Improving the Effectiveness Of Iot-Based Healthcare Monitoring and Control Devices using Deep Learning Models

Abstract

Internet of Things (IoT) is a boon to the healthcare industry, which is due to its inherent
advantages like ability of remote monitoring, remote actuation, high speed data processing
and low operation costs. IoT devices like electrocardiogram (ECG) monitor, continuous
blood pressure (CBP) monitor, continuous oxygen level monitor, etc. are being used by
Doctors worldwide to effectively monitor patient conditions and reduce overheads on the
currently overloaded nursing staff. Due to high availability of computational resources and
limited power constraints at hospitals and other healthcare stations, IoT devices can step up
their performance via high accuracy temporal data analysis and decision making in case of
even the slightest of anomalies. Some of the newer IoT systems have implemented such high
accuracy algorithms for improving their applicability in real time monitoring & control. But
the older systems need to be replaced to upgrade their performance, which raises a lot of
issues, including but not limited to cost of replacement, calibration, familiarity of nursing
staff with old equipment, etc. To reduce the effect of these issues this chapter proposes a
novel high accuracy, highly interfaceable, and low cost deep learning solution, which can be
integrated with both old and new healthcare monitoring devices to improve their efficiency.
Certain minimum application criteria are defined for a device to be eligible for interfacing,
and it is observed that more than 80% of currently working healthcare devices can be
upgraded via this architecture, and their effectiveness of monitoring & control can be
improved. The proposed architecture is found to be more than 99% accurate in terms of
parameter monitoring, and has excellent control exercising capabilities.