Motion analysis is often used to understand the knee joint biomechanics. This is mostly achieved using optoelectronic systems that precisely measure the displacement of lower-body reflective markers. However, their high cost and the necessity of a dedicated environment lead to more and more researchers using inertial sensors. Inertial sensors are indeed more affordable than optoelectronic systems and have increased their accuracy and stability over the years. Despite this infatuation, there still are no studies that show a precise 3D knee angle calculation during motion. This lack of accuracy is caused by magnetic disturbances that occur when inertial sensors composed of magnetometers are used. However, in the absence of magnetometers, a drift in the gyroscopes’ integrated signals also causes accuracy loss. This project’s goal is to assess the feasibility of inertial sensors including magnetometers in knee movement analysis. In order to achieve this, the movement detected by the inertial sensors (APDM) and the reference (VICON) is first compared using the rotation angle around the helical axis. Afterwards, the inertial and optical signals are calibrated and the knee angles are calculated during three squat trials. Results showed low repeatability for the chosen calibration method, mostly for the tibia. Moreover, differences greater than 50o for the knee angles were calculated between the inertial and optical systems for two trials out of three. Besides, subsequent analysis helped us understand that APDM inertial sensors provided erroneous quaternions for those two trials. However, further studies are required to understand what could be the cause of these quaternion calculation errors.
Blouin, J. (Author),
Hagemeister (Supervisor) &
Aissaoui (Co-supervisor),
25 May 2016Student thesis: Master's thesis › Master in Engineering: Engineering