An IoT System for Post-Myocardial Infarction Patients during Cardiac Rehabilitation (CR) in Indonesia

The American Heart Association (AHA) has demonstrated that involvement in Cardiac Rehabilitation (CR) can reduce Myocardial Infarction (MI) or heart attack recurrence. However, CR participation in Indonesia was found to be minimal after hospital discharge. Understaffing and lack of availability and access are some of the reasons. An efficient Internet of Things (loT) system could overcome inadequate staffing and improve the overall outcome of CR. The proposed loT system consists of multiple loT technologies and a wearable device (E4 wristband). This paper focuses on building a framework with the aim in developing an loT system for post-MI patients during CR in Indonesia and the development of a Convolutional Neural Network (CNN) based neural network model to classify human activity (HAR) based on labelled accelerometer data taken from the chosen E4 loT device. This paper also explains the process of training and testing of 2-dimensional Convolutional Neural Network and the development of selecting the wearable device. This paper also explains the medical and ethical guidelines for developing our framework. At the end of the study, it was observed that the proposed model had recorded a high classification accuracy. The CNN model had a classification accuracy of 96% of all six activities.