Integration of Person Identification and Respiratory Monitoring Based on Reconfigurable Intelligent Surface

This paper proposes a reconfigurable intelligent surface (RIS)-aided system capable of simultaneous human identification and respiration monitoring using ambient WiFi signals. The system associates physiological patterns with individual identities, enabling personalized services in healthcare, security, and humancomputer interaction. The integration of RIS effectively enhances signal quality by mitigating environmental interference and focusing wireless energy on the target. Transformer-based multitask learning (MTL) model is developed to simultaneously extract identity characteristics and estimate respiration rates from received channel state information (CSI). Experimental results demonstrate that the proposed system achieves a person identification accuracy of 93% and a respiration rate estimation error of 0.06 times per minute. These results highlight the potential of RIS-assisted wireless sensing to improve accuracy and robustness of noncontact biometric recognition. The proposed approach enhances the integration of wireless sensing and communication, offering a promising solution for next-generation smart healthcare and security applications.