Copy the page URI to the clipboard
Islam, Riasat; Bennasar, Mohamed; Holland, Simon; Mulholland, Paul and Price, Blaine
(2024).
Abstract
This study evaluated the versatility of the MoJoXlab in conducting clinical movement analysis using inertial sensors from various manufacturers, including low-cost, non-proprietary, and open-protocol wearable options. Data were collected from 15 healthy participants who performed a range of clinically relevant activities and exercises using two sets of sensors. Dynamic time warping analysis of the sensor signals suggested that the collected dataset could be used for further algorithm development. The findings demonstrate that the current iteration of MoJoXlab can perform movement analysis using quaternions from sensors of any manufacturer. However, the accuracy of the resulting joint angles is not yet suitable for clinical applications across all sensor types, and only Xsens and NGIMU sensors are currently supported. This study also explored the potential for reducing the number of sensors required by MoJoXlab, which currently uses seven sensors to calculate joint angles for three joints (hip, knee and ankle) on both sides of the body. The creation of a comprehensive databank for lower limb movement analysis algorithms was an additional outcome of this work. Further research and development are necessary to expand MoJoXlab's support for multiple sensor manufacturers and improve the accuracy of joint angle calculations for clinical applications.