Abstract
Key features of many optical signal processing systems match those characteristics appearing on the systolic array approach, namely, “synchrony” on the rhythmical computation of the data, “regularity” in terms of modularity and local interconnections, “temporal locality”, and “pipelinability”. Herein, we show that the R&F mathematical approach can be optically implemented for the systolic array processing of certain algorithms (e.g., matrix-vector multiplication).