GaitKeeper: Revolutionizing Gerontechnology with Artificial Intelligence and Augmented Reality for Enhanced Gait Analysis

Abstract

Purpose: GaitKeeper is an innovative mobile technology that uses artificial intelligence (AI) and augmented reality (AR) to enhance gait analysis in aging populations. GaitKeeper standardises assessments of gait speed, ensuring that each patient is uniformly evaluated. Standardisation boosts the reliability of data collected whilst also providing a consistent metric for monitoring patients over time. Gait speed is often referred to as the "sixth vital sign" due to its predictive value for functional capacity, mobility, cognitive decline, and overall mortality. (Fritz & Lusardi, 2009) By improving the precision and accessibility of gait speed measurements, GaitKeeper is set to revolutionise health assessments for older adults.

Method: GaitKeeper employs AR to create a standardized virtual gait laboratory that ensures uniform assessment conditions. Concurrently, GaitKeeper’s AI components analyse video recordings to capture 25 distinct joint positions. GaitKeeper simulates a 4-meter walkway, as recommended by the global guidelines for falls prevention and management, ensuring consistent assessment conditions across various clinical settings. (Montero-Odasso et al., 2022) This standardised testing environment improves the measurement of variables such as the acceleration phase, which previously lacked consistent measurement in traditional methods of measuring gait speed. GaitKeeper's efficacy and technological rigor were validated against Vicon and GaitRite through comparative analyses in both laboratory and clinical environments.

Results and Discussion: A technical evaluation at Dublin City University (n=30) measured GaitKeeper’s ability to capture accurate positional data from video, using the Vicon system as the verification gold standard. Additionally, a clinical evaluation at Tallaght University Hospital involving people with mild cognitive impairment (MCI) (n=30) compared gait speed data with specialist physiotherapist clinical observational studies and GaitRite systems during 4m walking under single and dual task conditions. The technical evaluation provides evidence of GaitKeeper’s ability to capture accurate gait data from video (n=30, p<0.001, error<3%). The clinical evaluation supports the accuracy in comparison with the gold standard (n=30, p<0.001, error<6%). These results demonstrate that GaitKeeper’s outputs are highly consistent with established gait analysis systems, showing less than 2% variance from Vicon and robust correlation coefficients with GaitRite (Pearson r = 0.72, Spearman ρ = 0.918). The integration of AI and AR not only facilitates precise gait measurements but also enhances the scalability and usability of gait analysis, supporting proactive health management in older adults.