Research on the Falling Detection System for Elderly-assistant and Walking-assistant Robot

Abstract

With the development of living standards, the problems caused by aging population have become more and more serious. At the same time, with the growth of age, the body function of the elderly is degenerating and the body flexibility is reducing, which leads to frequently occurring of falling in elderly. The research data shows that falls are the most common and unpredictable problem for old people, which are even life-threatening. Therefore, a falling detection system used in the walking-assistant robot is proposed to make the walking-assistant robot better service during the outdoor walking. Firstly, a general design plan of the falling detection for old people, which is used in walking-assistant robot, is introduced through the analysis of falling mechanism of elderly. Then, a acceleration sensor ADXL345 and the detection module of touch pressure sensors are used to collect data of acceleration and hand touch force of users. Next, the feature extraction is carried out on the above data, and the falling detection is delivered by time domain feature representation and recognition algorithm. At last, the test system is built for test verification, and the results show that the falling detection system has a stable and reliable performance and high detection accuracy rate, reaching 99.05%. It can detect the falling of elderly effectively.