Utilizing Edge AI in a Connected Healthcare SoS (System of Systems)

158
Artificial intelligence (AI), data mining, expert system software, machine and deep learning and another modern computer technologies concepts. Brain representing artificial intelligence with printed circuit board (PCB) design.

By Maria Palombini, Practice Lead, Director, Healthcare and Life Sciences, IEEE Standards Association (IEEE SA) and Srikanth Chandrasekaran, Practice Lead, Senior Director, Foundational Technologies, IEEE Standards Association (IEEE SA)

In healthcare, every second matters – including the ability to process data that can alert and enable smarter and faster decisions by healthcare providers. 

The demand for real-time, cost-effective, and efficient services continues to push technological innovation, including the transformation to edge computing infused with Artificial Intelligence (AI). The evolution of Edge AI is empowering healthcare systems with immediate, real-time notifications from IoMT devices, such as wearables and biosensors. When abnormalities are detected, such as a patient falling or a warning signal from a pacemaker, real-time notifications alert caretakers or healthcare providers. Clearly, this technology has revolutionized the approach to healthcare. 

But there remain more opportunities. A connected healthcare system is composed of thousands of connected devices, operating systems, networks both within and outside of a facility with multiple parties in different areas accessing, utilizing, storing, and computing the data. This presents a daunting challenge: how can we enable this infinite number of links to be connected and informative? While the standardization of interoperable devices might someday be achieved, an alternative path may be a System of Systems (SoS) approach.

The Evolution of Edge AI

The evolution and ubiquity of more powerful computers and advances in AI applications has spurred the rapid advance of edge computing in numerous industries, including healthcare. Edge AI can process growing amounts of data at the point where it’s created.

Unlike Cloud AI whereby data is sent from a device or system to the cloud, Edge AI analyzes massive amounts of data locally. The more obvious benefit is the reduction of latency, or delay in transmission of data, because the data doesn’t have to be transmitted to the cloud. Consider the continuous streams of data from IoT devices as well as the enormous data from medical digital images from MRIs, CAT scans, nuclear medicine, CT scans, and ultrasounds. All of it must be transmitted, captured, and analyzed. Edge AI can decipher data more in real-time, slashing time for providers to make actionable patient care decisions.

Additionally, there are other benefits. Because Edge AI does not require transmission of large amounts of data to execute its processing ability at the edge, organizations can potentially realize reduced costs for data communication. Security and privacy are always concerns for providers and patients. With Edge AI, most data remains in its location where it is captured, thus avoiding the risk of transmitting data across vulnerable internet, data and storage networks.

System of Systems (SoS) & Swarming Data

Healthcare organizations are embracing the benefits of Edge AI, but a greater opportunity remains to unlock the barriers that would allow organizations to connect thousands of devices and distributed data processors in a way that could arm providers with better ability to better diagnose and treat patients. Instead of exclusively relying on solving the data and device inoperability challenges, some are advancing the development of a System of Systems (SOS) approach with the help of Edge AI.

Think of SoS as a superstructure or dashboard, leveraging AI to coordinate and aggregate data from many different points, often referred to as “swarm data.” The SoS approach would aims to connect an organization’s multiple, disparate systems by leveraging Edge AI and algorithms designed to harvest insights from data – not the actual raw data. In this approach, the data remains in its point of origin; and system of systems would be aggregating the insights generated from the trained AI algorithms which provide the value needed to make the necessary, faster, and smarter decisions from the use of these multiple devices. 

Processed at the edge from these different systems, these insights are presented in a way that enables swarm learning – sourcing larger datasets of insights – which can empower more accurate diagnoses and better treatment decisions. Additionally, an Edge AI-powered SoS can allow devices to make immediate decisions without human intervention such as an insulin pump automatically dispensing a dose when a sensor alert signals the patient needs to be dosed.

The Future of Learning & Treatment

In the early stages of development, Edge AI-powered SoS systems face significant barriers to overcome include the lack of standardized data and protocols, overall effective distributed data management, and ensuring privacy and security at the edge to ensure system-wide and patient trust. Despite these challenges, a connected healthcare system featuring swarm learning holds great promise as one means of using technology to improve a more efficient, intelligent delivery to achieve improved patient experiences and outcomes.

Maria Palombini currently leads the IEEE SA Healthcare & Life Sciences Practice working with a global community of stakeholder volunteers who are committed to establishing trust and validation in tools and technologies that will change the approach to discover therapies, deliver care, and ultimately enable a sustainable and universal quality of care for all. She holds a B.S. and B.A. from Rutgers College and an M.B.A. from Rutgers Graduate School of Business at Rutgers University.

Srikanth Chandrasekaran leads the IEEE SA Foundational Technologies Practice, focused on

developing key programs which address core issues of security, identity, trust and building end-to-end trustworthy devices and systems across emerging areas such as IoT, smart cities, sensors

and blockchain. Sri also heads the standardization activities for IEEE SA for the Asia Pacific region.

Additionally, he heads the IEEE Blended Learning Program effort, driving the development of an eLearning platform, focused on bridging skills for students in current and emerging technologies as well as lateral skilling of industry professional. Sri earned a B.S. in physics from Madras University, India, and a post-graduate degree in electrical communication from Indian Institute of Science in Bangalore, India.