Abstract:
Mathematical models of components of a vector-matrix optoelectronic multiplier are considered. Perturbing factors influencing a real optoelectronic system – noise and errors of radiation sources and detectors, nonlinearity of an analogue–digital converter, nonideal optical systems – are taken into account. Analytic expressions are obtained for relating the precision of such a multiplier to the probability of an error amounting to one bit, to the parameters describing the quality of the multiplier components, and to the quality of the optical system of the processor. Various methods of increasing the dynamic range of a multiplier are considered at the technical systems level.