Incorporating model-based reasoning in interactive maintenance aids

An approach to generating an intelligent maintenance decision aid which is implemented in a tool called POINTER (portable interactive troubleshooter) is presented. The approach involves developing a mathematical model of the system to be tested, and then reasoning about the system to determine which test to choose and what one can deduce from the test outcome. The mathematical model is based on the flow of information through the system and the type of information available at each test point in the system. The corresponding test dependencies are then identified. This model is used to guide the fault-isolation process. In addition to the basic fault-isolation capability, the model provides tremendous flexibility for diagnosing a system. Failures can be hypothesized. Test choices can be overridden or delayed. Various optimization criteria can be specified. Each of these capabilities is discussed