Application of a nonparametric gait model and feature generation based on insole data

Abstract

Gait and balance disorders pose significant risks of injury and negatively impact quality of life, especially among the elderly and individuals with neurological conditions. In this work, a novel feature generation method based on a nonparametric gait model (NPWM) for plantar pressure data is proposed for early fall risk assessment. Unlike traditional gait analysis methods that rely on fixed parameters or predefined models, the proposed approach directly extracts spatiotemporal features from real world data collected by wearable plantar pressure sensors. These features include raw spatial distributions, gait temporal characteristics, balance stability metrics, energy expenditure features, and informative feature ratios. Experimental results demonstrate that the proposed method achieves a fall risk prediction accuracy of 0,9, offering significant advantages over conventional clinical assessment methods.