Effect of phase shifted signal propagated in optical fiber into minor laser neural network.

In this simulation study, a primary optical neural network called light brain, is proposed via theoretical simulation based on a feed-forward model. Calibration is assumed for the nonlinear behavior in such a network by a semiconductor laser of the VCSEL type. The proposed part of the network is assumed to be constructed from three lasers: the first laser is assumed to be an influencer, followed by two modulated follower lasers. In such an argument, an input layer is present. The transmitting medium is an optical fiber, which separates the influencer laser signal into two new signals. The first signal is subjected to specific phase shift fluctuations (0, 90, and 185 degrees) and then detected and uploaded via modulation as a radio frequency (RF) through laser DC pumping to one of the two followers. The second signal phase is kept constant, detected, and modulated directly to the second follower. Finally, both of those two new signals are mixed optically. The impact of each signal on the overall signal that occurred nonlinearly inside both of these two lasers is evaluated, where they are expected to be correlated. This can be modeled to approximate more complex functions within neurons and neural networks. Results indicated that a higher phase shift value introduced the most spikes and chaotic behavior, i.e., more transmitted signal security inside the simulated electrochemical signal. Spectrum broadening is observed due to this interaction with a width value of 0.005 THz. [ABSTRACT FROM AUTHOR]