The 15th IEEE International Conference on Electron Devices and Solid-State Circuits (EDSSC)

511 / 2019-03-04 10:44:29

A Wearable Fall Detection Device with Accelerometer and Gyroscope

Fall detection, Wearable device, feature extraction, Acceleration, Posture, low-power design

Plenary track > MEMS and Sensors

With the increasing aging population in China, the protection of the elderly from accidental injuries has become an urgent problem to be solved. Fall is one of the most harmful behaviors to the elderly, hence the judgment of falls is particularly critical. About this problem, we design a wearable fall detection device with inertial sensors acceleration and gyroscope, which takes advantage of support vector machine （SVM） algorithm to detect and judge the fall behavior. The algorithm builds fall models through collected acceleration and posture data of the human body both falls and non-falls, and then classifies the falls and the non-falls respectively. The experimental results show that the accuracy, recall, precision, and F1 with SVM algorithm on this wearable device are 95.43%, 93.38%, 92.76% and 93.05%. From this we know that the detection effect is quite good and the wearable device can be put into use for the elderly and monitoring system in hospital.