| As Chinese aging society is coming, the percentage of elder people is also increasing, and the living of the elder people has become an important social problem.According to statistics, there are more than 1/3 of the elder people have had the experience of falling, especially in older age groups. The falling down of human body is a direct reason of injury for the elder people. And if the injured elder people were ignored for a long time, it may cause more serious result, even death. In the United States, the number of people who died of falling is more than ten thousand each year.In this thesis, the design and development of a human fall detector based on pattern recognition was proposed. When the elder people carried our human fall detector, the body falling will start the alarm immediately. Obviously, the proposed detector in this thesis can promote the safety of life for the erder people.In this thesis, kinds of the human fall detection schemes were studied, and a kind of pattern recognition algorithm based on support vector machine was proposed. First of all, the MMA8451 three axis accelerometer observe and detect the dynamic acceleration of the human body, and the acceleration characteristic values are collected to identify the fall behavior. If any falling is detected, the alarm will start immediately. And the alarm includes voice alarm and wireless alarm. The voice alarm can reduce the false alarm and omission by using the cancel option active help option.For wireless alarm, the GPS module and GPRS module are used, the positioning information will be sent to the binding mobile phone through the GPRS wireless network.A large number of experimental data show that: the accuracy of the proposed system can reach 96.9%. And with the further optimization of the detection algorithm and the increasing of the statistic data, the effectiveness of the proposed system will promote too. In brief, the human fall detector in the thesis can achieve good performance. |
PDF Full Text Request