Fabric movement artifact in smart clothing with direct integration of IMU technology: preliminary reliability analysis of upper limb and trunk kinematics

This thesis work explores the reliability of a motion capture system based on IMUs (Inertial Measurement Units) integrated into smart clothing in a preliminary analysis. The main objective is to characterize errors in the reconstruction of joint angles resulting from movements of the fabric on the trunk relative to the skin, during movements of the upper limbs and trunk. First, using a reflecting marker grid and stereophotogrammetry, the trunk fabric deformations during some fundamental movements were investigated. The results revealed that the largest deformations occur mainly in the regions under the armpits and along the back, affecting the accuracy of joint angle measurements. To evaluate errors in joint angle reconstructions in the same investigation, metrics such as RMSE and multiple correlation coefficients CMC were used, demonstrating that the IMU system presents differences compared to the gold standard represented by the optoelectronic system (RMSE from 1.5° to 19.5°). Secondly, errors in joint angles due to fabric motion were also examined, extracting soft tissue artifacts and differences between global reference systems from the total error, highlighting that fabric motion is the main source of error (RMSE from 0° to 16.8°). This research provides a basis for optimizing the design of smart clothing intended for biomechanical analysis and clinical use.