TY - JOUR
T1 - Multi strategy fusion enhanced channel estimation algorithm based on deep learning
AU - Liu, Xia
AU - Huang, Liang
AU - Mei, Xiaojun
AU - Saeed, Nasir
AU - Wang, Feng
AU - Zhang, Yuxuan
AU - Ma, Xue
AU - Weng, Congyan
N1 - Publisher Copyright:
© 2025 The Authors
PY - 2025/7
Y1 - 2025/7
N2 - The increasing frequency of maritime activities has fueled a growing demand for advanced wireless communication systems, making accurate channel estimation a crucial technology. Traditional channel estimation algorithms often face limitations when dealing with noise factors. To address this issue, we propose an enhanced channel estimation algorithm based on deep learning, which integrates multiple strategies and is named the IMBP algorithm. This method simulates the insertion of pilot signals at the receiving end and combines the efficiency of mean filter. Additionally, it utilizes random forests to optimize end-to-end information transmission and adjusts strategies through dynamic thresholds. Simultaneously, by incorporating the powerful feature learning capability of deep learning in channel estimation, it upgrades traditional linear mapping to nonlinear mapping. The simulation results demonstrate that the IMBP algorithm proposed in this paper significantly reduces BER in communication, demonstrating superior performance.
AB - The increasing frequency of maritime activities has fueled a growing demand for advanced wireless communication systems, making accurate channel estimation a crucial technology. Traditional channel estimation algorithms often face limitations when dealing with noise factors. To address this issue, we propose an enhanced channel estimation algorithm based on deep learning, which integrates multiple strategies and is named the IMBP algorithm. This method simulates the insertion of pilot signals at the receiving end and combines the efficiency of mean filter. Additionally, it utilizes random forests to optimize end-to-end information transmission and adjusts strategies through dynamic thresholds. Simultaneously, by incorporating the powerful feature learning capability of deep learning in channel estimation, it upgrades traditional linear mapping to nonlinear mapping. The simulation results demonstrate that the IMBP algorithm proposed in this paper significantly reduces BER in communication, demonstrating superior performance.
KW - Channel estimation
KW - Deep learning
KW - Maritime communication technology
KW - Strategy integration
UR - http://www.scopus.com/inward/record.url?scp=105004222263&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=105004222263&partnerID=8YFLogxK
U2 - 10.1016/j.asej.2025.103416
DO - 10.1016/j.asej.2025.103416
M3 - Article
AN - SCOPUS:105004222263
SN - 2090-4479
VL - 16
JO - Ain Shams Engineering Journal
JF - Ain Shams Engineering Journal
IS - 7
M1 - 103416
ER -