The Academic Achievement of the National Key Laboratory of Mobile Communication of Southeast University won the Best Paper Award of "China Communications"

Publisher:沈如达Release time:2018-11-29Times of browsing:791

At the IEEE/CIC (International Conference on Communications in China) held in Beijing on August 17, 2018, the paper “Deep Learning for Wireless Physical Layer: Opportunities and Challenges, vol. 14, no. 11, pp. 92-111, Nov. 2017 (Tianqi Wang, Chao-Kai Wen, Hanqing Wang, Feifei Gao, Tao Jiang, Shi Jin)” has won the Best Paper Award of China Communications in 2018. The first author, Tianqi Wang (Mentor: Shi Jin), is a master student of the National Mobile Communication Research Laboratory in Southeast University. This paper was awarded by the editor-in-chief, Prof. Jianhua Lu of Tsinghua University, and is one of the two best papers selected from more than 700 papers published from 2015 to 2017.

This paper reviews the application and development of AI technology in the physical layer of wireless communication in recent years. With the rapid development of information technology, wireless communication is facing new requirements such as multi-scenarios, high reliability and low latency, which constantly challenges the traditional algorithms and system architecture of the wireless physical layer. On the other hand, the breakthrough of deep learning in computer vision, natural language processing and other fields, leads to the research of applying artificial intelligence technologies to the physical layer of wireless communication. By virtue of the characteristics of data-driven modeling, deep learning based models are promising to assist conventional physical layer algorithms and solve the difficulty of channel modeling. This paper extensively and thoroughly investigates the latest academic research achievements of applying deep learning to wireless physical layer at home and abroad, and summarizes the common schemes and performance advantages of deep learning for modulation recognition, channel decoding, signal detection and so no.  The details of a new end-to-end wireless communication framework are also presented. This paper reveals the feasibility and great potential of applying learning for wireless communication, and also points out current problems and challenges, such as lack of reliable theoretical analysis, specialized architectures for wireless communications, and specific implementation schemes, and thus indicates the future research and development direction in this field.

China Communications (ISSN1673-5447) is an SCI monthly English magazine for the worldwide engineers and scholars of industries, universities, and research institutes in the field of information & communication technology (ICT). It is hosted by China Institute of Communications (CIC) and IEEE Communication Society (IEEE ComSoc), and provides online access to the latest papers, reports advanced research achievements, publishes review papers, and introduces major technology applications.