An application of Bayesian reasoning to improve functional test diagnostic effectiveness

This paper describes a software package that embodies a Bayesian reasoning engine and modeling schema to significantly improve the ability to discern the defective component causing a failed functional test. This software approach brings to functional test similar diagnostic capabilities that have become familiar to test engineers working with X-ray, automatic optical inspection (AOI) and in-circuit test (ICT) test technologies. This software package, known as Fault Detective, provides significantly improved diagnostic accuracy as compared to human efforts, and works with exactly the same data set as is currently available for diagnostic purposes. The model is based on the interaction of the functional test suite with the product functional block diagram. This approach also means that the software package is highly independent of the technology behind the system being diagnosed.