Combined optimization of area and testability during state assignment of PLA-based FSM´s

Stuck-at and crosspoint faults in PLA´s, introduce combinational and sequential redundancies into PLA-based FSM´s, this affecting the testability of these FSM´s. In this paper, we propose a new state assignment algorithm for PLA-based FSM´s called EARTH, which considers both aspects of area minimization and testability of the resultant PLA´s with the fault model containing single stuck-at and/or single cross-point faults. We have also developed an algorithm that checks for redundancies in a PLA-based FSM. We have found the redundancies in FSM´s that have been state encoded using EARTH with our redundancy detector and areas of these MCNC Benchmark FSM´s. EARTH has yielded 5 times less undetectable faults on these FSMs than NOVA (approx) with areas 2% more than NOVA (approx), on an average