The University of British Columbia

The authors present a unified probabilistic model of count-based compaction that relates the probability of occurrence of the counted events to a circuit's fault detection probabilities. This model allows an identical treatment of all the different count-based techniques proposed to date, e.g. ones, transitions, edges, and spectral coefficients, by essentially reducing all techniques to simple ones-counting. From a Markov model of ones-counting, the authors derive asymptotic aliasing probabilities, and for finite test sequence lengths they developed a computation technique for determining the aliasing associated with the specifically mentioned schemes, as well as more general count-based compaction techniques, under various error models