Wi-Fi-Based Tracking of Human Walking for Home Health Monitoring

• Published in: IEEE internet of things journal Vol. 9; no. 11; pp. 8935 - 8942
• Main Authors: Lin, Chi-Lun, Chang, Wen-Jia, Tu, Guan-Hua
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.06.2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Channel state information (CSI); Coordinates; Diseases; Errors; Gait; gait analysis; health monitoring; Human motion; Internet of Things; Legged locomotion; Monitoring; Parkinson's disease; Physicians; Signs and symptoms; Straight lines; Telemedicine; Walking; Wireless communication; Wireless fidelity; wireless networks; Wireless sensor networks
• ISSN: 2327-4662; 2327-4662
• DOI: 10.1109/JIOT.2021.3118429

Neurodegenerative disorder diseases, such as Parkinson's disease (PD), are progressive, and their motor symptoms develop slowly. Assessing the motor functions of PD patients relies on the physician's experience and subjective judgment. However, the disease's actual condition may not fully present during the clinical examination due to the motor fluctuation or symptom variation in a day. This article proposed a wireless detection method that can enable long-term monitoring and quantification of walking, one of the challenging metrics to be evaluated in clinics. We utilized the channel state information (CSI) from two Wi-Fi links to construct a 2-D coordinate system in space to locate, track, and capture the gait information of a walking subject, including walking distance, step length, and cadence. The results showed that the estimation errors were 0.02-0.16 m (17.84%-38.48%) for step length, 0.01-0.16 Hz (1%-11%) for cadence, and less than 8.84° for walking direction when a subject walked along a straight line. The method could also detect walking that constantly changes direction (circular path) with localization errors of 0.31 (CW) and 0.50 m (CCW). Moreover, the proposed method could distinguish two subjects when they simultaneously walked along nonpredefined paths. The method can be further developed as a home health monitoring technology to provide long-term and reliable data without personally identifiable information for physicians to make more precise treatments.