Self-Evolving Maintenance Knowledge Bases

Intelligent Automation Systems, in collaboration with the Georgia Institute of Technology proposes the development of Self-Evolving Knowledge Base architecture. Specifically, this project will consist of 1) Analysis to identify maintenance related issues such as failure modes, criticality, etc for a selected test bed that will comprise of a critical subsystem or component; 2) Systematic knowledge engineering steps to characterize the maintenance, logistics, and inventory related issues, and to elicit domain as well as control knowledge; 3) Development of a novel architecture and algorithms for the integration of the control knowledge of self-evolution for a continuous improvement process in a feedback manner; 4) Development of testing plan to validate and verify the self-evolving architecture on a subsystem or component; and 5) Software demonstration of the framework. This proposal involves a novel artificial intelligence technique of reasoning and learning in episodic manner. This methodology forms the core of what is presented as a comprehensive CATER (CAse-based TEmporal Reasning) solution that will exhibit self-evolution process by improving the performance of knowledge-base by observing the feedback generated from multiple sources, such as maintenance logs, IETMs, inventory footprint, component statistics, etc. The solution has a built in capacity to reduce the uncertainty that is inherent in the PHM and maintenance processes.