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 curve based on the collected footsteps data, resulting in high accuracy 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 this method paves the ground for future research endeavors on up-scaling low-resolution sensors to higher resolutions and improving floor-sensor-based localization.