Integration of neural networks with diagnostic expert systems
D. Hornig, R. Aschenbrenner, Rajiv Enand · 2002
The integration of expert system and neural network technologies is a promising approach to solving diagnostic and field service problems. A hybrid system has shown the feasibility of integrating these two technologies. It uses a neural network to perform an initial diagnosis via acoustic signal recognition, and uses an expert system to perform follow-up tests leading to a specific diagnosis. The hybrid successfully diagnoses a simulated mechanical fault using acoustic information and expert-level knowledge, demonstrating that a standard low-cost platform can support a combination of neural network, expert system, and data acquisition software. This hybrid technology has potential applications in diagnostics and prognostics applications where the available diagnostic evidence includes both signal and symbolic information. The hybrid technology is particularly appropriate for situations that require rapid development and cost-effective maintenance of the diagnostic system.>