Location optimisation for IMUs in the design of motion tracking soccer tights based on Soft Tissue Artefact analysis

Abstract

INTRODUCTION: It has been known that effects of Soft Tissue Artefact (STA) cause inaccuracy of motion tracking when using skin-attached markers for kinematic studies (1) (2). However no study has been found to compare the displacement of different marker locations for the implementation of IMUs in soccer clothing. In this study, the displacements of seven marker locations were compared to segments of a Plugin Gait system (3) during acceleration and deceleration trials. The seven marker locations were determined by a preliminary study on comfort and risk of different marker locations during soccer-specific movements, resulting in a ranking of the marker locations.
METHODS: Four able bodied subjects (2M 2F) performed five repetitions of both acceleration (ACC; 0% - 100%) and deceleration movements (DEC; 100% - 0%). Optoelectronic measurements were done in an indoor space surrounded by eight optoelectronic VICON cameras (100 FPS). Sixteen markers placed on bony landmarks were used to model joint centre coordinates of the hip, knee and ankle to simulate bone segments of the thigh and shank to serve as a reference frame. C3D data of markers was processed in Matlab® to calculate displacement both perpendicular and parallel to the reference frame. Parallel displacement is measured relatively and measured from the knee joint centre due to non-constant segment length caused by the modelled hip joint centre.
RESULTS: The total number of data points after eliminating outliers is 963 (488 perp. and 475 para.). Each data point represents the maximum shift for each marker per trial for both legs as the range of distances between the marker and the reference frame. The markers with significant lowest mean displacement according to an ANOVA parametric test (with Post Hoc Scheffe, p = 0.05) are the ones located on the lateral side of the shank and thigh (both 2/3 distal). Significant higher displacement occurs on the Semimembranosus, Vastus Lateralis and Gastrocnemius. Markers located on the lower hamstring (7 cm proximal to knee joint centre, posterior) and Popliteus are neither in the significantly lowest or highest subset of markers.
CONCLUSION: Optoelectronic markers on the lateral side of the thigh and shank cause lower amount of perpendicular and parallel errors in kinematic measurements compared to five other marker locations. As the lateral shank was also recommended in the preliminary study, it would be a good marker location to both minimize comfort issues and measurement errors caused by STA. Lateral thigh locations, however, were perceived to cause discomfort. The designer should consider using the marker location on the Popliteus to balance between lowest displacement and comfort. It should be noted that displacement highly depends on type of movement and its vector, of which parallel displacement values could be affected by systematic errors of the Plugin Gait model.