Abstract—The revolutionary advancements of Fifth Generation (5G) technology have redefined wireless communication systems, establishing a critical platform for integrating Artificial Intelligence (AI) with modern telecommunications. This paper emphasizes leveraging deep learning to enhance Automatic Modulation Recognition (AMR) within the physical layer, especially in noncooperative scenarios. Amid the increasingly complex and shared electromagnetic spectrum, AMR is crucial for efficient signal processing. Researchers address signal sparsity challenges by proposing Compressed Sensing (CS), enabling modulation identification in the compressed domain without full signal reconstruction. This paper introduces an innovative deep learning framework, CS- A Bidirectional Long Short-Term Memory (Bi-LSTM), combining CS with bidirectional long short-term memory networks. This architecture ensures high bandwidth signal acquisition via nonuniform low-rate sampling, excels in contextual feature extraction and addresses long-term dependencies. A SoftMax classifier further refines classification accuracy. Simulation results confirm that this groundbreaking approach surpasses existing deep learning models in AMR tasks, establishing a new benchmark for intelligent wireless communication systems.


Keywords—Fifth Generation (5G), Deep Learning (DL), Long Short-Term Memory (LSTM), Bidirectional Long Short-Term Memory (BI-LSTM), Automatic Modulation Recognition (AMR), Compressed Sensing (CS)


Cite: Hossam M. Kasem, Haithem S. Khallaf, and Sherief Hashima*, “Compressed Automatic Modulation Recognition Deep Learning Network Based on Bi-LSTM (CSBi-LSTM)," Journal of Communications, vol. 20, no. 3, pp. 358-370, 2025.


Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).