Analysis of a shuffling detection device for fall prevention

Abstract

According to the World Health Organization, falls are the second leading cause of accidental or unintentional injury deaths worldwide. In order to address this issue from a fall prevention perspective, I began developing a footwear device to give feedback to the user on their walking. In this iteration, I've identified characteristics to distinguish between shuffling and walking strides for implementation in a threshold-based algorithm and the device created was effective in making this distinction with a 74.65% accuracy and 73.17% sensitivity. The errors identified are those common to threshold-based algorithms and can be corrected in future iterations with a more sensitive threshold and stride detection method. Moving forward, the device can be improved in order to be integrated into a feedback device for the intended user.