1. Design of Force Sensitive Resistor (FSR) embedded insole for phase detection during human gait and its classification
- Author
-
Pant, U, Baral, S, Gupta, A, and Shrestha, P L
- Abstract
Gait analysis is the study of human motion during walking or running. It is often used to evaluate and treat individuals with movement disorders or to assist in the design of prosthetics and other assistive devices. The phase detection aspect of gait allows for the quantification of gait characteristics, assessment of effectiveness of interventions and monitoring the changes in gait over time. The objective of the paper is the design and development of a Force Sensing Resistor (FSR) embedded insole that can be placed in shoes for detection of phases during the human gait and its classification. The classification using FSR is achieved by placing sensors at strategic locations on the foot, to detect the forces applied during walking. The data collected by these sensors are used to identify specific phases of the gait cycle—particularly heel strike, midstance, toe-off and swing, by analysing the patterns of data during movement. The layout and number of FSRs is determined through literature review and is followed by the insole fabrication. The wireless transmission of gait data through the FSRs is actuated for ten healthy adults. Machine Learning (ML) program is consequently developed based on all available classification algorithms of supervised learning namely—Logistic Regression, Naïve Bayes, Decision Tree, Random Forest and Support Vector Machine (SVM). The algorithm is trained and tested on separate FSR data sets from which the predictive decision in each gait cycle is attained. In the phase classification aspect, SVM, Naïve Bayes, and Random Forest achieved the highest accuracy of 91.54%, followed by Decision Tree at 90.77% and Logistic Regression at 89.23%. In conclusion, the optimal number of FSRs and their location in smart insole for gait phase detection is resolved. Gait phase classification is also attained. FSR embedded insole thus, provides a promising solution in accurately identifying and categorizing gait phases.
- Published
- 2024
- Full Text
- View/download PDF