Year: 2026
Type: article
Abstract: Automatic Modulation Recognition (AMR) is crucial and challenging for Cognitive Radio (CR) in 5G and 6G wireless communication scenario demanding with high speed and low latency. Existing Deep Learning (DL) based methods have achieved higher accuracy compared with traditional methods, without model solving the issue of Orthogonal Frequency Division Multiplexing (OFDM) system, which is widely applied in modern wireless communication scenarios. To address this issue, this paper proposes a specially designed Fourier Adaptive Filter Attention (FAFT) framework based on Frequency domain contrastive regularization and adaptive filter. The proposed method achieves to inject the character from time-series and frequency domain of signals specializing in different signal noise ratio (SNR) and different type of communication systems. Experimental results show that, its accuracy can reach 89.1% at SNR of 20 dB under urban channels with 91.8% at SNR of 14 for public dataset RML2016.10a and Multipath effect and Doppler effect for OFDM signals. Compared with state-of-the-art, the proposed method significantly reduces the computational complexity while maintaining recognition accuracy, demonstrating its significant feasibility in practical scenarios.
Source: Scientific Reports
Authors: Yuepeng Li, Xiaogang Tang, Lu Wang, Hongji Xing
Institution: Space Engineering University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,690
Award: G3061847787
OpenAlex ID: W7133846641