Page 66 - 2024F

UEC Int’l Mini-Conference No.53 59

Estimation of lower limb muscle activity during walking in elderly

people using markerless motion capture technology

1
Theodore REANGPUSRI and Hidetaka OKADA 2
1 Blekinge Institute of Technology (BTH), Karlskrona, Sweden
2 Department of Mechanical Engineering and Intelligent Systems, , The University of
Electro-Communications, Tokyo, Japan

Keywords: Markerless Motion Capture, OpenPose, AnyBody Modeling, Gait Analysis, Elderly Mo-
bility, Gender Differences, Age-Related Gait Changes.

Abstract

Mobility decline is a major concern in elderly individuals, leading to an increased risk of falls and
reduced quality of life. Traditional motion capture systems require physical markers and specialized
equipment, making them expensive and less accessible for clinical use. This study explores the feasibility
of markerless motion capture using OpenPose for gait analysis, combined with AnyBody musculoskeletal
modeling to estimate muscle forces and joint loads.
Thirteen elderly individuals (6 females, 7 males) aged 77 to 92 years were recorded walking at
a self-selected pace. OpenPose was used to detect key gait events such as heel strikes and toe-offs,
and gait cycles were normalized to 101 interpolated data points per subject. AnyBody Modeling
System estimated biomechanical parameters, including muscle activation and joint forces, to analyze
gait characteristics.
Results indicate that male participants exhibited slightly higher mean gait metrics compared to
females, with moderate variability between individuals. Age-related analysis showed no strong correla-
tion between age and gait metrics, suggesting that gait characteristics remain relatively stable in this
elderly age range. OpenPose effectively tracked gait cycles without requiring physical markers, while
AnyBody provided meaningful estimations of muscle activation and joint forces, supporting its potential
application in rehabilitation and clinical assessments.
This study demonstrates the feasibility of using markerless motion capture for gait analysis in elderly
individuals. While minor gender differences were observed, the effect of aging on gait characteristics was
minimal. The integration of OpenPose and AnyBody Modeling provides a non-invasive, cost-effective
alternative for clinical gait analysis, fall prevention strategies, and AI-driven rehabilitation monitoring.