Design Of Wi-Fi Networks Based Body Fall Detection System

Fall has become the second major accident injury death cause.In these death related to fall, most people are more than 65-year-old. So the fall has become one of the most harmful factors to the elderly. Therefore, in the face of increasingly prominent global aging problem, developing a portable, accurate judgment, real-time data acquisition system for body fall detecting is of great economic and social value.The paper designed a fall data acquisition and analysis system based on low-cost MEMS inertial sensors and Wi-Fi networks. The system adopts STM32F407 as primary controller, Single chip SOC chip ESP8266(Espressif inc.) as Wi-Fi RF module, low power consumption chip MPU9150(InvenSense inc.) as motion capture inertial sensor. Firstly, we wear the designed sensor node on the body segment, acquire real-time accurate body motion data by MPU9150, including acceleration, velocity and attitude angle, and develop digital fusion filtering algorithm that output real-time accurate body motion data in the process of movement. Then, The real-time accurate body motion data from various parts of the body are sent to PC through ESP8266 Wi-Fi RF node, while PC sends personalize commands to sensor node. Finally, we design the data analysis software platform on PC, which can real-time display, playback, visualization analysis measure, save and read each node motion data waveform.Using the system, in the experiments of simulating human body fall and daily life behavior, after observing and analysing the real-time accurate body motion data wave-form each part from wireless sensor node, we find that the system is reliable, stable and high transport precision on wireless real-time data transmission aspect, and data analysis software platform on PC can export motion data, and invoke Matlab to analysis and design the fall detection algorithm. Using the acceleration threshold algorithm and vertical velocity threshold algorithm in the analysis of fall experiment were tested and obtained good experimental results and can effectively detect the fall behavior. Comparing to the traditional data acquisition system, the fall data acquisition and analysis system can solve problems of cable transmission, and soft real-time data analyzing capability, and false judgment easily because of single body motion data acquisition node.