Gait cycle prediction model based on gait kinematic using machine learning technique for assistive rehabilitation device

Adi Izhar Che Ani, Zakaria Hussain, Mohd Ikmal Fitri Maruzuki, Mohd Suhaimi Sulaiman, Ahmat Adam Abd Rahim


The gait cycle prediction model is a key technique to control assistive rehabilitation devices such ad exoskeleton or orthosis. At the current state, the development of human gait model uses mathematical approaches, while the dynamic characteristics of human physiology give limitation to the current approach. The existing human gait cycle prediction models require extensive kinematic and kinetic data of human body as input parameters, and the measurement of them needs special equipment, thus enable to be used in an assistive rehabilitation devices is challenging. In this study, three different machine learning algorithms named Gaussian Process Regression (GPR), Support Vector Machine (SVM) and Decision Tree (DT) were used to develop human gait models which potential to predict gait cycle. The input parameters for the machine learning algorithm used to develop the model are subject height, weight, hip and knee angle and ground reaction force (GRF), which consider as minimal input parameters. The models developed were further enhanced by introducing different sliding window data from hip and knee angle and ground reaction force for better gait cycle model prediction. DT with sliding window data (t − 3) was ranked as the best gait cycle prediction model due to it providing the lowest Root Means Square Error (RMSE) of 3.3018 and highest R-Squared (R-Value) of 0.97. The prediction model based on hip and knee angle and GRF was a viable approach of the gait cycle for assistive rehabilitation device control.


Assistive Rehabilitation; Devices; Gait Cycle Prediction; Mechine Learning;



  • There are currently no refbacks.

View IJAI Stats

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.