United International University

Nowadays, the burden of disease has shifted dramatically from communicable to non-communicable ailments, a trend particularly significant in Bangladesh. Deaths resulting from such diseases have a major impact on people's lives. Among all non-communicable diseases, cardiovascular disease is highly prevalent in our country. However, many people cannot afford the cost of regular check-ups.

Our paper proposes a prototype demonstrating a mechanism for identifying patients with cardiac disease. Our prototype consists of different sensor modules to collect data, such as Heart BPM, Cholesterol, and ECG readings. We utilize a Machine Learning model to classify patients based on the collected data. The prediction is made visible to the patient instantly so they can take necessary precautions beforehand.

To date, we have managed to collect data from Heart Rate sensors, ECG sensors, Cholesterol sensor modules, and Blood Pressure modules. Currently, we are working on integrating the Glucose sensor module.

The death rate in Bangladesh due to various types of non-communicable cardiac diseases is rapidly increasing every year. These diseases are chronic and progress slowly, often reaching a critical stage that is difficult to manage. Consequently, many people experience sudden heart attacks or only discover their condition when it is too late.

In this paper, we propose an IoT and Machine Learning-based method for predicting cardiac status. This method involves collecting essential data from the human body using IoT devices (sensors) and transferring this data to the cloud, where it is stored securely with user authentication. The data obtained from the human body are then normalized before applying machine learning algorithms to calculate and predict the overall condition of a patient's heart.

So far, we have been able to collect heart rate, ECG signals, and cholesterol levels through IoT devices from the human body. The results obtained for heart rate and cholesterol levels are quite satisfactory. However, we are still facing some challenges with interpreting the ECG signals, which we are actively working on resolving.