Gait Analysis And Classification Of Restricted Knee Based On Human Feet Electrostatic

Gait analysis is a kind of biometric detection and analysis technique according to human walking posture. Due to its features of intuitive and real-time, it causes widespread concern in the field of rehabilitation medicine and clinical diagnosis.Non-contact body electrostatic detection technology detects human electrostatic field,and it can detect human movement within the effective range. When people walk,changes in the electrostatic field of the human body contain characteristics of human walking, this may provide a new source of information for gait feature extraction and human gait analysis.In this paper, we develop an electrostatic sensor and use it to obtain human gait electrostatic signal when people walk. And we use frequency domain information of gait electrostatic signal to extract gait features in three restricted circumstances of knee, and classify the three kinds of gaits. Compared to traditional gait analysis, this method has the advantages of non-contact, ease of use, no need to wear any device and no subject to environmental factors such as light interference. In the research of gait frequency domain information, a gait electrostatic signal acquisition system is set up. And we design the gait simulation experiment of knee joint disease patients and obtain the gait electrostatic signal with the knee joint restricted in normal, limited and semi confinement condition, with which a small gait electrostatic signal database was established. The frequency domain information of three kinds of signals is obtained by the fast Fourier transform. We analyze the characteristics of the three kinds of signals in frequency domain and prove the rationality of using the frequency domain information for classification. By using principal component analysis(PCA), the dimension of frequency domain information is reduced and characteristic parameters are got. We select k nearest neighbor algorithm for classification and recognition and its highest recognition rate is 84.75%. Results show that the frequency domain information of the human gait electrostatic signal can reflect the activity of the knee joint. This may provide a theoretical basis for the treatment of diseases of the knee rehabilitation.This article will be helpful to establish a new technical means to assess the degree of rehabilitation of knee joint disease and will play a role in the exploration of the human lower limb bone joint rehabilitation research.