The Real Time Motion Pattern Recognition of Lower Limb Based on sEMG signals

In this work, we proposed a Back-propagation neural network (BPNN) method based on surface electromyo-graphy(sEMG) signals for recognition of movement in human locomotion. The method only needs six sEMG signals to help detect hemiplegia patient walking. We use experiment first verify the availability of the sEMG data acquisition system. Then quantify the system with a gold standard optoelectronic system Vicon-nexus (VN). The Hip and Knee Joint Angles has a good correlation coefficient and mean squared error of volunteers locomotion. In contrast to the existing methods, the proposed model not only avoids the complex sensor system but also ensures the accuracy of recognition. Meanwhile, the aim of this paper is to build the regression model which relates the multichannel sEMG signals to human lower limb joint angles in order to develop a more natural study for the patients.