Carlos Henrique da Silva Santos

Free Visual FDTD 2D Simulator to Support the Telecommunication Teaching-Learning Process

This work presents the integration of different computational techniques to specify and design a new computational electromagnetic simulator, called Free Visual FDTD 2D (FV-FDTD-2D), to support the Telecommunications teaching-learning process. Some Software Engineering, human-computer interface (HCI) and an advanced numerical method, based on Finite Difference Time Domain (FDTD), were successfully integrated. It provides an efficient processing and usable software, which detailed documents for developers and comprehensive graphical output results. The FV-FDTD-2D has been adopted in some technical lectures related to telecommunications, presenting successful evaluations by the attempted public.

Human-Computer Interface Techniques to Design and Evaluate an Electromagnetic Simulator

This work presents the interrelation of software engineering and computational semiotics applied to establish a new procedure to design and specify electromagnetic simulator, contributing to define a possible solution for a complex usability problem for engineers and scientists who wants to use this type of computational approach. It is justified by the large number of Physical properties necessary to perform any numerical simulation, specially, for telecommunications. In this work the main integrated techniques are Service Oriented Architecture (SOA), Nielsen principles with severity levels, signs representations from computational semiotics and a cognitive route technique, and they were integrated to establish a procedure to design and test this type of numerical simulator, allowing to identify some breakpoints in these simulator projects before its delivery. These integrations present successful results, eliminating interface errors and failures, improvements in human-computer interactions and, consequently, it contributes perform a faster procedure to develop an interface for an electromagnetic simulator, which can be applied in science and didactic procedures.

Evolutionary strategy algorithm applied to optimize micro-to-nano coupler devices

This work presents the integration of finite element method and artificial intelligence approaches to model and optimize non-intuitive micro-to-nano couplers. The non-intuitive term is related to the columns arrangements into the photonic couplers, which best optimized solution coupled 73.94%. Another important advantage is the possibility to fabricate this device into Unicamp facilities.

Towards an integrated evolutionary strategy and artificial neural network computational tool for designing photonic coupler devices

Abstract Photonics has been widely explored in computing and communications, mainly to rationalize the relationship between device size minimization and data processing/transmission maximization. Generally driven by optimization and modeling techniques, the design of photonic devices is often performed by bio-inspired algorithms integrated to electromagnetic solvers, which have achieved advances but is still time-consuming. As an alternative to a costly finite element method (FEM) solver, a multilayer perceptron (MLP) neural network is proposed for computing power coupling efficiency of photonic couplers, originally designed through an integrated evolutionary strategy (ES) and FEM routine. We address the ES-FEM design of two efficient couplers, present the MLP implementation and the MLP training and testing over the routine generated datasets, and measure MLP and FEM runtime. MLP suitably predicted the power coupling efficiency of a variety of unknown couplers on tests. The measured runtime showed MLP is ∼105 faster than FEM. In conclusion, MLP is a potential tool to be integrated to ES on the design of such photonic couplers.

On the accuracy of approximate descriptions of discrete superpositions of Bessel beams in the generalized Lorenz-Mie theory

This paper aims to investigate one of the most important schemes, viz. the localized approximation (LA), for describing arbitrary-shaped beams in the generalized Lorenz-Mie theory. Our focus is on a specific class of non-diffracting beams called discrete frozen waves, which are constructed from superpositions of Bessel beams and can be designed to provide virtually any longitudinal intensity pattern of interest. Recently, the LA was applied to frozen waves allowing, for the first time, for the analysis of light scattering problems from spherical scatterers and the subsequent determination of the physical/optical quantities in optical trapping. Since the LA cannot be rigorously applied to Bessel beams, it is of interest to determine whether those results and predictions previously established in the literature are reliable or not. To do so, we rely on exact descriptions of frozen waves and establish the limits to the validity of such an approximation scheme. It is revealed that, although serious doubts can be raised against its use, specially due to cumulative errors, the LA is much more robust than previously thought, and it may serve well to both paraxial and non-paraxial Frozen Waves, under certain circumstances.

Parallel bio-inspired algorithms in computational electromagnetics applications

This work presents the integration of parallel bio-inspired algorithms and numerical techniques to optimize electromagnetic devices for microwave and photonics applications. The results were focused on parallel computing impact in optimization efficiency and runtime simulation analyses.