.png){kind=icon}
.png){kind=icon}
Return
Multi-Domain Dynamic Filtering Network for Multicarrier Modulation Recognition
.png){kind=icon}
PRE
AI
DOI：10.1109/LWC.2025.3616110.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
With the rapid evolution of multicarrier technologies, the coexistence of various multicarrier modulations (MCMs) has emerged as a critical scenario in modern wireless systems. The inevitable inter-user interference arising from spectral resource scarcity makes MCM recognition critical for efficient spectrum monitoring. However, current algorithms predominantly rely on single-domain representations, failing to exploit cross-domain complementary information and consequently suffering from limited recognition accuracy. To address this challenge, we propose the multi-domain dynamic filtering network which incorporates delay-Doppler domain analysis together with both time domain and time-frequency domain information, achieving significant accuracy improvements in multicarrier signal recognition. Furthermore, a dynamic filtering module is proposed to adaptively amplify critical frequency components through frequency domain filtering, thereby enhancing feature discriminability. Finally, simulations adopting a comprehensive dataset including state-of-the-art MCM schemes demonstrate that the proposed methodology achieves superior performance gain in complex wireless scenarios, when compared with only time domain or time-frequency domain method.
Keywords:
Multicarrier modulation recognition
affine frequency division multiplexing
feature fusion
deep learning
Journal
I
IF:
5.5
Papers:
665
Citations:
0