Assessment of a Multi-Sensor FBG-Based Wearable System in Sitting Postures Recognition and Respiratory Rate Evaluation of Office Workers

• Published in: IEEE transactions on biomedical engineering Vol. 70; no. 5; pp. 1673 - 1682
• Main Authors: Zaltieri, Martina, Lo Presti, Daniela, Bravi, Marco, Caponero, Michele Arturo, Sterzi, Silvia, Schena, Emiliano, Massaroni, Carlo
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.05.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Bayes Theorem; Biomedical monitoring; Biomedical optical imaging; Bragg gratings; fiber bragg gratings; Fiber gratings; Humans; Modular systems; Monitoring; Musculoskeletal diseases; Musculoskeletal disorders; occupational disease; Parameters; Physical stress; Posture; Protocols; Recognition; Respiration; Respiratory Rate; Sensors; Sitting Position; sitting posture detection; Torso; Wearable Electronic Devices; wearable system; Wearable technology
• ISSN: 0018-9294; 1558-2531; 1558-2531
• DOI: 10.1109/TBME.2022.3225065

Due to prolonged incorrect sitting posture, upper body musculoskeletal disorders (UBMDs) are largely widespread among sedentary workers. Monitoring employees' sitting behaviors could be of great help in minimizing UBMDs' occurrence. In addition, being primarily influenced by psycho-physical stress conditions, respiratory rate (RR) would be a further useful parameter to delineate the workers' state of health. Wearable systems have emerged as a viable option for sitting posture and RR monitoring since enable continuous data collecting with no posture disturbances. Nevertheless, the main limits are poor fit, cumbersomeness, and movement restriction resulting in discomfort for the user. In addition, only few wearable solutions can track both these parameters contextually. To address these problems, in this study a flexible wearable system composed of seven modular sensing elements based on fiber Bragg grating (FBG) technology and designed to be worn on the back has been proposed to recognize the most common sitting postures (i.e., kyphotic, upright and lordotic) and estimate RR. The assessment was performed on ten volunteers showing good performances in postures recognition via Naïve Bayes classificator (accuracy >96.9%) and agreement with the benchmark in RR estimation (MAPE ranging between 0.74% and 3.83%, MODs close to zero, and LOAs between 0.76 bpm and 3.63 bpm). The method was then successfully tested on three additional subjects under different breathing conditions. The wearable system could offer great support for a better understanding of the workers' posture attitudes and contribute to gathering RR information to depict an overall picture of the users' state of health.