectional Long Short-Term Memory Networks (BiLSTM), trained to predict Channel State Information (CSI) for gamma–gamma distributed FSO channels. Simulations are conducted under weak, moderate, and strong turbulence, demonstrating that the proposed method significantly outperforms classical approaches in terms of normalized mean square error (NMSE) and bit error rate (BER). The results show that the proposed method achieves robust performance, especially under strong turbulence, with lower computational complexity compared to iterative methods like EKF.