An RFID-based Human Spine Behavior Monitor Method

The spine is an important component of the human skeletal system, playing a crucial role in maintaining balance, facilitating movement, and protecting the nervous system. This paper presents a method for monitoring human spinal behavior using ultra-high frequency passive RFID(Radio Frequency Identification) technology. This method aims to predict the occurrence and changes of diseases, providing earlier health interventions. The study summarizes the geometric principles underlying human spine physiological movements and designs corresponding models to capture spine motion patterns. A RFID phase data preprocessing framework is devised to address the phase periodicity, p ambiguity, and multipath interference issues commonly encountered when collecting physical layer data with COTS RFID readers, enabling more accurate restoration of phase data. Additionally, a spine bending angle prediction model based on polar coordinates is developed to mitigate the phase interference caused by tag rotation during motion. Experimental results demonstrate that this method can monitor human spine behavior without distance limitations, in everyday environments, without requiring individuals to wear complex sensor devices. Furthermore, the method exhibits robust monitoring performance and accuracy even in complex scenarios.