RF Vision

RF vision uses radio frequency signals to create images and extract information about the environment and human activity, aiming to overcome limitations of traditional optical vision systems. Current research focuses on developing high-resolution 3D imaging techniques using mmWave radar, coupled with advanced signal processing and machine learning algorithms, including diffusion models and correlated knowledge distillation, to improve accuracy and robustness in tasks like pose estimation and object recognition. This technology offers significant potential for applications requiring privacy-preserving sensing, such as human activity monitoring and gait recognition, and for operating in challenging conditions where optical systems fail. The ability to achieve this through low-cost sensors further expands its potential impact.