Tackling cost optimization in testable design by forward inferencing

Abstract

The authors describe how the forward chaining mechanism of an expert system's inference engine is employed to achieve automated cost optimization in designing testable circuits. The characteristics of forward chaining are used to explore the set of applicable testability strategies. This exploration is preceded by a testability analysis, using a set of testability rules with the emphasis on maximization of fault coverage. The analysis locates the places in the design that are not optimally accessible, taking into account the characteristics of the concerning parts, such as function and design style. A testability synthesis is accomplished by the forward chaining inference engine. A conflict set is compiled, containing rules which relate to testability strategies that will relax the violations of the analysis rules. The forward chaining mechanism in combination with the structural testability analysis directly drives the testability strategy that makes the best use of the available circuit resources. >