Similar articles

Abstract: Signal-to-noise ratio (SNR) fluctuations significantly affect spectrum sensing performance in wireless communications. Traditional convolutional neural network (CNN) exhibits limited feature extraction capabilities and inefficient feature utilization at low SNR levels, leading to suboptimal spectrum sensing performance. This paper proposes a spectrum sensing method based on a multi-scale feature fusion network (MSFFNet) to address this issue. First, the proposed method employs a multi-scale feature extraction block (MSFEB) to capture multi-scale information from the input data comprehensively. Next, an adaptive feature screening strategy (AFSS) highlights key features while suppressing redundant information. Finally, a multi-level feature fusion mechanism (MLFFM) optimizes and integrates features across scales and levels, enhancing spectrum sensing performance. Simulation results demonstrate that compared to other methods, the proposed approach achieves superior performance in low-SNR communication scenarios. At an SNR of -14 dB, the detection probability P_d reaches 0.936, while the false alarm probability P_fa is only 0.1. Furthermore, this paper constructs a multi-level mixed-SNR dataset to simulate real communication environments and enhance the robustness of spectrum sensing.

[1]Bai WW, Zheng GQ, Mu Y, et al., 2025. Cooperative spectrum sensing method based on channel attention and parallel CNN-LSTM. Dig Signal Process, 158:104963.

[2]Cai LN, Cao KT, Wu YP, et al., 2022. Spectrum sensing based on spectrogram-aware CNN for cognitive radio network. IEEE Wirel Commun Lett, 11(10):2135-2139.

[3]Dosovitskiy A, Beyer L, Kolesnikov A, et al., 2021. An image is worth 16×16 words: transformers for image recognition at scale. Proc 9^th Int Conf on Learning Representations.

[4]Gai JX, Xue XF, Wu JY, et al., 2021. Cooperative spectrum sensing method based on deep convolutional neural network. J Electron Inform Technol, 43(10):2911-2919 (in Chinese).

[5]Gai JX, Xue XF, Nan RX, et al., 2022. Spectrum sensing method based on residual dense network. J Commun, 42(12):182-191 (in Chinese).

[6]Gao ZH, Li YF, Chen Z, et al., 2024. Intelligent spectrum sensing of consumer IoT based on GAN-GRU-YOLO. IEEE Trans Consumer Electron, 70(3):6140-6148.

[7]Geng Y, Huang JY, Yang JX, et al., 2022. Spectrum sensing for cognitive radio based on feature extraction and deep learning. J Phys Conf Ser, 2261(1):012016.

[8]Gou JP, Chen Y, Yu BS, et al., 2024. Reciprocal teacher-student learning via forward and feedback knowledge distillation. IEEE Trans Multimed, 26:7901-7916.

[9]He ZH, Jin G, Wang YJ, 2022. A novel grey wolf optimizer and its applications in 5G frequency selection surface design. Front Inform Technol Electron Eng, 23(9):1338-1353.

[10]Hong S, Xu WQ, 2025. Spectrum sensing in very low SNR environment using multi-scale temporal correlation perception with residual attention. Sensors, 25(2):528.

[11]Hou QB, Zhou DQ, Feng JS, 2021. Coordinate attention for efficient mobile network design. Proc IEEE Conf on Computer Vision and Pattern Recognition, p.13713-13722.

[12]Hu J, Shen L, Sun G, 2018. Squeeze-and-excitation networks. Proc IEEE Conf on Computer Vision and Pattern Recognition, p.7132-7141.

[13]Janu D, Singh K, Kumar S, et al., 2023. Hierarchical cooperative LSTM-based spectrum sensing. IEEE Commun Lett, 27(3):866-870.

[14]Jiang W, Zhou KQ, Sarkheyli-Hägele A, et al., 2022. Modeling, reasoning, and application of fuzzy Petri net model: a survey. Artif Intell Rev, 55(8):6567-6605.

