A data-driven machine learning integrated wearable medical sensor framework for elderly care service

Abstract

Recently, Wearable Medical Devices (WMDs) is a promising medical technology to improve patient treatment in various medical perspectives. However, the traditional model-driven framework faces challenges on narrow reconstruction ratio and compression ratio due to improper measurement of signals in sensors. In this research, a Machine Learning assisted Integrated Data-driven framework (MLA-IDDF) that can acquire signal features based on personalized characteristics from elderly patients to improve compressive sensing performance with a reduced number of measurements for precise model. The system coordinates the collection and processing of raw sensor data from a variety of sources in a common information database. MLA-IDDF develops semantic models to describe patient’s conditions, and decision-making mechanisms based on the interpretation of collected data that provides accurate observations about elderly patients. The Simulation analysis at the lab scale shows the performance of the system and user satisfaction assessments confirmed that the system is superior to other conventional systems.