Modeling Human Walking for Step Detection and Stride Determination by 3-Axis Accelerometer Readings in Pedometer

Pedometers are increasingly being used for purposes like assessment of physical activity, rehabilitation, disease management and as a utility for the visually impaired. These devices rely on algorithms that consist of measuring distance covered by the number of steps taken. Conventional estimations of distance covered rely on empirically obtained relations to arrive at their data. In this paper, we present a means of estimating the distance covered by the wearer through a more realistic model using sensor readings from a 3-axis accelerometer present in most pedometers as well as in modern phones and tablets. First comes step detection where each individual step taken by the wearer is identified. Subsequently, we determine stride length, based on the inverted pendulum model and the periodicity of accelerations generated by humans. Lastly, we obtain the total distance covered during the walk. Finally, we analyze the results with respect to currently existing models.