Agent-Based Health Monitoring System

IAI has successfully used principal component analysis to detect even subtle latent faults in a subsystem. An advantage of the PCA for this application is that the system can be real-time, even when there are very large data sets to be analyzed. IAI has also successfully used software autonomous agents to provide robust, flexible, optimized control of large complex decentralized systems such as air traffic control, battle management, etc. The innovation of this proposal is the integration of these previously separate technologies to provide system health management and self-reliant systems. The system to which we will apply this innovation is a power grid where nodes can be power producers, power consumers, power transmission elements, and in some cases nodes can be either producers or consumers at different times. Our software agents interact with each other based on a contract net paradigm. IAI has been working on agent-based systems for a decade and has an extensive library of tools and techniques to design and implement agents interacting via a contract net. These tools and techniques will be used in the proposed work and will allow us to implement a simulation of a non-trivial system even during Phase 1.