Elman neural network identify elders fall signal base on second-order train method

This paper presented a Elman neural network identify elders fall signal base on second-order train method. The older human signal analysis have been a research topic health care fields that algorithms build in wearable device real time detect fall situation. However, the user dynamic sport signal nonlinear relationship between time and environmental uncertainty that are not easy to obtain user fall signal. The Elman neural network has recursive neurons weight that record previous signal and current signal to train neurons. Therefore, we utilize feature of Elman neural network to identify user fall signal. Moreover, we used second-order train method to optimal recursive weight. In the experimental results, the Elman neural network identify human fall signal. In addition, we used wearable device combined BLE (Bluetooth low energy) feedback output response real time.