Research On Fault Diagnostics For Multiple State Machine Model

Along with the further study about FSM(Finite State Machine), application based on FSM model has been widely used in all fields, therefore the correctness and reliability of the state machine model is becoming more and more demanding and the test of FSM is becoming more and more important.Most of the study about the test of FSM focuses on the generation method of the test case. However,after finding a fault feature,it’s still a problem worthy to study that how to determine the fault location through fault feature.Existing methods about fault diagnosis of FSM mainly base on the assumption of single fault and two fault model of output fault and transiton fault,and assume that the test case are all executable.Combining with the reality of the state machine test requirements,we add the situation of null output and excutable for the input,so the FSM can excute by test case generated with random way.And we raise the transition with empty output and the transiton that cannot be transformed on the FSM on the basis of traditional model.With the new assumption and new fault model, we formed a set of improved fault diagnosis methods and algorithm on FSM,CFSM(Communicating FSM) and ECFSM(Embedded CFSM) model.After simplifying the process of the existing methods,we divides the process into three stages,which are preprocessing,determining fault and identifying fault.The preprocessing stage gets some basic information and generate the initial set of suspicious.The determing fault stage is used to find all the possible fault by the way of positive and reverse analysis method,with which we present the fault determing algorithm. In the stage of identifying fault,we use additional test cases to identify which possibility is the single fault. We give some specific solutions and principles and the solution under different situation.After comparing the different feature of three model,we present different fault determing algorithm.After that,we analyse the correctness and accuracy of the algorithm.In the end, we present a specific example to describe the implementation process of the algorithm.Results indicate that the algorithm can efficiently and accurately diagnose the fault.In this paper, the fault diagnosis method we propose enriches the research of fault diagnosis method based on state machine,so it has great theoretical significance.And it is very useful for the test of system based on FSM, CFSM and ECFSM model, so it has a good practical significance.