Data Stream Management for CPS-based Healthcare: A Contemporary Review

Abstract

ABSTRACT Data stream management (DSM) for cyber-physical systems (CPSs) provides good quality care services in the medical domain. This is a very prominent field of research that includes sensing, processing, and networking of various medical devices. The DSM for CPS is a combination of computation using Cyber world (computers or WBANs), including wearable sensors and smart meters as well as communication between the processes through the networks. Data analytics and mobile computing include the usage of wireless sensors which plays a very significant role while handling uncertainties of data stream in healthcare domain. This paper presents a comprehensive review of DSM techniques, including the problem of concept drift in the healthcare domain using CPS, and the challenges associated with the domain. The complete taxonomy characterizes and classifies all the components and methods required for data management in healthcare. This paper provides a glimpse of futuristic techniques used for DSM in view of concept drift while handling real-time data management in healthcare and also identifies the fields for future research. The prime objective of this review is to provide a solution to aggregate health data streams using DSM, generated from different sources, clean and normalize them, and improve them for analysis, diagnosis, pattern identification for data analytics and complex event processing. It is supposed that the techniques discussed in this paper are relevant and useful for further research in the area of DSM for CPS in healthcare.