Enhancing Accuracy of Low-Cost Floor Sensor Data for Human Localization Using the Human SLIP Model

Automatic indoor human tracking has gained significant research attention due to the growing demand for enhanced services in smart home environments. In this study, we present a novel method utilizing low-cost floor sensors to estimate an individual’s position in a closed environment. The proposed approach involves calculating the center-of-mass (COM) curve based on the collected footsteps data, resulting in acceptable accuracy with low computational cost for both curved and straight paths. Additionally, an innovative approach based on a human walking model is introduced, effectively reducing floor sensors’ output error by up to 26%, specifically for straight paths. We believe that this method paves the ground for future research endeavors on upscaling low-resolution sensors to higher resolutions and improving floor-sensor-based localization.