Stride Length Estimation with Motion Data from the Contralateral Foot

Stride length estimation plays an important role in gait analysis, rehabilitation, and mobility assessment, offering insights into walking efficiency, balance, and musculoskeletal health. Gait depends on the coordinated movement of the stride foot and the contralateral foot. However, most studies have focused solely on estimating stride length using data from the stride foot, leaving the contribution of the contralateral foot largely unexplored. This study addresses this gap by investigating how capturing kinematic data from the contralateral foot can enhance stride length estimation using motion data from a smart insole system consisting of an IMU embedded in the mid-region of the insole and an LSTM model. A leave-one-subject-out evaluation was conducted with data from 10 healthy subjects to assess stride estimation under three conditions: utilizing IMU data from both feet, from the stride foot only, and from the contralateral foot only. The best-performing test case, with an average mean percentage absolute error of 7.67%, was achieved when using data from both feet. Notably, kinematic data from the contralateral foot alone yielded a competitive error of 7.75%, further highlighting its relevance. These findings demonstrate the importance of integrating motion data from contralateral feet to enhance stride length estimation while ensuring robust and reliable gait analysis. This study provides key insights for future research in wearable sensing, gait analysis, and AI-driven movement modeling, contributing to advancements in lower-limb mobility assessment and rehabilitation applications.