Comprehensive validation of a wearable foot sensor system for estimating spatiotemporal gait parameters by simultaneous three-dimensional optical motion analysis

Abstract

Background Use of a wearable gait analysis system (WGAS) is becoming common when conducting gait analysis studies due to its versatility. At the same time, its versatility raises a concern about its accuracy, because its calculations rely on assumptions embedded in its algorithms. The purpose of the present study was to validate twenty spatiotemporal gait parameters calculated by the WGAS by comparison with simultaneous measurements taken with an optical motion capture system (OMCS). Methods Ten young healthy volunteers wore two inertial sensors of the commercially available WGAS, Physilog®, on their feet and 23 markers for the OMCS on the lower part of the body. The participants performed at least three sets of 10-m walk tests at their self-paced speed in the laboratory equipped with 12 high-speed digital cameras with embedded force plates. To measure repeatability, all participants returned for a second day of testing within two weeks. Results Twenty gait parameters calculated by the WGAS had a significant correlation with the ones determined by the OMCS. Bland and Altman analysis showed that the between-device agreement for twenty gait parameters was within clinically acceptable limits. The validity of the gait parameters generated by the WGAS was found to be excellent except for two parameters, swing width and maximal heel clearance. The repeatability of the WGAS was excellent when measured between sessions. Conclusion The present study showed that spatiotemporal gait parameters estimated by the WGAS were reasonably accurate and repeatable in healthy young adults, providing a scientific basis for applying this system to clinical studies.

Predicting Gait Parameters of Leg Movement with sEMG and Accelerometer Using CatBoost Machine Learning

“…The average shoe size among participants was 23.9 ± 0.6, varying from 23 to 25 cm. The four electrodes were placed on the gastrocnemius and vastus lateralis muscles of the left and right legs for sEMG measurements, the ACC was placed on the outer thigh for recoding the thigh motion, and the GaitUp Physilog ® [28][29][30] shoe-worn inertial sensors were positioned on the tongues of shoes for gait status measurement. Participants evaluated their physical health before participating in the experiment.…”

Gait characteristics of individuals with incomplete cervical spinal cord injury by the wearable gait analysis system

“…Eight representative spatiotemporal gait variables were used for analysis; walking speed (m/s), cadence (steps/min), stride length (% height), stance phase time (% gait cycle), double support time (% gait cycle), loading phase time (% stance), foot-at phase time (% stance), and pushing phase time (% stance). The de nitions of each variable were provided by the instruction manual [14] (Table 1). To avoid acceleration or deceleration effects, the initial and last two steps of each walking were excluded from the analysis.…”

“…Due to the technological advancement, wearable sensors on feet can clarify spatiotemporal gait characteristics easily with great validity and repeatability comparable to optical motion analysis system [14] and are currently used for many clinical studies [11,15]. Thus, the purpose of the present study was to reveal (1) the difference in gait performance between individuals with ICSCI and control participants, and (2) the correlation between the gait variables and walking speed or the balance function in the ISCI group in a clinical setting using wearable sensors.…”

Pathological Gait Analysis With an Open-Source Cloud-Enabled Platform Empowered by Semi-Supervised Learning-PathoOpenGait

“…However, their high costs and extensive space requirements limit their applicability in clinical and domestic settings [17]. Alternatively, systems based on wearable devices might be relatively more affordable [12], [18]- [20]. However, the additional equipment an individual must wear can be disruptive, and the setup is often inconvenient [21].…”

“…Moreover, the location where wearable devices are attached can influence estimation performance [22]. While devices attached to the feet might provide more effective signals for gait analysis than those attached to the knees, the former can be more uncomfortable for users than the latter [20].…”