A wearable gait analysis system based on "Lab-in-Shoe" intelligent footwear
10.3760/cma.j.cn115530-20231231-00287
- VernacularTitle:基于"Lab-in-Shoe"智能鞋的可穿戴步态分析系统研究
- Author:
Ji HUANG
1
;
Xu WANG
;
Xin MA
;
Wenming CHEN
Author Information
1. 复旦大学工程与应用技术研究院，上海 200433
- Keywords:
Gait;
Biomechanics;
Ankle injuries;
Wrarable technique
- From:
Chinese Journal of Orthopaedic Trauma
2024;26(8):705-710
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To develop a wearable gait analysis system based on "Lab-in-Shoe" intelligent footwear for quantitative assessment of gait dysfunction in patients with ankle injuries.Methods:In this study, integration of inertial sensors and insole-type plantar pressure distribution sensors into footwear formed the hardware core of the "Lab-in-Shoe" intelligent footwear system. In terms of algorithms, acceleration data from the inertial sensors were integrated to obtain spatial parameters of gait. The insole-type plantar pressure sensors were employed to acquire the data concerning foot pressure distribution, as well as temporal parameters and mechanical parameters of gait, including support phase, swing phase, and zero velocity moments. To validate the accuracy of this system, 8 young and healthy participants [age: (25.6±1.3) years; height: (175.4±2.2) cm] were recruited for gait data collection in an optical motion capture laboratory. By comparing the gait data between the "Lab-in-Shoe" intelligent footwear system and the gold standard Vicon optical motion capture system, the effectiveness and reliability of the intelligent footwear system were respectively tested. Additionally, a calibration experiment was conducted for the sensing units of the plantar pressure sensors to examinate the accuracy of the pressure data.Results:The tested system accurately captured the following gait parameters of the participants: step length, step width, step frequency, walking speed, gait phase division, foot pressure distribution, and center of pressure curve, among other core spatiotemporal gait parameters. Moreover, the system demonstrated its ability to replicate the dual-foot posture during gait. Compared with the gold standard Vicon optical motion capture system through Bland-Altman, the Lab-in-Shoe smart shoe system showed stride length mean error within 3.12% (range: 2.76% to 4.24%) across 3 different walking speeds [slow speed (0.68±0.05) m/s, preferred speed (1.10±0.07) m/s, and fast speed (1.40±0.13) m/s]. 90% of the results fell within the 95% limits of agreement, indicating good consistency. The intraclass correlation coefficients (ICC) for stride parameters within the slow, preferred, and fast walking speed groups were 0.93, 0.917, and 0.893, respectively, indicating good reliability. The calibration data of multiple sensor units from the plantar pressure sensors all fell within the 95% linear regression range, with a correlation coefficient of r=0.949 ( P<0.05). The plantar pressure data collected by the intelligent footwear system presented a distinct bimodal characteristic. Conclusions:The "Lab-in-Shoe" smart shoe system developed by our institute is capable of collecting and calculating gait parameters conveniently and quickly, and demondtrates good reliability and validity across different walking speeds. Therefore, it is valueable for large-scale gait data collection in a clinical setting.
