Multi-frequency continuous-wave radar dataset for vital signs monitoring across body orientations and subject-radar distances

Contactless monitoring of human respiration using Continuous-Wave (CW) Doppler radar has emerged as a promisingsolution for unobtrusive healthcare sensing. However, the robustness of CW bio-radar measurements is strongly influencedby acquisition geometry and operating frequency, particularly in short-range scenarios. This work presents a multi-subject,multi-frequency CW bio-radar dataset for respiratory monitoring, collected under a structured experimental protocol thatsystematically varies subject–radar distance and body orientation. The dataset comprises synchronized radar and a referenceequipment measurements acquired from 20 human subjects using CW radar at 2.4 GHz, 5.8 GHz, and 24 GHz, with distancesranging from 0.5 m to 2.0 m and multiple controlled orientations. Reference respiratory signals were recorded using a wearablebiosignal acquisition system, enabling direct validation of radar-derived respiratory rates. By providing raw radar signals, anddetailed metadata, this dataset supports reproducible research on body shape positioning performance analysis, spectraldiversity, and robustness evaluation, facilitating the development, benchmarking, and validation of signal processing anddata-driven methods for contactless vital signs monitoring.