A patient-specific single sensor IoT-based wearable fall prediction and detection system using machine learning

Abstract

Technologies in health care are increased effectively. Normally, old peoples have a serious problem, i.e., falls, making the elder people fall and injured severely. In health care, this is a serious problem and has become a challenge through worldwide. There are many studies and research to detect the fall in old age person by using various sensors and several types of cameras. Several sensors were implemented to predict fall detection, and prediction mainly works on the channel state information (CSI), giving less efficiency on the results. In this paper, using Machine learning (ML) methods, the study analysis to detect falls is implemented. By using machine learning algorithms, the efficiency gains will up to 96%. The main advantage of the system is it can be used as a wearable device. In the market, MEMS sensors are available to detect the acceleration changes in a person. The alert for the doctor is achieved by MY MQTT. The previous data is used to store in an excel sheet or pen drive, and then the mems sensor of X, Y values are taken and at some point of time extract the data by analyzing the accelerometer data and previous data. From that, if fall detection takes place, it alerts like alarm sound and also it sends a message to the caretaker and hospital people. If a fall is not detected, it once again checks the entire process.