[15]Kim H, Kim YJ, Kim WT, 2023. Multitask learning-based deep signal identification for advanced spectrum sensing. Sensors, 23(24):9806.

[16]Li JC, Fang FM, Mei KF, et al., 2018. Multi-scale residual network for image super-resolution. Proc 15^th European Conf on Computer Vision, p.527-542.

[17]Li JX, Wang YJ, Liu Z, et al., 2025. Metaformer-based lightweight neural network for massive MIMO CSI feedback. IEEE Wirel Commun Lett, 14(2):275-279.

[18]Li SP, Han YZ, Gaber J, et al., 2024. A multiantenna spectrum sensing method based on HFDE-CNN-GRU under non-Gaussian noise. Int J Intell Syst, 2024:1085161.

[19]Liu C, Wang J, Liu XM, et al., 2019. Deep CM-CNN for spectrum sensing in cognitive radio. IEEE J Sel Areas Commun, 37(10):2306-2321.

[20]Ma QY, Xiao J, Shao ZQ, et al., 2025. Frequency-learning adversarial networks based on transfer learning for cross-scenario signal modulation classification. Front Inform Technol Electron Eng, 26(5):816-832.

[21]Pan GL, Wu QH, Zhou B, et al., 2025. Spectrum prediction with deep 3D pyramid vision transformer learning. IEEE Trans Wirel Commun, 24(1):509-525.

[22]Peng TQ, Yang SY, Feng ZX, et al., 2024. Spectrum sensing via residual dilated network and horizontal shift attention for cognitive IoT. IEEE Internet Things J, 11(22):36817-36828.

[23]Pham VT, Zniyed Y, Nguyen TP, 2024. Efficient tensor decomposition-based filter pruning. Neur Netw, 178:106393.

[24]Qin F, Zain AM, Zhou KQ, et al., 2025. Hybrid weighted fuzzy production rule extraction utilizing modified harmony search and BPNN. Sci Rep, 15(1):11012.

[25]Saraswathi M, Logashanmugam E, 2024. Chicken swarm optimization modelling for cognitive radio networks using deep belief network-enabled spectrum sensing technique. PLoS ONE, 19(8):e0305987.

[26]Sutton PD, Nolan KE, Doyle LE, 2008. Cyclostationary signatures in practical cognitive radio applications. IEEE J Sel Areas Commun, 26(1):13-24.

[27]Urkowitz H, 1967. Energy detection of unknown deterministic signals. Proc IEEE, 55(4):523-531.

[28]Xiang YH, Zhou KQ, Sarkheyli-Hägele A, et al., 2025. Parallel fault diagnosis using hierarchical fuzzy Petri net by reversible and dynamic decomposition mechanism. Front Inform Technol Electron Eng, 26(1):93-108.

[29]Xu LW, Li YF, Gulliver TA, 2025. Federal underwater acoustic spectrum sensing algorithm based on DCYOLO. Expert Syst Appl, 275:126928.

[30]Yan H, Zhong CQ, Wu YH, et al., 2023. A hybrid-model optimization algorithm based on the Gaussian process and particle swarm optimization for mixed-variable CNN hyperparameter automatic search. Front Inform Technol Electron Eng, 24(11):1557-1573.

[31]Ye SQ, Zhou KQ, Zain AM, et al., 2023. A modified harmony search algorithm and its applications in weighted fuzzy production rule extraction. Front Inform Technol Electron Eng, 24(11):1574-1590.

[32]Zeng YH, Liang YC, 2009. Eigenvalue-based spectrum sensing algorithms for cognitive radio. IEEE Trans Commun, 57(6):1784-1793.

[33]Zhang WS, Wang Y, Chen X, et al., 2024. Spectrum transformer: an attention-based wideband spectrum detector. IEEE Trans Wirel Commun, 23(9):12343-12353.

[34]Zhivomirov H, 2018. A method for colored noise generation. Roman J Acoust Vibrat, 15(1):14-19. https://rjav.sra.ro/index.php/rjav/article/view/40