Human Activity Recognition: Trends and Challenges

Internet of Things (IoT) usage has increased due to advancements in information and communication systems. IoT technologies in contemporary healthcare applications connect doctors and patients for automated and knowledgeable everyday activity monitoring for older adults and older adults. The use of wearable body sensors and mobile devices to track personal health care is becoming more widespread. Wearable sensor technology is one of the key IoT advancements in healthcare tracking systems. IoT integration in healthcare has also sparked the development of intelligent apps like smart healthcare and sophisticated healthcare monitoring systems. Human Activity Recognition (HAR) has received a lot of interest from pervasive computing researchers working on smart healthcare systems. Patients with heart disease, obesity, and diabetes must engage in regular physical activity as part of their therapy plan. Some people with mental illnesses must monitor their everyday physical activity to avoid unfavorable conditions. As a result, identifying and keeping track of physical human activity can help a health professional evaluate a patient’s behavior in terms of their health. HAR system must, therefore, be made accessible to end users to protect their health. Over the past few decades, much work has gone into developing a strong and reliable framework for action recognition and prediction. In this chapter, we discuss the trends and challenges of machine learning-based HAR systems and offer solutions for them.