1.北京邮电大学人工智能学院,北京 100876
2.泛网无线通信教育部重点实验室,北京 100876
3.鹏城实验室, 广东 深圳 518055
牛凯(1976年生),男;研究方向:信息论与极化码、智能信号处理;E-mail:niukai@bupt.edu.cn
纸质出版日期:2025-01-15,
网络出版日期:2024-10-08,
收稿日期:2024-06-18,
录用日期:2024-07-14
移动端阅览
牛凯,鲁延鹏,董超.面向工业互联网的语义编码传输方法及应用[J].中山大学学报(自然科学版)(中英文),2025,64(01):51-60.
NIU Kai,LU Yanpeng,DONG Chao.Semantic coding transmission method and application for industrial internet[J].Acta Scientiarum Naturalium Universitatis Sunyatseni,2025,64(01):51-60.
牛凯,鲁延鹏,董超.面向工业互联网的语义编码传输方法及应用[J].中山大学学报(自然科学版)(中英文),2025,64(01):51-60. DOI: 10.13471/j.cnki.acta.snus.ZR20240204.
NIU Kai,LU Yanpeng,DONG Chao.Semantic coding transmission method and application for industrial internet[J].Acta Scientiarum Naturalium Universitatis Sunyatseni,2025,64(01):51-60. DOI: 10.13471/j.cnki.acta.snus.ZR20240204.
通过将语义通信技术引入工业网络,构建了一个面向工业互联网的语义编码传输系统。系统中设计了语义编解码器以提取信源中的语义信息,相对于传统通信系统,基于语义信息的通信有更高的信息压缩效率与更高的符号差错容忍能力。同时引入信源信道联合编解码器,以信源信道联合编码的方式将语义信息转化为信道符号传输,进一步提升系统对工业网络通信资源的利用效率。所有编解码器均构建在深度神经网络架构Transformer上,确保了编解码器对语义信息的理解能力及系统的泛化能力。在工业药品生产场景中,对该系统进行测试,结果显示:相较于传统通信方案,该语义编码传输系统在图像重建质量和传输处理速度方面均有显著提升。且系统对下游任务的性能影响极小,保证了工业生产中如缺陷检测等关键任务的准确性。
The integration of semantic communication technology into industrial networks is proposed and a semantic coding transmission system for the industrial internet is established. Within this system,a semantic codec is developed to extract semantic information from the source. Compared with traditional communication systems, semantic-based communication offers higher information compression efficiency and greater symbol error tolerance. Furthermore, a co-codec for joint source-channel coding have been introduced, which enhances the utilization efficiency of industrial network communication resources through co-coding techniques. All codecs are built upon leading deep neural network architecture “Transformer”, ensuring their ability to comprehend semantic information and maintain network universality. The system has been tested in real-world scenarios within industrial drug production facilities, demonstrating significant improvements in image reconstruction quality and transmission processing speed compared to traditional communication schemes. Additionally, minimal impact on downstream task performance ensures accuracy in critical tasks such as defect detection in industrial production processes.
工业互联网数据压缩与传输语义通信深度学习
industrial internetdata compression and transmissionsemantic communicationsdeep learning
牛凯,戴金晟,张平,等,2021.面向6G的语义通信[J].移动通信,45(4):85-90.
美国通用电气公司,2012. 工业互联网:突破智慧和机器的界限[R].美国通用电气公司.
张平,牛凯,姚圣时,等,2023.面向未来的语义通信:基本原理与实现方法[J].通信学报,44(5):1-14.
BALLé J, CHOU P A, MINNEN D, et al, 2020. Nonlinear transform coding[J]. IEEE J Top Signal Process, 15(2): 339-353.
BALLé J, MINNEN D C, SINGH S, et al, 2018. Variational image compression with a scale hyperprior[EB/OL]. [2018-03-01].https://doi.org/10.48550/arXiv.1802.01436https://doi.org/10.48550/arXiv.1802.01436.
BOURTSOULATZE E, KURKA D B, GüNDüZ D, 2019. Deep joint source-channel coding for wireless image transmission[J]. IEEE Trans Cogn Commun Netw, 5(3): 567-579.
DAI J, WANG S, TAN K, et al, 2022. Nonlinear transform source-channel coding for semantic communications[J]. IEEE J Sel Areas Commun, 40(8): 2300-2316.
DENG J, DONG W, SOCHER R, et al, 2009. Imagenet: A large-scale hierarchical image database[C]//IEEE Conference on Computer Vision and Pattern Recognition. Miami, FL,USA:IEEE Computer Society: 248-255.
GüNDüZ D, ERKIP E, 2008. Joint source-channel codes for MIMO block-fading channels[J]. IEEE Trans Inf Theory, 54(1): 116-134.
GüNDüZ D, QIN Z, AGUERRI I E, et al, 2022. Beyond transmitting bits: Context, semantics, and task-oriented communications[J]. IEEE J Sel Areas Commun,41(1):5-41.
LI J Q, YU F R, DENG G, et al, 2017. Industrial internet: A survey on the enabling technologies, applications, and challenges[J].IEEE Commun Surv Tutor, 19(3): 1504-1526.
LIU Z, LIN Y, CAO Y, et al,2021. Swin transformer: Hierarchical vision transformer using shifted windows[C]// IEEE/CVF International Conference on Computer Vision. Montreal, QC, Canada:IEEE Computer Society: 10012-10022.
MINNEN D, BALLé J, TODERICI G,2018. Joint autoregressive and hierarchical priors for learned image compression[EB/OL].Advances in Neural Information Processing Systems. [2018-09-08].https://doi.org/10.48550/arXiv.1809.02736https://doi.org/10.48550/arXiv.1809.02736.
NIU K,DAI J,YAO S, et al,2022. A paradigm shift toward semantic communications[J]. IEEE Commun Mag,60(11): 113-119.
P’NG C,GREEN J,CHONG L C, et al,2019. BPG: Seamless, automated and interactive visualization of scientific data[J]. BMC Bioinformatics, 20(1): 42.
QIN Z, TAO X, LU J,et al,2021. Semantic communications: Principles and challenges[EB/OL]. [2022-06-27].https://doi.org/10.48550/arXiv.2201.01389https://doi.org/10.48550/arXiv.2201.01389.
RYAN W E,2004. An introduction to LDPC codes[R].Coding and Signal Processing for Magnetic Systems 5.2 .CRC Press: 1-23.
SHANNON C E,WEAVER W, 1971. The mathematical theory of communication[M]. The University of Illinois Press.
VASWANI A, SHAZEER N, PARMAR N, et al, 2017. Attention is all you need[EB/OL]. Advances in Neural Information Processing Systems. [2023-08-02].https://doi.org/10.48550/arXiv.1706.03762https://doi.org/10.48550/arXiv.1706.03762.
WIEGAND T, SULLIVAN G J,BJONTEGAARD G,et al,2003. Overview of the H. 264/AVC video coding standard[J].IEEE Trans Circuits Syst Video Technol,13(7): 560-576.
YAO S, LU Y,NIU K,et al,2024. Semantic information processing for interoperability in the industrial internet of things[J]. Fundam Res, 4(1): 8-12.
0
浏览量
71
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构