A Hybrid Heuristic-Aided Algorithm of Serial Cascaded Autoencoder and ALSTM for Channel Estimation in Millimeter-Wave Massive MIMO Communication System

Abstract

Channel estimation is a general issue for downlink transmission in millimeter-wave (mmWave) multiple-input multiple-output (MIMO) devices. To achieve the merits of mmWave massive MIMO devices, the channel state information (CSI) is very necessary. However, it is hard to attain the downlink CSI in the corresponding device, which results in training overhead. To overcome the particular issue, this paper proposes a new method as a serial cascaded autoencoder with attention-based long short-term memory (SCA-ALSTM), where the attributes are tuned using the iterative of reptile search and dingo optimizer (IRSDO) to derive the multiobjective function with multiple constraints such as root mean square error (RMSE), mean square error (MSE), normalized mean square error (NMSE), bit error rate (BER), and spectral efficiency (SE). The proposed SCA-ALSTM model leverages the power of attention mechanisms to focus on important information within the input data, allowing for more accurate channel estimation. By incorporating the IRSDO hybrid model, the SCA-ALSTM system can efficiently fine-tune the parameters to improve channel estimation accuracy while minimizing training overhead caused by evaluating a high amount of channel factors. Finally, the experimentation is accomplished with conventional algorithms and proved that the developed model helps to improve the channel estimation accuracy while reducing training overhead. By leveraging the developed model, channel estimation may be enhanced regarding accuracy and efficiency with reduced computational complexity. Moreover, it can better handle the complexities of non–line-of-sight (NLOS) channels, leading to improved estimation accuracy. Thus, the system outperforms the channel estimation to raise the efficiency.