Yang Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Yang Li has a total of 17 co-authors. Here are the top co-authors by number of shared publications:
- Caijun Zhong
College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou, China
1 shared publication - Congyu Li
Beijing University of Posts and Telecommunications
1 shared publication - Guangxu Zhu
Shenzhen Research Institute of Big Data
1 shared publication - Jiamo Jiang
China Academy of Information and Communications Technology, Beijing, China
1 shared publication Jianzhong Charlie Zhang
Samsung Research America
1 shared publication- Jinlong Zhang
Beijing University of Posts and Telecommunications
1 shared publication
Latest Publications
- 13 November 2025
- Published by Springer Nature
- p. 395-407
High-Accuracy Action Recognition in Practical WiFi Systems: Platform Implementation, Dataset Construction, and Algorithm Design
- 13 November 2025
- Published by Springer Nature
- p. 395-407
High-Accuracy Action Recognition in Practical WiFi Systems: Platform Implementation, Dataset Construction, and Algorithm Design
- Open Access
- 10 February 2023
- Published by MDPI AG in Applied Sciences
- Vol. 13 (4) , 2319
A Joint Channel Estimation and Compression Method Based on GAN in 6G Communication Systems
Due to the increasing popularity of communication devices and vehicles, the channel environment becomes more and more complex, which makes conventional channel estimation methods further increase the pilot overhead to maintain estimation performance. However, it declines the throughput of communication networks. In this paper, we provide a novel two-stage based channel estimation method by using generative adversarial networks (GANs) to handle this problem in orthogonal frequency division multiplexing (OFDM) systems. Specifically, the first stage aims to learn the mapping from a low-dimensional latent variable to the real channel sample. During the second stage, an iterative algorithm method is designed to find the optimal latent variable by matching the pilot channels of a real channel and generated channel. Then, the data channels are recovered based on the learned mapping relationship between the latent variable and the real channel sample. The simulation results show that our proposed method can achieve a performance gain of more than 2 dB with a pilot reduction by when SNR is 10 dB, by comparing with the widely used Wiener filter interpolation method. In addition, as the low-dimensional latent variable can be obtained simultaneously, it can also be used for reducing the feedback overhead.- 10 February 2023
- Published by MDPI AG in Applied Sciences
- Vol. 13 (4) , 2319
Open AccessA Joint Channel Estimation and Compression Method Based on GAN in 6G Communication Systems
Due to the increasing popularity of communication devices and vehicles, the channel environment becomes more and more complex, which makes conventional channel estimation methods further increase the pilot overhead to maintain estimation performance. However, it declines the throughput of communication networks. In this paper, we provide a novel two-stage based channel estimation method by using generative adversarial networks (GANs) to handle this problem in orthogonal frequency division multiplexing (OFDM) systems. Specifically, the first stage aims to learn the mapping from a low-dimensional latent variable to the real channel sample. During the second stage, an iterative algorithm method is designed to find the optimal latent variable by matching the pilot channels of a real channel and generated channel. Then, the data channels are recovered based on the learned mapping relationship between the latent variable and the real channel sample. The simulation results show that our proposed method can achieve a performance gain of more than 2 dB with a pilot reduction by when SNR is 10 dB, by comparing with the widely used Wiener filter interpolation method. In addition, as the low-dimensional latent variable can be obtained simultaneously, it can also be used for reducing the feedback overhead. - Open Access
Wireless Image Transmission with Semantic and Security Awareness
Semantic communication is an increasingly popular framework for wireless image transmission due to its high communication efficiency. With the aid of the joint-source-and-channel (JSC) encoder implemented by neural network, semantic communication directly maps original images into symbol sequences containing semantic information. Compared with the traditional separate source and channel coding design used in bitlevel communication systems, semantic communication systems are known to be more efficient and accurate especially in the low signal-to-the-noise ratio (SNR) regime. This thus prompts an critical while yet to be tackled issue of security in semantic communication: it makes the eavesdropper more easier to crack the semantic information as it can be decoded even in a quite noisy channel. In this letter, we develop a semantic communication framework that accounts for both semantic meaning decoding efficiency and its risk of privacy leakage. To achieve this, targeting wireless image transmission, we on the one hand propose an JSC autoencoder featuring residual for efficient semantic meaning extraction and transmission, and on the other hand, propose a data-driven scheme that balances the efficiency-privacy tradeoff. Extensive experimental results are provided to show the effectiveness and robustness of the proposed scheme.
Open AccessWireless Image Transmission with Semantic and Security Awareness
Semantic communication is an increasingly popular framework for wireless image transmission due to its high communication efficiency. With the aid of the joint-source-and-channel (JSC) encoder implemented by neural network, semantic communication directly maps original images into symbol sequences containing semantic information. Compared with the traditional separate source and channel coding design used in bitlevel communication systems, semantic communication systems are known to be more efficient and accurate especially in the low signal-to-the-noise ratio (SNR) regime. This thus prompts an critical while yet to be tackled issue of security in semantic communication: it makes the eavesdropper more easier to crack the semantic information as it can be decoded even in a quite noisy channel. In this letter, we develop a semantic communication framework that accounts for both semantic meaning decoding efficiency and its risk of privacy leakage. To achieve this, targeting wireless image transmission, we on the one hand propose an JSC autoencoder featuring residual for efficient semantic meaning extraction and transmission, and on the other hand, propose a data-driven scheme that balances the efficiency-privacy tradeoff. Extensive experimental results are provided to show the effectiveness and robustness of the proposed scheme.
- 10 April 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 465-470
Model-driven Machine Learning Approach for Mobility Classification in Intelligent 5G Network
Channel information is essential to unleash the benefits of 5G New Radio (NR) by enabling network intelligence that adapts transmissions to users’ channels. In this paper, we propose model-driven feature design and use support vector machine (SVM) to classify users’ speed range. Our model-driven features are designed based on stochastic channel modeling. Multiple features are derived from time-domain cross-correlation and time-domain auto-correlation function of the sounding reference signals. The classifier is trained and verified with extensive standard compliant simulation channels at different SNR levels and speeds, and attains greater than 90% accuracy.
- 10 April 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 465-470
Model-driven Machine Learning Approach for Mobility Classification in Intelligent 5G Network
Channel information is essential to unleash the benefits of 5G New Radio (NR) by enabling network intelligence that adapts transmissions to users’ channels. In this paper, we propose model-driven feature design and use support vector machine (SVM) to classify users’ speed range. Our model-driven features are designed based on stochastic channel modeling. Multiple features are derived from time-domain cross-correlation and time-domain auto-correlation function of the sounding reference signals. The classifier is trained and verified with extensive standard compliant simulation channels at different SNR levels and speeds, and attains greater than 90% accuracy.