Human body lower limb motion movement classified identification method based on single-channel myoelectricity signals

Abstract

The invention belongs to the technical field of non-stationary nonlinear surface myoelectricity signal processing and lower limb motion movement classification, and in particular relates to a human body lower limb motion movement classified identification method based on single-channel myoelectricity signals. According to the method, myoelectricity signals obtained by a single channel obtained at a certain single lower limb muscle group which is related to daily lower limb motion movements are adopted, myoelectricity signal data corresponding to four different tested typical lower limb motion movement states including leg extending, leg bending, walking supporting and walking swinging are acquired, then wavelet transform is carried out on the signals, so that wavelet coefficients are obtained, and singular value decomposition is carried out on each layer of wavelet coefficient, wherein the decomposition result serves as a characteristic matrix; finally, a quadruple classification problem is transformed into a plurality of duplex classification problems by adopting a tree-shaped support vector machine duplex classifier, so that a classification identification result is obtained. The verification result under the clinical test data shows that the feature extraction and classification method provided by the invention has a very good effect for single-channel myoelectricity signals of the four daily lower limb motion movements.

Description

technical field [0001] The invention belongs to the technical field of non-stationary nonlinear surface electromyography signal processing and lower limb movement action classification, and in particular relates to a single-channel electromyography signal-based classification and identification method for human lower limb movement actions. Background technique [0002] Human limb movement is an important part of human daily life activities. For special populations (such as hemiplegic patients or elderly people, etc.), daily life activities (especially lower limb movements such as walking, standing, squatting or sitting, etc.) often require external assistance to be successfully completed. A large number of scholars are engaged in research on power-assisted robot systems to meet the daily life assistance needs of the above-mentioned population. Since the surface electromyography signal (EMG) can directly represent the electrophysiological response of daily life activities, m...

AI Technical Summary

Problems solved by technology

The identification method of myoelectric signals proposed in this patent application is not closely integrated with practical applications, nor can it be directly used for the classification and identification of human lower limb movement movements with a single channel of myoelectric signals as the signal source

Images