Intelligent Wearable Healthcare Monitoring Framework: Trends in Sensor-Deep Learning Approaches

Introduction

The detection of ailment symptoms at an early stage is an essential element of any automated or artificially controlled diagnostic system (Stokes et al.,2022). The design of such a predictive diagnostic tool is a challenging task and it shall provide assistance to the medical professionals to obtain prior analytics regarding the state of an ailment of a patient. It is also crucial in playing a significant role in formulating probable treatment mechanisms. Further, it is likely to help in monitoring and deriving appropriate decisions about the nature of ailment, treatment and recovery (Dash et al., 2019). Safeguarding one’s health is optimal for longevity and overall well-being. Many people among us lose their lives due to different health issues. Diabetes and high blood pressure (BP) are the two most common and dangerous diseases that impair human body functionality and increase the risk of cardiovascular disease (Sharma et al., 2020). Early detection and classification of diabetes and high blood pressure into their respective categories is therefore critical for effective patient treatment. Furthermore, accurate monitoring of diabetes and blood pressure can provide an opportunity to protect the patient from cardiovascular diseases (Chatrati et al., 2020).

All medical devices have been digitized as a result of advancements in information and communication technology (ICT). These digital devices make living easier and more comfortable. As a result, people use various devices in their daily lives, such as smartphones and wearable sensors. Smartphone’s may include sensors that can be used to gather data about the human body. Wearables have the potential to collect a massive amount of patient vitals (Peral et al., 2018). Both of these devices can be used to monitor the patient in real time. Several healthcare systems have been proposed till date to monitor diabetes and blood pressure patients using smartphones and wearable sensors (Gia et al., 2019; Saravanan et al., 2017; Alifan etal., 2018; El-Sappagh et al., 2019; Ali et al., 2018; Siddiqui et al., 2018; Su et al., 2019; Arakawa 2018). However, these systems are not well-equipped to derive user-specific decisions in complementary supplementary roles after collecting data in real-time (Vijayan et al., 2021). Furthermore, digital devices generate a massive amount of ever-increasing healthcare data, which existing systems find difficult to process, manipulate, disseminate and store for accurate patient health monitoring and predictive diagnosis (Dash et al., 2019). Furthermore, extracting and effectively analysing valuable information from healthcare data has become a new challenge for existing healthcare monitoring systems. The data generated by these devices is unstructured, making it difficult to manage for chronic patient monitoring. Furthermore, the amount of data is growing at an exponential rate, necessitating massive storage space. To accurately monitor diabetes and blood pressure patients, store their healthcare data, and perform predictive analysis on structured and unstructured data, a smart methodology and a cloud-based healthcare architecture are required.

The proposed work is related to the development of an intelligent healthcare framework using wearable devices, social networking platform, internet of things (IoT), and cloud technologies to improve patient survival prediction with minimal human intervention and process driven feature engineering. By the help of this type of system and wearable sensor, health monitoring can be done for a patient who is living in a remote area because it would be difficult for such patients to regularly visit hospitals or go for check-ups on a regular basis. So, this system can reduce the risk of life and improve the health status of a person. Existing healthcare monitoring systems have limitations in extracting valuable information from sensor data, shared content, and networking, as well as difficulties in effectively analysing these.