Toward a neuro-fuzzy inference system approach for the performance analysis of all-optical half-adder based on CPA

Abstract

The half-adder (HA) is one of the most useful combinational logic functions for data processing. We report on demonstrating an all-optical HA comprising AND and XOR all-optical logic gates (AO-LGs) without nonlinearity and thus with arbitrarily low intensity, ultrafast light-by-light control. The proposed method engages the coherent interaction of optical waves on a metasurface nanostructure, known as coherent perfect absorption (CPA), which is only 50 nm thick, to efficiently control the absorption of near-infrared (700–800 nm) pulses. Particle Swarm Optimization (PSO) determines the most efficient dimensions of metasurface unit cells. Another significant innovation of this article is the utilization of an adaptive neuro-fuzzy inference system (ANFIS) instead of a simple threshold. The flexibility of the designed fuzzy system allows it to be used for any CPA system, applied in the implementation of logic gates. Achieved figures of merit of AO-LGs, including contrast ratio (CR = 89.77 dB) and extinction ratio (ER = 87.53 dB), prove the impressive effects of ANFIS on the gates’ efficiency. Smaller dimensions, lower power consumption, and a reduced logic state range compared to other HAs are additional specifications of the designed HA, achieved through optimized metasurface-based CPA and ANFIS.