"WiFi-Rehab Dataset"

"This dataset provides WiFi Channel State Information (CSI) data for device-free upper-limb rehabilitation assessment. The data were collected using commodity Intel 5300 network cards during standardized arm-lift rehabilitation exercises. Five range-of-motion grades are included, corresponding to standing still and arm-lift angles of approximately 45\u00b0, 90\u00b0, 135\u00b0, and 180\u00b0. Each class contains CSI samples collected from multiple volunteers under indoor settings, making the dataset suitable for WiFi-based human activity recognition, postoperative rehabilitation monitoring, and range-of-motion grading tasks.The dataset can support research on CSI signal preprocessing, fine-grained motion recognition, subject-independent evaluation, and lightweight deep learning models for home-based rehabilitation systems. It is intended to facilitate reproducible research on privacy-preserving, low-cost, and unobtrusive rehabilitation assessment using ambient wireless signals."