Built-in self-test generator design using nonuniform cellular automata model

The paper presents a new test vector generator construction technique for built-in self-test (BIST). The technique is based on the cellular automata model nonuniform cellular automata (NUCA). In NUCA, cell neighbourhoods are not predefined but decided for each cell dynamically by the test vector set. The problem of finding the minimum NUCA topology that can generate a given precomputed test vector sequence is worked on and reduced to independent set-covering problems. Also, a polynomial time algorithm that decides on a small, but not minimal, cellular topology is introduced. Simulations using benchmark circuits showed that the hardware component cost of a test vector generator based on the NUCA model is considerably less than the cost of a single programmable logic array (PLA) approach