A prediction method of speed-dependent walking patterns for healthy individuals

• Published in: Gait & posture Vol. 68; pp. 280 - 284
• Main Authors: Fukuchi, Claudiane A., Duarte, Marcos
• Format: Journal Article
• Language: English
• Published: England Elsevier B.V 01-02-2019
• Subjects: Gait analysis; Kinematics; Kinetics; Prediction methods; Regression analysis; Walking; Walking; Regression analysis; Kinetics; Gait analysis; Kinematics; Prediction methods
• ISSN: 0966-6362; 1879-2219
• DOI: 10.1016/j.gaitpost.2018.12.006

•A new method is presented to predict joint angles and moments at any gait speed.•Predicted data at a given gait speed showed good agreement with experimental data.•The method produces unbiased reference data to compare subjects at different speeds.
Gait speed is one of the main biomechanical determinants of human movement patterns. However, in clinical gait analysis, the effect of gait speed is generally not considered, and people with disabilities are usually compared with able-bodied individuals even though disabled people tend to walk slower.
This study proposes a simple way to predict the gait pattern of healthy individuals at a specific speed.
The method consists of creating a reference database for a range of gait speeds, and the gait-pattern prediction is implemented as follows: 1) the gait cycle is discretized from 0 to 100% for each variable, 2) a first or second-order polynomial is used to adjust the values of the reference dataset versus the corresponding gait speeds for each instant of the gait cycle to obtain the parameters of the regression, and 3) these regression parameters are then used to predict the new values of the gait pattern at any specific speed. Twenty-four healthy adults walked on the treadmill at eight different gait speeds, where the gait pattern was obtained by a 3D motion capture system and an instrumented treadmill.
Overall, the predicted data presented good agreement with the experimental data for the joint angles and joint moments.
These results demonstrated that the proposed prediction method can be used to generate more unbiased reference data for clinical gait analysis and might be suitably applied to other speed-dependent human movement patterns.