To support NASA’s needs for health management technologies to increase safety and mission effectiveness for future space habitats, American GNC Corporation (AGNC) and Louisiana Tech University (LaTech) are proposing the “BRoad Advanced Intelligent Networked (BRAIN) System” consisting of: (i) an innovative Analysis Kernel with evolving cognition based on optimized deep neural networks and collaborative learning; (ii) recognition of new patterns and system trends by a novel Retrospective Change Point Detection (CPD) method for change analysis in temporally evolving systems; (iii) Human-System Integration Subsystem to support active learning to provide a friendly environment for human feedback; and (iv) Distributed Awareness Environment for consoles and optional mobile devices with voice activated commands and effective health information presentation (i.e. faults, data, features, fault cause-effect information, and maintenance actions).
The BRAIN system is tailored for processing NASA ISHM data. A cognitive approach is addressed, where deep learning schemes provide diagnostics capability and support to prognostics considering complex systems interrelations, big data, and uncovering unknown relationships among features. AGNC’s embedded Collaborative Learning Engine is infused and expanded to accommodate human feedback as a methodology for active learning and to process previously unknown system degradation. Two cases scenarios are approached to verify and validate the BRAIN’s core technologies: (a) power system monitoring and interrelations with sensors in a fluid distribution system and (b) leakage detection in pipelines of a fluid distribution system.
The BRAIN system provides an innovative cognitive engine, diagnostic capability for distributed software and hardware architectures, and human-system interfaces. Target applications are NASA’s Life Support Systems (ECLSS), Extravehicular Activity Systems, and Biological Life Support systems (e.g. electrical power systems, environmental monitoring, etc.). Since it coexists with NASA ISHM, applications also include ground and launch systems, distribution systems, various testing facilities, and rocket engine ground testing.
Applications include manufacturing plants, large power plant and power transfer systems, structural monitoring, oil refineries, machinery spaces monitoring, aircraft avionics equipment monitoring, robotic/unmanned systems, military assets in ships and submarines, smart buildings, and others. With some modifications the BRAIN System has the potential to be applied to the Internet of Things (IoT).