Abstract
The US Navy Integrated Diagnostic Support System (IDSS) includes an adaptive diagnostic subsystem (ADS) that provides an online fault-isolation capability using multiple reasoning techniques; a local adaptation mechanism is provided for updating the numerical parameters (e.g. relative failure probabilities) associated with the system model. A more global and exhaustive feedback can be accomplished through the feedback analyser (FA) tool, capable of symbolic as well as numerical learning and providing a differential or comparative analysis among systems of the same type. The ADS and FA tools, as precursors to a complete prognostics capability, are described in terms of their operational, functional, and performance features.