Generative AI for Wireless Systems
Generative models grounded in the structure and physics of radio signals.
I am currently an Assistant Professor in the School of Artificial Intelligence at Beijing University of Posts and Telecommunications (BUPT). Before joining BUPT, I was a postdoctoral researcher at the Institute of Trustworthy Networks and Systems, Tsinghua University, led by Prof. Yunhao Liu. I received my Ph.D. from Tsinghua University in 2024 under the supervision of Prof. Zheng Yang, and my B.E. from BUPT in 2019.
How can generative AI understand the physical world through radio signals?
Wireless Sensing ISAC Embodied Navigation
Generative models grounded in the structure and physics of radio signals.
Reliable perception from signals already present in the environment.
Connecting perception, maps, and action for agents in the physical world.
A time-frequency diffusion framework for generating diverse, high-quality time-series RF signals.
Domain-adaptive Wi-Fi fall detection using environment-independent features and visual supervision.
Standalone Wi-Fi 6-DoF pose tracking designed for reliable indoor drone flight control.
A public-camera-based system that localizes users, attaches semantic context, and provides indoor navigation.
A multimodal diffusion framework that combines commodity Wi-Fi CSI and RGB images for robust depth estimation.
An invited presentation on generative methods for radio signals and their applications to wireless sensing, signal denoising, and channel estimation.