Research on FSO modulation classification algorithm based on deep learning

Abstract

For FSO communication atmospheric turbulence has a large impact on signal modulation, the convolution-profile stellar data image conversion algorithm proposed in this paper performs data conversion on the received constellation maps, so that they retain more original signal feature images. A classification network based on the channel attention mechanism is proposed to classify the modulated signals by extracting the feature information in the image through the residual structure, and the attention mechanism assigns different weights of the channel features. Under the same data conversion algorithm, the proposed classification network achieves the highest recognition accuracy of 96.286%.

References

[1] Sun Y L, Liu J. Automatic recognition of signal modulation based on deep learning[J]. Journal of Qilu University of Technology, 2020, 34(01): 58-63 (in Chinese)

[2] Zeng C Z, Jia X, Zhu W G. Modulation classification of communication signals[J]. Communications technology, 2015, 48(03): 252-257 (in Chinese)

[3] Lu Y X, Zhang H Y, Liang Y, et al. Indoor visible light channel model of convolutional neural network based on GAN fingerprint database[J]. Journal of optoelectronics·laser, 2023, 34(11): 1201-1209 (in Chinese)

[4] Chen D, Wang C H, Yao B Y. Wireless optical subcarrier modulation identification based on fuzzy clustering and BP neural network[J]. Laser & optoelectronic progress, 2018, 55(06): 61-66 (in Chinese)

[5] Yang H J, Xu J L, Liu Y L. Modulation recognition of optical communication system based on pattern recognition[J]. Laser journal, 2019, 40(03): 109-112 (in Chinese)

[6] Li T Y, Li M, Zhou M O, et al. Signal detection in ACO-OFDM free-space optical communication based on deep learning[J]. Journal of optoelectronics·laser, 2023, 34(01): 68-73 (in Chinese)

[7] Li Y, Dong Q, Duan L Y. Research on state pattern recognition of an optical communication system based on cloud computing[J]. Laser journal, 2021, 42(05): 88-91 (in Chinese)

[8] Wang Y, Lv T T, Yang H. Classification of underwater communication modulated signals based on deep residual network[J]. Digital ocean & underwater warfare, 2020, 3(03): 242-249 (in Chinese)

[9] Zhu L, Han H M, Peng H. Research on QPSK intelligent receiver model based on deep learning[J]. Computer measurement and control, 2024, 32(02): 213-218 (in Chinese)

[10] Gao L, Ni H. Research on fault diagnosis of optical communication system based on improved convolutional neural network[J]. Laser magazine, 2023, 44(08): 161-165 (in Chinese)

[11] Yun L, Tu Y, Zheng D, et al. The application of deep learning in communication signal modulation recognition[C]. 2017 IEEE/CIC International Conference on Communications in China (ICCC), October 22–24, 2017, Qingdao, China, 2017, New York: IEEE 8330488

[12] Zhang W T, Cui D. Training images generation for CNN based automatic modulation classification[J]. IEEE access, 2021, 9: 62916-62925

[13] Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks[C]. International Conference on Neural Information Processing Systems, December 3–6, 2012, Lake Tahoe, Nevada, USA, 2012, Vancouver: NIPS 1097-1105

[14] Szegedy C, Liu W, Jia Y, et al. Going deeper with convolutions[C]. Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition, June 8–10, 2015, Boston, MA, USA, 2015, New York: IEEE 1-9

[15] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[EB/OL]. (2015-04-10) [2024-01-29]. https://arxiv.org/pdf/1409.1556.pdf

[16] He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]. Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition, June 26-July 1, 2016, Las Vegas, NV, USA, 2016, New York: IEEE 770-778