A systematic approach of designing fault-tolerant systolic architectur
es is proposed in this paper. In this approach, redundant computations
are introduced at the algorithmic level by deriving three versions of
a given algorithm. Fault-tolerant systolic array is constructed by me
rging the corresponding systolic array of the three versions of the al
gorithm, The merging method attempts to obtain the fault-tolerant syst
olic array at minimal cost in terms of area and speed, It is based on
rescheduling input data, rearranging data flow, and increasing the uti
lization of the array cells. The resulting design can detect and toler
ate all single permanent and temporary faults and the majority of the
multiple fault patterns with high probability.