Action recognition method based on GST and VL-MOBPNN

Abstract

The invention relates to an action recognition method based on GST and VL-MOBPNN. The method comprises the steps: acquring sample data of human lower limb electromyographic signals; performing time-frequency generalized S transformation on an electric signal to obtain a time domain cumulative characteristic curve when the time resolution is relatively high and a frequency domain cumulative characteristic curve when the frequency resolution is relatively high; obtaining a feature vector of the signal; and inputting the feature vector into a learning rate variable quantity back propagation neural network for identification and classification to obtain a classification result. According to the invention, time-frequency generalized S transformation is adopted to analyze signals more meticulously, so that the problem that the window width of a Gaussian window is fixed in a traditional time-frequency analysis method is solved; Gaussian window adjustment parameters are introduced into time-frequency generalized S transformation, the inverse proportion change rate of the Gaussian window width along with the frequency can be flexibly adjusted according to the frequency distribution characteristics and the time-frequency analysis emphasis of the electromyographic signals in practical application, the method can better adapt to analysis and processing of specific signals, and therefore the signal processing capacity is improved.

Description

technical field [0001] The invention relates to the technical field of pattern recognition, in particular to an action recognition method based on GST and VL-MOBPNN. Background technique [0002] Surface electromyography (sEMG) is a non-stationary one-dimensional time-series bioelectric signal of neuromuscular series activities guided, amplified and recorded from the muscle surface through electrodes, which can reflect neuromuscular activities. Different body movements are produced by different muscle contraction patterns, and there are also differences in the characteristics of EMG signals. Different action patterns can be distinguished by analyzing the characteristics of the EMG signal. EMG signals have been widely used in many fields such as clinical medicine, kinematics, biomedicine and engineering, and have become ideal control signals for functional electrical stimulation in intelligent prosthetics. [0003] The key to action pattern recognition based on EMG is how t...

AI Technical Summary

Problems solved by technology

However, the window width of the short-time Fourier transform is fixed, and the window width cannot be changed according to different signal components.

Images