Lower limb knee joint continuous motion estimation method based on electromyographic signals

Abstract

The invention relates to a lower limb knee joint continuous motion estimation method based on electromyographic signals. Firstly, human lower limb knee joints are collected at low speed; the femur biceps femoris in a medium-speed and fast movement mode; Quadruple femur muscle, Femoral lateral muscle, Intra-femoral muscle, the electromyographic signals and the real-time angles of the hemiplegia muscles and the thigh muscles are acquired; The method comprises the following steps: firstly, extracting a wavelet coefficient, a root mean square, a sorting entropy and other features of a signal as input, then combining the three features into a new feature as input, carrying out normalization processing on feature data, and finally, carrying out prediction by using a least square support vector machine regression model through comparison of different methods. Experimental results show that different characteristics have different relative prediction performances in different motion modes, prediction results combined with the three characteristics are obviously superior to those obtained by adopting any one of the three characteristics independently, prediction is very accurate through a least square support vector machine model, and finally an ideal prediction model is obtained.

Description

technical field [0001] The invention belongs to the field of pattern recognition, and relates to a method for pattern recognition of electromyographic signals, in particular to a method for estimating continuous motion of knee joints of lower limbs based on electromyographic signals. Background technique [0002] Human-computer interaction technology is a popular cutting-edge science and technology nowadays, and surface electromyography (sEMG) is the input signal source of human-computer interaction. The surface electromyography signal is a non-stationary weak signal, which is superimposed on the skin surface by a group of action potential sequences jointly generated by the muscle and its related motor units when it is excited. The acquisition is simple and non-invasive, and has become a hot research field in human-computer interaction technology. [0003] The research on surface EMG mainly focuses on the two processes of feature extraction and pattern recognition. The cor...

AI Technical Summary

Problems solved by technology

At present, there are not many studies on continuous motion estimation of joints using EMG signals at home and abroad, and there is a large research space

Images