Research Of Fall Detection Technology Based On 3-Axis Accelerometer

As the proportion of elder people of our society grows bigger, the movement safety of elderly has become a social problem. For the people who suffer from harmful falling, one of the best steps he can take is ensuring that reliable and immediate help is available to reach him at all times. So it is very important to set up a perfect detecting and alarming system that can monitor the daily movement of the people with falling potential. The main purpose of this research is to design a realizable portable fall detection device which can measure and analysis the real-time body gesture and movement by electronic technology, and if it detects the user's falling movement, the device can send out alarm with the user's geographical location information to some health watch response organization. Obviously, it will be very valuable for the senior that this device can not only improve their personal independence and life quality but also reduce the pressure of their children and the society.Based on the comparison analysis of the literature about the fall detection technology or associated theory, we bring out the design of fall detection device which adopts the technology of 3-axis accelerometer, embedded MCU, CDMA wireless communication and gpsOne with careful consideration of function and cost. The accelerometer measures and collects the user's acceleration and transfer data to MCU in which an embedded algorithm analyzes the user's movement and monitors their body status. If the device decides that the user is undergoing falling down, it will first start the function of geographical locating by gpsOne, and after that it will send out the alarm with this location information by SMS. How to assess the acceleration data, figure out the user's movement and give out the arbitration of falling is the main problem. Furthermore, we introduce the alarm processing mechanism to decrease the possibility of false positive or false negative cases. All the fall detection algorithm and humane interface have been implemented by a practical hardware platform and the relevant embedded firmware at last.After we constructed a prototype, we design a series of experiments to evaluate its performance. The result shows that the fall detection device can measure the user's gesture and when it detects falling movement, it can make arbitration in 20 seconds and send out an alarm with geographical location information. The average accuracy rate of the judgments is 88.57%. The successful implementation of this fall detection device will make great contribution to improving the personal movement safety of the elderly and be very significant for create harmonious atmosphere of our society.