Skip to main content Start main content

Dr YANG Lei and his team received the Best Paper Award (Runner-Up) and Graduate Award at ACM MobiCom 2023

3 Nov 2023

Dr YANG’s team was recognised at the award presentation ceremony of ACM MobiCom 2023

Our PhD students received the graduate award at ACM MobiCom 2023

The spatial spectrums, also known as the multipath profile, show how strong the RF signal is from a particular direction composed of the azimuthal and elevation angles. (a) shows the scene, in which the TX may be located at any position, but the RX equipped with a 4 × 4 antenna array is fixed at a corner; (b) shows the point cloud created by LiDAR, which is only used for the conventional ray-tracing algorithm; (c) compares the synthesis spectrums generated by different algorithms when the TX is located at four different positions. The ground truth is obtained using the antenna array.

The physical structure of RFNN. An RFNN network is a physical neural network consisting of a 𝐾-layer metasurface. Each layer has 𝑁 RF elements, which are regarded as neurons. They receive the signals from neurons on the previous layer and retransmit them with altered phase shifts to the next layer.

The research paper titled "NeRF2: Neural Radio-Frequency Radiance Fields" published by Dr Lei YANG’s team was awarded the Best Paper Award (Runner-up) at the 29th Annual International Conference on Mobile Computing and Networking (ACM MobiCom 2023). The co-authors of the paper are Dr Zhenlin AN, postdoctoral fellow and Ms Qingrui PAN and Mr Xiaopeng ZHAO, our PhD students.

The paper proposes a model called Neural Radio-Frequency Radiance Fields (NeRF^2), which is designed to predict the propagation of Radio Frequency (RF) signals in complex environments. The team trained the NeRF^2 model with a few signal measurements to predict the behaviour of a signal at any position when the location of a transmitter is known. The study demonstrated the effectiveness of NeRF^2 in applications such as indoor localisation and 5G MIMO technology. Read the paper:


Additionally, the team also received a Graduate Award at ACM MobiCom 2023 for a demo developed by Dr YANG, Dr AN, Mr ZHAO and two COMP PhD students, Mr Jingyu TONG and Mr Sicong LIAO. This demo, namely “Radio Frequency Neural Networks for Wireless Sensing”, introduces an innovative architecture called RF Neural Network (RFNN) that emulates the functionalities of a conventional neural network in proximity to or within sensors. Leveraging reconfigurable metasurfaces, RFNN enables high-speed calculations on raw RF signals, offering processing capabilities at the speed of light. The initial findings highlight RFNN's potential in various wireless sensing applications, including material classification, localisation, and gesture recognition, while operating with minimal power consumption. This research holds promise for optimising the sensing capabilities of IoT devices.


ACM MobiCom is a renowned global conference focused on mobile computing and wireless networking. The conference, sponsored by ACM SIGMOBILE, showcases cutting-edge research in areas such as mobile cloud computing, 5G networks, and IoT. It is well-regarded for its rigorous review process and attracts attendees from both academia and industry.

Your browser is not the latest version. If you continue to browse our website, Some pages may not function properly.

You are recommended to upgrade to a newer version or switch to a different browser. A list of the web browsers that we support can be found here