Eventos Anais de eventos
COBEM 2019
25th International Congress of Mechanical Engineering
INFLUENCE OF BIOMECHANICAL PARAMETERS ON THE GAIT IDENTIFICATION OF EXOSKELETON WITH UNSUPERVISED CLUSTERING ALGORITHM
Submission Author:
Carlos Eduardo Oliveira , SP
Co-Authors:
Carlos Eduardo Oliveira, Júlio César Moraes Fernandes, Marcos Silveira
Presenter: Carlos Eduardo Oliveira
doi://10.26678/ABCM.COBEM2019.COB2019-0772
Abstract
The objective of this work is to study the kinematic behavior of an exoskeleton-like structure of lower limbs in order to identify gait patterns. Specifically, the influence of the length ratio between the tibia and the femur is shown, and the k-means unsupervised clustering algorithm is used to identify gait patterns using quantitative biomechanical metrics. The k-means algorithm was used with 18 metrics, variation of R and combinations of patterns of foot and pelvis. The results show that the k-means unsupervised algorithm was not able to correctly identify all four patterns when all 18 metrics were used. Further analysis showed that the use of fewer selected metrics can give very good identification. For example, when using only metrics S_1 and A_C it was possible to obtain 100% correct identification. In general, metrics A_C, S_1, x_C, x_{Cmax} and c_{min} were the more successful in prediction of gait patterns.
Keywords
exoskeleton gait characterisation, biomechanical parameters, k-means clustering
