Rui Wang
About
Dr. Rui Wang is an Associate Professor and Ph.D. Supervisor at the Department of Electronic and Electrical Engineering, SUSTech. With research experience in both academia and industry, he has published over 120 papers in IEEE journals and conferences, focusing on communication and computing system optimization. He actively participates in wireless standard development and holds 80+ patents and applications. His current research focuses on prototype-verified ISAC technologies and Wi-Fi standardization.
Research Projects
DoDTrack: Indoor Mobile Devices Tracking via Difference-of-Doppler
Existing indoor tracking schemes often require extensive fingerprint training or strict synchronization/wide bandwidth (ToF/AoA). We propose DoDTrack, which uses only Difference-of-Doppler (DoD) measurements from distributed antennas sharing one oscillator, canceling CFO/SFO and removing the need for transmitter cooperation or a known starting position. It achieves a median tracking error of 0.34 m in a 6 m × 6 m indoor area.
DopMap: Doppler-Driven Indoor Mapping via mmWave Communication Signals
DopMap reconstructs indoor wall layouts using millimeter-wave communication signals. Conventional wireless mapping often requires wideband ranging or tight synchronization between transmitters and receivers. DopMap instead combines Doppler signatures from a moving scatterer with signal direction measurements to identify an initial wall and map the surrounding walls. In experiments, measurements from four transmitter positions achieved a mean distance error of 0.29 m for the additional walls, without wideband ranging or transmitter–receiver clock synchronization.
Dual-Attention and Adversarial Transfer Networks for Sim-to-Real Cross-Orientation Wireless Sensing
Millimeter-wave activity recognition degrades when the user's orientation changes, yet labeling measured data for every orientation is costly. S2M-Sense synthesizes orientation-diverse training data from single-orientation motion capture with a physics-guided channel simulator, extracts orientation-robust features from dual-link Doppler spectrograms with a dual-attention network, and bridges the simulation-to-reality gap through adversarial unsupervised transfer learning. On 60.48 GHz measured data, accuracy reaches 88.33% with simulation-only training and 95% after transfer with 16 unlabeled samples.