The Future of Connected Healthcare: Scalable AI Frameworks for Event-Driven Wireless Sensor Systems

Abstract

Robyn Alexandar

A scalable artificial intelligence (AI) framework is the next critical development in harnessing the benefits of connected healthcare systems. Event-driven wireless sensor systems for pervasive health monitoring can collect data on a vast number of parameters under real-world conditions. The sensor data, when accumulated in large volumes longitudinally, can be used to develop more personalized management strategies for patients, as well as better inform and monitor public health treatments, care delivery strategies, and research investigations than traditional cross-sectional measures. Relatedly, the sensor systems can aid caregivers or practitioners in understanding the effects of medical procedures, new drugs, and behavioral modification. Interpretation of sensor data presents a considerable analytical problem. Pure numerical approaches by identifying linear mathematical relations with outcomes are not an efficient or effective strategy. This is because there is a considerable lag between the acquisition of data and the development of medical signals. A more efficient approach is to combine data science methods that take inspiration from how humans perceive data.

In this paper, we exemplify the deployment of a scalable AI cloud platform for the real-time analysis of a wireless sensor system that captures human physiological data relevant to a number of therapeutic areas. Moreover, we develop an eventdriven cloud platform so investigators and clinicians can monitor their study 'dashboard' for changes and patterns when wearable health sensors capture significant events. Participants, who in the context of this paper are patients with a chronic disease and cognitive symptoms, will find comfort in monitoring and interpreting their real-time data in addition to maintaining patient engagement with the clinical research.

PDF