AN APPROACH TO DESIGNING CHECKING EXPERIMENTS BASED ON A DYNAMIC MODEL

Abstract

This chapter describes an approach to designing checking experiments based on a dynamic model. In the usual form of machine identification experiment, the experimenter is to determine the behavior of an unknown sequential machine by applying an input sequence and recording the output sequence. The problem of finding an input sequence, which allows the experimenter to characterize completely the behavior of the unknown machine, cannot be solved unless the machine is strongly connected and an upper bound on the number of its distinguishable states is known. It is always possible to devise an experiment that allows the prediction of all future behavior of the machine. To test a machine thoroughly, it is necessary to devise a machine identification experiment, which assures that the machine under test has a given state diagram, to the exclusion of all other state diagrams that can result from a fault. The checking experiment consists of a homing experiment, followed by a series of tests, which check the response to each state and transition by means of the distinguishing experiment. The gedanken-experiments on machines are usually classified as either preset or adaptive.