The smartphone as a gait recognition device impact of selected parameters on gait recognition

This paper aims to identify the impacts of a couple of parameters on gait recognition when a build-in smartphone accelerometer is used. The use of different types of shoes impacts significantly gait recognition while the matching rate on a different surface e.g. grass has only a minor impact. A correlation between accelerometer's data and the phone position was identified. For this, data originating from the Z-axis as well as from the X-Y-Z-axis was used together with Dynamic Time Warping (DTW) for template generation and matching tests.