Design And Implementation Of Fall Detection System Based On Triaxial Acceleration

With the continuous development of aging of population, the health and life quality of old people need to get more attention. However, falling down has becoming one of the most significant reason for old people’s illness and death. If the olds’falling down can be found in time, many tragedies can be avoidedBased on the three-axis acceleration sensor, the main research content of this article is,1) Designed a new fall detection algorithm.2) Designed and implemented a fall detection system prototype..Fall detection algorithm’s data source of this paper is three-axis acceleration sensor in the waist, based on the Zigbee wireless communication technology, working in the unlicensed 2.4 GHz frequency, and sending data to the terminal PC by the radio frequency way. All activities can be divided into falling down, slight movement and resting. In order to better distinguish falling down from the other two kinds of behavior, this article suggests a new way to solve the problem:Regarding the mix of acceleration peak value, three-axis signal changing range and the energy used by movement as character, comparing the character vector which is composed of three character values with falling down practice sample, judging whether the sample to be tested is in the falling down vector area or not.Fall detection system prototype of this paper is designed based on the c# language. Mainly realized the detections of user management, data acquisition, data processing, data analysis, motion status report and fall send alarm and so on.According to the result of the experiment, the accuracy rate of the method of detection proposed in the report could run up to 93 percent. It is a kind of fall detection algorithms, which more accurate and effective, whatever orientation sensor worn, recognition rate keeps steady state. On this basis, design and implement the fall detection alarm system prototype. If falling down’s happening, alerted signal can be sent to family members or medical organization through mobile network for quick relief.Research results and innovation points of this article has the following several aspects:1) New fall detection algorithm of pattern recognition, which can effectively avoid the false alarm caused by body position change.2) Using Zigbee network technology for data collection and transmission, greatly improved the fall monitoring distance.3) Verified for the proposed fall detection algorithm by the experiments.