On the Sequential Diagnosibility of a Class of Digital Systems,
Piero Maestrini, Liu,C L · Defense Technical Information Center (DTIC) · 1978
We use a model that was first introduced by Preparata, Metze, and Chien. In this model, a digital system is partitioned into a certain number of units, each of which can be at one of two possible states, fault-free and faulty. A configuration of a system is an assignment of either the fault-free or the faulty state to each unit in the system. We assume that each unit in the system possesses a certain amount of computational resources to enable it to test is a binary signal which depends on the state of the testing and the tested units. In particular, we assume that: (1) if a fault-free unit is tested by a fault-free unit, a signal 0 will be generated; (2) if a faulty unit is tested by a fault-free unit, a signal 1 will be generated; and (3) if a fault-free or a faulty unit is tested by a faulty unit, either a signal 0 or a signal 1 will be generated. (In other words, the signal generated by a faulty testing unit is completely unreliable.) A diagnosis experiment is one in which every unit tests all the units it is capable of testing once. The outcomes of the tests are referred to as a syndrome.