Related Experiment Video
Updated: Jul 5, 2026
08:56
Clinical Assessment of Spatiotemporal Gait Parameters in Patients and Older Adults
Published on: November 7, 2014
Detection of tripping gait patterns in the elderly using autoregressive features and support vector machines
Daniel T H Lai1, Rezaul K Begg, Simon Taylor
1Department of Electrical and Electronic Engineering, The University of Melbourne, Parkville Campus, Melbourne, Victoria 3010, Australia. d.lai@ee.unimelb.edu.au
Journal of Biomechanics
|April 25, 2008
Summary
This study introduces an intelligent gait detection system to identify elderly individuals at risk of tripping falls. The system accurately detects at-risk gait patterns using minimum toe clearance, enabling early preventive measures.
Area of Science:
- Gerontology
- Biomedical Engineering
- Gait Analysis
Background:
- Elderly individuals are at high risk for falls, leading to significant medical costs and mortality.
- Tripping falls in the elderly are often caused by underlying gait abnormalities.
- Current methods for fall risk assessment may not be sufficiently sensitive to subtle gait changes.
Purpose of the Study:
- To develop and validate an intelligent gait detection system (AR-SVM) for early screening of elderly individuals prone to tripping falls.
- To identify specific gait characteristics, namely minimum toe clearance (MTC), indicative of fall risk.
- To enable timely preventive interventions by detecting at-risk gait patterns.
Main Methods:
- The proposed system integrates an autoregressive (AR) model with a support vector machine (SVM) classifier.
- Input data consists of digital signals derived from consecutive minimum toe clearance (MTC) measurements during steady-state walking.
- The system was evaluated on 23 elderly individuals (13 healthy, 10 with a history of falls) using treadmill walking data.
Main Results:
- A fourth-order AR model accurately distinguished fallers from non-fallers using at least 64 MTC values.
- Integrating AR model coefficients into an SVM classifier improved detection accuracy significantly, requiring less than 1 minute of walking data.
- The AR-SVM system achieved 95.65% detection accuracy with 16 MTC samples, demonstrating high efficiency.
Conclusions:
- The developed AR-SVM system offers a fast and efficient method for detecting tripping gait characteristics in the elderly.
- The system requires minimal data (a small number of strides and MTC measurements) for accurate risk assessment.
- This technology has the potential to facilitate early detection and preventive strategies, reducing fall-related injuries and costs.
