Enhanced Modelling for Detection of Daily Activities of Aged People for Health Care

Abstract

Millions of individuals around the world are sixty years of age or older. Global concerns about aging populations have been impacting social welfare and healthcare systems. As people get older, people find it more difficult to carry out their daily tasks, which affects their mental and physical health. Additionally, there are increasing events affecting the elderly, especially in and around their homes, with serious consequences including hospitalization. The most prevalent cause of injuries among the elderly is falls. Human Activity Recognitions (HARs) are techniques that use (ML ) algorithms to identify human activities from sensor data and categorize human behaviors such as standing, walking, and running. HARs are crucial for human behavior analysis and human-computer interaction in applications related to serious illness identification, patient rehabilitation, and healthy lifestyles. A growingly useful use of MLis the prediction of human behavior, where computers that keep an eye on routines and step in when there is a crisis or a shift in behavior may be very helpful to the elderly. Despite the fact that many academics use ML approaches, only few studies about HARs of elderly people exist. By concentrating on tracking human motions utilizing perceived accelerometer and gyroscope data, this effort aims to close the aforementioned gap. According on the results of experiments, this paper proposes the Scaled Regressive Human Activity Detection Model (SRHADM), an ML -based schema that can accurately identify human activities 90% of the time. Human centered applications such as actively supported living and home monitors can be properly utilized with the help of SRHADM implementations. Since it could shed light on the possible therapeutic or research uses of HARs, this knowledge might be important for clinical applications.