Design Of Wearable Human Activity-Monitoring System Based On Accelerometer

Currently, with the increase of the incidence of diseases like obesity, high blood sugar and high cholesterol, people begin to pay more attention to their own health. One easiest way to keep healthy is to do more exercise. Considering the current situation that the wearable concept is popular and MEMS inertial sensor technology is mature, this paper intends to design a wearable activity-monitoring system based on accelerometer.Design of Hardware and PC software: This paper takes STM32F405 as the control chip, uses MMA8451 Q collecting motion data, takes W25Q64 to save daily activity data and adopts the HM-12 Bluetooth module to communicate with Android smart phones.It uses VS2010 to develop the PC software, intending to show waveform diagram and save the activity data which makes preparation for subsequent algorithm analysis; takes Eclipse to develop Android interactive interface App, for displaying the body activity parameters.Design of gait algorithm: The paper selects thigh as the research object, use the length of acceleration as major feature, and choose the discrete Fourier low-pass filter algorithm. We analyze the characteristics of the human body daily activity, adopt peak-detection method, present seven features to raise the gait recognition algorithm.Experiments on nine kinds of daily activity including walking forward/backward, running,downhill, uphill, riding bicycles, downstairs, upstairs, stepping on the same place are carried out, intending to verify the validity of the algorithm.Design of sleep monitoring algorithm: It selects thigh as the research object. We use turning angle as the main feature, adopt peak-detection method and four other features to present the algorithm. Experiments are carried out to verify the validity of recognition.