Pilot Study on Frailty Detection Using Non‐wearable Gait Sensing

To develop a system for early frailty detection, extracting frailty‐related parameters through gait sensing using non‐wearable sensors was investigated. Gait parameters, such as step length, were obtained using passive infrared sensors and footstep sounds were recorded using a floor‐mounted microphone. Fifty‐seven subjects were classified into robust, pre‐frail, and frail groups based on the Japan Cardiovascular Health Study (J‐CHS) criteria, and the parameters were compared across groups to assess their association with frailty. Frailty detection was attempted using the k ‐nearest‐neighbor algorithm with the extracted parameters, resulting in an F1‐score of up to 80.4%. These results suggest the feasibility of frailty detection using non‐wearable gait sensing.