Research And Implementation Of Fall Detection Technology Based On Android Platform

With the development of human society, social population structure has been changing from “youth” to “eldly”. An aging population has brought new problems, and the life quality and health problems of the old people have caused extensive concern of society. Especially fall of the elderly has evolved into a more notable problem. In this situation, various fall detection technology and systems have arisen to solve the problem that the old people fall. Fall detection system can not only benefit the elderly, but also can reduce the pressure of society and the children, and that has important economic and social significance.At present, fall detection systems can be divided into two classes: context-aware fall detection system and wearable fall detection system. Existing systems, however, due to equipment limitations or imperfect research theory and so on, have some shortcomings. Context-aware fall detection system cannot ensure users’ privacy, have a limited monitoring scope and high cost of equipment installation. Some wearable fall detection system equipment are inconvenient, some cannot detect in real time, and some have unreasonable algorithm. The key of fall detection is the performance of the algorithm, therefore this paper mainly researches on fall detection algorithms, and realize the fall detection system based on Android platform.This paper specifically researches on fall detection algorithm and automatic system for help, namely, through the accelerometer and gyroscope sensor collecting user’s movement information, Data preprocessing, fall detecting and automatically calling for help if a fall happens. The main work is as follows:(1) After collecting movement data, remove the noise through the data processing.(2) Using two-stage joint fall detection algorithm, can ensure the sensitivity of detection, improve the efficiency of detection, and reduce the complexity of the algorithm.(3) Calling for help is divided into two stages, preparation stage and Action stage, to further ensure the stability of the system.(4) Choosing android, which are widely used, as the realization platform for the system, can effectively reduce the cost and development complexity.(5) Different age volunteers participate in the experiment, and simulate fall and daily actions for data acquisition and system verification.The main innovation of this paper is the new fall detection algorithm. It presents a two-stage joint fall detection algorithm which combining threshold method and support vector machine(SVM) classification method. The first-level detection based on threshold value method ensures the detection efficiency of the algorithm, and the secondary detection based on support vector machine(SVM) classification method greatly meets the needs of accuracy in fall detection, so that the progressive combination of them ensures the real-time and accuracy of the whole algorithm. A large number of actual experimental prove that the fall detection system, which is proposed in this paper, has a higher accuracy and a reliable algorithm, and it has a certain reference of further research on the fall detection system.