Accurate Fall Localization for Patient based on GPS and Accelerometer Sensor in Outside the House

Falling of the elderly patients was increased every year and can cause head injury or broken bones. The elderly cannot move or making calls to the medical emergency center by himself after fall, which leads to increase the danger of fall injuries. This paper aimed to design and implemented a wearable fall detection system (FDs) which is used for detecting the falling for the elderly and helps them after a fall by making automatic communication between them and caregivers. A proposed system which includes the Bio-sensor, localization module based on GPS, controller unit, and GSM unit will alert the caregivers by sending messages contained his/her health status and geolocation (latitude and longitude coordinates) immediately when he or she passes a selected threshold of falling. The proposed system was validated by using statistical analysis based on mean absolute error. The fall detection system appeared the ability to distinguish between fall and normal daily activities of the elderly with an accuracy of 99.2% based on the proposed algorithm which compared acceleration magnitude and fall threshold values, called the FDB-FTAM algorithm. The result showed, the proposed system was satisfactory performance and more accurate when get the patient location with 1.938×10
^-5
°
, and 1.327×10
^-5
°
of mean absolute error for latitude and longitude information respectively.