Study on Novel Self-diagnosis at System Level

Abstract

This research suggests unconventional approach to system level self-diagnosis. Self-diagnosis at the system level has traditionally centred on determining the state of the units that are tested by other system units. The suggested approach, on the other hand, relies on the results of tests done by a system unit to determine its own condition. Since a unit examines its own status, which is fundamental in self-testing, such diagnosis is similar to self-testing in many ways. In contrast to self-testing, the proposed approach has a unit evaluate itself based on tests performed on other system units rather than on itself. Different diagnosis models with varied testing assignments and different faulty assumptions, such as permanent and intermittent faults, as well as hybrid-fault situations, are investigated. The diagnosis algorithm for determining the state of the unit has been devised, and its validity has been verified by computer simulation. Conventionally, for self-testing and self-diagnosis it is implicitly assumed that a unit includes some fault-free subsystem capable of executing the diagnosis algorithms correctly.