Wearable sensor makes possible today to obtain motion measurement for health monitoring, physical activity assessment, or sport performance evaluation in ecological environment. One of them, the inertial measurement unit (IMU), is extensively used to obtain measurements relative to human motion. The IMU, composed of 3D gyroscope, accelerometer, and generally magnetometer, can be used to provide an orientation in a global coordinate system after fusion of the different sensor data. The principal issue to obtain the human movement measurement is the sensor-to-segment (S2S) calibration, which provides the orientation of the IMU relatively to the segment on which it is fixed (Picerno 2017). This calibration aims to define the localization of the segment axes in the IMUs technical coordinate system. Different approaches have been made in the literature to perform this calibration. Static calibration defines segment axes during specific static postures by making assumptions regarding the alignment of the axes relatively to the vertical (Palermo et al. 2014). Functional calibration hypothesises that the subject can perfectly rotate around segment axes during specific movements (Favre et al. 2008). Anatomical calibration uses a supplementary device to locate the segment axes in the IMU coordinate system (Picerno et al. 2008). Today, there is no consensus on which calibration method is the most appropriate to perform S2S calibration. The aim of this study is then to compare different S2S calibration methods for lower limb segments. In the present study, the segment axes are defined during different movements and postures deduced from markers tracked with optoelectronic system, not by IMU measurement, in order to distinguish errors due to the measurement system from errors due to the calibration method by itself. The axes, thus, obtained are compared with axes obtained with a method of reference proposed for optoelectronic measurement. This method of reference is based on recommendations formulated by the ISB (Wu et al. 2002) and on the recommendations the most recently formulated in the literature (Kainz et al. 2017). The difference in segment axes orientation will then be exclusively due to the axe definition method.
© Copyright 2020 Computer Methods in Biomechanics and Biomedical Engineering. Taylor & Francis. Published by Taylor & Francis. All rights reserved.