基于android的跌倒检测算法研究及实现-控制科学与工程专业论文.docxVIP

AbstractWith Abstract With the frequent increase of population aging and the”empty nest”phenomenon， the elderly health problems have become increasingly prominent．Among the many factors that threaten the physical and mental health of the elderly,fall down is the primary risk factor．How to effectively detect the failing down in the elderly，SO as to reduce the health damage caused by falling down has become a new research focus in the world． Tllis paper mainly study the fall detection algorithm based on body posture data LPCC feature，through the DTW(dynamic time neat)method to match the human body real-time LPCC feature of optimal mode，to accurately detect the fall state，and use Matlab to simulate and verify the algorithm．And the paper further realize this fall detection algorithm on the Android smartphone，and timely send messages or call the police automatically when detecting user fall down，and then make the fall was found and treated in time．The main work is as follows： Firstly,using STM32 microcontroller and mpu6050 MEMS sensor chip built a body posture data acquisition system，to collect the user’S various daily motion and the original experimental data of the falling posture． Secondly,to process the coordinate correction and complementary filtering of the original experimental data and time frequency analysis，extract falling time·frequency domain features，finally determine the short-time LPCC feature of posture data Call be beRer describe and distinguish between the whole process feature of the body posture changes． Again，using the body posture data of short-term LPCC feature as feature parameters，present the matching template training algorithm based on DTW algorithm and the fall detection algorithm based on pattern matching，realize and verify the algorithm in the Matlab platform by collecting a large number of dam samples，and the experimental results show that the validity and reliability of the algorithm． Finally,transplant the above algorithm into the Android