Judgment motion by pressure clustering of limited nodes (three nodes) based on K-means

Proposed a new algorithm that based on K-means to distinguish the motion states that have large similarity exactly. Identify the motion behavior from the collection of gait, characterization and identification of gait. For the problems that the raw data volume of foot is big and the latitude of data is high and its application is so difficult, in this article we distribute the nodes of feet based on the mechanical characteristics of the human organism, and combined with pressure sensors plantar to acquisition pressure of feet. Due to distinguishing the motion, so the data collected must be in continuous period of large amounts, through the analysis of existing algorithms, offered to an improved algorithm that based on existing K-MEANS algorithm, and identified the state of motion through the combination of pressure and frequency, and can improve the accuracy of the algorithms. To test and verify the algorithm through the pressure values, which can effectively verify the effectiveness of the algorithm and can accurately distinguish the state.