摘要 |
The present invention relates to a surface electromyography (sEMG) signal-based gait phase recognizing method adaptively selecting properties and channels using a classifier, which is most accurate, by generating classifiers as much as the number of all cases of a method for extracting muscle and properties of each test subject per a gait phase through a training process and calculating the accuracy of the generated classifiers using separate gait data. The sEMG signal-based gait phase recognizing method includes: (a) a step of receiving sample signals of the EMG signals of the multiple channels and generating the classifiers classifying the EMG signals per the channel; (b) a step of receiving test signals of the EMG signals of the multiple channels, calculating the accuracy of the classifiers by testing the classifiers, and selecting the classifier in accordance with the accuracy; and (c) a step of recognizing the gait phase with the selected classifier. By the method, the surface electromyography signal-based gait phase recognizing method is capable of processing the gait phase recognition faster than an existing method using all properties and all channels and considerably increasing the recognition accuracy by adaptively using the proper property and the proper channel among the multiple properties. |