Deep Analytics for Data in Cyber-Physical Systems

Visual Analytics Managers (VAMs) for Cyber-Physical Systems (CPS) will allow users to gain insight into the CPS with state-of-the-art analytics and artificial intelligence and provide human-in-the-loop interaction with the CPS. Our VAM toolkit builds around a Scala Akka actor framework, allowing designers to define the sensors and actuators present in the CPS, select from a rich set of analytic and machine learning tools for processing data, and select views to display and interact with the data. The toolkit uses a publish-subscribe protocol over established low-level communication protocols. Using the Scala Akka actor framework allows VAMs built with our toolkit to scale to include millions of data streams. Toolkit scripts use actor factories to create source code and then compile them into bytecode for execution in a Java Runtime Environment (allowing for platform independence). In Phase I we will demonstrate our approach"s feasibility by implementing scripts for the highest risk functionality to build the core of a Casualty Analytic Real-time Evaluation System (CARE). CARE will perform data analytics, visualization, and machine learning informed suggestions to aid in tactical combat casualty care and triage. During Phase II we will complete the toolkit and a fully functioning CARE System.