An Efficient Machine Learning Based Elderly Monitoring System

Due to the rapid growth of the population and the challenges associated with it in the health and social sectors, realtime elderly monitoring systems play a vital opportunity to monitor and control the health and fitness of the elderly. The system proposed in this paper is a portable device that can be used to monitor and analyze various daily life activities such as resting, walking, and falling. An MPU6050 sensor module with a 3-axis gyroscope and a 3-axis accelerometer is used to detect the movements of the elderly and the collected data is transferred to a real-time database. This sensor module is lightweight and requires low power consumption but, it is low sensitivity to body movements, which can lead to improper positioning. To avoid this, an effective machine learning algorithm is developed and trained to obtain accurate estimation. The result of this study, the error value calculated by using the formula of Mean Squared Error of all movements and accuracy of the system is 99.9%.