Research Interests

Wireless Perception

I am interested in enabling wireless signals, such as WiFi CSI and mmWave radar, to perceive humans and scenes beyond the limits of cameras. My current focus is WiFi-CSI-based 3D human pose estimation, where wireless signals provide privacy-preserving perception in camera-denied environments such as darkness, occlusion, and through-wall scenarios. In the long term, I aim to extend these ideas toward mmWave / RF-based human and scene understanding, including pose, depth, occupancy, and semantic perception.

RT-Based Synthetic Data Generation

Large-scale real wireless sensing data is difficult to collect, annotate, and reproduce across environments. I am interested in building physics-grounded synthetic RF data pipelines using ray tracing and simulation, with the goal of reducing dependence on costly real-world data collection. I am especially interested in sim-to-real calibration, where synthetic wireless signals are adjusted to better match real-world hardware, materials, antenna configurations, and multipath propagation.

Cross-Domain Generalization

A central challenge in wireless sensing is domain shift: models trained in one room, device setup, antenna placement, or environment often fail in another. I aim to study RF-native self-supervised and representation learning methods that generalize across environments, subjects, devices, and sensing modalities. My long-term goal is to develop scalable and generalizable wireless perception models that can transfer from WiFi CSI to mmWave radar and other RF sensing systems.