Force Myography based Continuous Estimation of Knee Joint Angle using Artificial Neural Network
Recent development in lower limb intelligent prosthetics has led to building an interface for active training in rehabilitation. In this paper, a force myography (FMG) based method has been used to model and estimate the knee joint angle during walking at four different self-selected speeds i.e. slow, normal, fast and run. An eight-channel in-house FMG data acquisition was developed to collect the data wirelessly. A trained artificial neural network (ANN) was fed with FMG data to estimate the knee joint angle. Till now, the experiment was performed for one subject. Root mean square error (RMSE) and the cross-correlation coefficient were used to quantify the accuracy of the proposed method. The results hold promise for FMG based affordable robotic rehabilitation in the future.
Paper
Full text
Force Myography based Continuous Estimation of Knee Joint Angle using Artificial Neural Network
Semantic Scholar · Engineering · 2019
Abstract
Recent development in lower limb intelligent prosthetics has led to building an interface for active training in rehabilitation. In this paper, a force myography (FMG) based method has been used to model and estimate the knee joint angle during walking at four different self-selected speeds i.e. slow, normal, fast and run. An eight-channel in-house FMG data acquisition was developed to collect the data wirelessly. A trained artificial neural network (ANN) was fed with FMG data to estimate the knee joint angle. Till now, the experiment was performed for one subject. Root mean square error (RMSE) and the cross-correlation coefficient were used to quantify the accuracy of the proposed method. The results hold promise for FMG based affordable robotic rehabilitation in the future.