IoT-based Architecture for Automatic Detection of Fall Incident using Accelerometer Data

I Wayan Wiprayoga Wisesa (Institut Teknologi Sumatera)
Genggam Mahardika (Institut Teknologi Sumatera)

Indonesian Journal of Artificial Intelligence and Data Mining
Vol 3, No 2 (2020): Spetember 2020

Fall is an unintentional incident that could happened in our daily life. For the elderly, fatal fall incident might increase the risk of death. There is a need to quickly do the first aid after fall incident occur. IoT based architecture made it possible to monitor fall incident remotely. The monitoring device records the activity and object movement using tri-axial accelerometer sensor attached to user’s waist. The system implemented simple thresholding technique based on total acceleration recorded over time. Various scenarios were performed in order to test the system including normal daily activities and fall incident. Using sensitivity and specificity measurement to evaluate the system, the proposed system achieved the value of 98% and 96% respectively.