Information quality control for network centric ship maintenance

This research formulates and prototypes an intelligent multi-user mechanism, called an Information Broker, for context dependent time critical manipulation of dispersed and transient maintenance data embedded in autonomous systems where semantic heterogeneity is the norm. This paper presents the fundamentals of a relational calculus which models a prespecified set of relations across the electronic data sources as a distributed search tree of meta-data terms. The semantic distance between meta-data terms is quantified to represent semantic precision. The incoming health monitoring and operational data are classified into observable or predictable maintenance event sets. When a maintenance event is identified the Information Broker is called upon to support unstructured information queries to find related diagnostic tools, built-in-tests, procedures, policy and planing information with specified precision. An Information Broker has been prototyped as part of a real-time maintenance decision support tool, deployable on-board the USS Yorktown for online condition assessment, interactive fault isolation and maintenance decision support for power plants and electric motors