A Novel Approach for Statistical Parameter Estimation and Test Pattern Generation

The importance of signal probability concept has been recognized for testability analysis and for power estimation. Monte Carlo based simulations forms a class of computational algorithms that utilizes the idea of repeated random sampling for the estimation of unknown parameters. These simulations are mainly targeted to estimate the probabilities of various outcomes in a process that are characterized by random variables. This work focuses on the estimation of statistical parameters for signal arrival time distributions at the output nodes for ISCAS' 89 benchmarks. It uses MAX and SUM operator for identification of critical paths within the design. Test patterns were generated for the transition delay faults and compaction was carried out. The performance metrics compared includes power, fault coverage and test coverage. The test coverage reported was 98.55%. A maximum of 5% increase was observed in the average capture switching activity after compaction of test patterns.