Design and Research of Gait Recognition Method of Upper Knee Prosthesis Based on KNN Algorithm

Abstract

Due to the influence of diseases, accidents and other factors, the number of thigh amputations is increasing year by year, and the demand for intelligent and diversified artificial limbs is also increasing. Aiming at the poor accuracy of gait recognition of prosthesis, this paper takes the upper part of the knee of healthy people as the main research object, collects acceleration signal and gyroscope signal, and performs wavelet packet denoising on them. KNN algorithm was used to construct the classification model for the collected examples to be classified. Three typical gait, stationary, flat walking and stair climbing were selected, and k nearest neighbor samples of unknown samples were studied to predict the category of unknown samples, namely the gait of healthy people. It provides an accurate gait recognition condition for the research of dynamic prosthesis.