Prof. Yan Luo as the PI officially awarded a NSF grant for Secure Data Architecture on Sep. 3, 2015


Title: CICI: Secure Data Architecture: STREAMS: Secure Transport and REsearch Architecture for Monitoring Stroke Recovery
Award #: 1547428
Investigator(s): Yan Luo Yan_Luo@uml.edu (PI), Martin Margala (Co-PI), Xinwen Fu (Co-PI), Yu Cao (Co-PI)
Amount: $499,858.00

Abstract

The rehabilitation of stroke patients is a long but critical process for their long-term wellness. Monitoring patients with wearable sensors and web cameras can support at-home rehabilitation by reducing the risk of events such as accidental falls and inappropriate dietary intake. Such sensor-generated live data streams about patient status and activities are processed at data centers for real-time analytics, helping healthcare professionals to respond to patients' needs quickly and effectively. Since the data streams may contain electronic Protected Health Information (ePHI), they must be protected so that transmission and usage conform to security and privacy regulations, such as Health Insurance Portability and Accountability Act (HIPAA) and applicable state laws. Therefore, it is important to investigate advanced networking and computing technologies to meet these security requirements, which are critical for bringing sensors and data analytics from research to clinical environments.

This research plans to address the security challenges in transferring and processing patient related sensor data by developing a Secure Transport and REsearch Architecture for Monitoring Stroke Recovery (STREAMS), a technical proof-of-concept implementation, to secure end-to-end sensor data streams using secure software defined networking and elastic compute and storage resources. STREAMS will be the first prototype of a secure network architecture to provide advanced data analytics-based healthcare to stroke patients in a realistic clinical environment. This project addresses issues in securing heterogeneous sensory ePHI patient data. It captures the workflows of patient data analysis and defines a role-based security enforcement framework to apply access policies. A Secure SDN controller will be designed to authenticate, identify, and direct encrypted data streams to ensure the data streaming over the network are HIPAA compliant, provide guidance in provisioning of compute resources at the cloud, and apply the most appropriate decryption algorithms based on the role of users, priority, types and source of the sensor data stream, as well as network conditions. A generalizable secure hardware and software architecture collects, encrypts, decrypts, stores, transports, analyzes, and maintains the integrity and availability of the data from these multimodal sensors to enable them to be fused using analytics algorithms to learn about patient activities that are highly relevant to stroke recovery. The highly interdisciplinary project team consists of healthcare professionals, medical researchers, computer scientists, IT staff, engineering staff, and industrial partners.