Analytical Modelling of Spectral Response of Crystalline Structure Using Machine Learning

We present a new approach to design photonic crystal based optical filters using machine learning based mathematical model. The presented optical filter device finds its application in near infrared spectral range. The design and spectral response of the filter can be predicted using the proposed ma

Project Title

Analytical Modelling of Spectral Response of Crystalline Structure Using Machine Learning

Project Area of Specialization

Artificial Intelligence

Project Summary

We present a new approach to design photonic crystal based optical filters using machine learning based mathematical model. The presented optical filter device finds its application in near infrared spectral range. The design and spectral response of the filter can be predicted using the proposed mathematical model which can considerably reduce simulation time and efforts. The numerical simulation of the optical filter device along with its spectral results and mathematical modeling are described.

Project Objectives

• The aim is to identify the spectral response of colors in the nature of crystalline structures.

• Basic crystalline structure will be numerically simulated by varying its structural properties over a wide range of values and the spectral responses will be recorded against each parameters.

• The structural properties of the crystal are the input data values gives by the user and the spectral response values are the output parameters of the system.

Project Implementation Method

The whole project will be implemented through Python language using Linux environment. The All-possible symmetric arrangements of particles in three-dimensional space may be described by the finite space groups. The Crystalline Structure can determine with the help of different properties, such as Reflective Index, Lattice Constant, Depth and Radius. In this project the first vital role is to generate the data by using MEEP. Once data is generated than this generated data train the model using Machine learning and then get the results This project works fully automate the program and train the model for prediction using Machine Learning techniques. If we train a model so there will be no need of the simulations again and again so that we can save the system resources and the time consumption.  While using such Machine Learning techniques they develop mathematical models so, at the end of program we can compare our simulated results with the that mathematical model to improve our accuracy.

Benefits of the Project

 Now a days, Machine Learning, Neural Networks and Deep learning sciences are used in different fields such as Medicines, Biochemistry, Processing of DNA, Nano technologies.

Today’s Artificial Intelligence has more requirements like speed and fast processing. So that’s way the world is moving towards the photonic crystals. In photonic crystals the information will be processed with Optical frequency, Photonic processors and Photonic Quantum Computers and many Periodic structures will be used to process Optical Frequency.  For all these processes we use special category of Photonic Crystals, working on photons which process light frequencies. So, the developers present an integrated ML approach and show how Machine networks (MNs) can streamline the design process and provide a unique, robust, time-efficient, and accurate characterization capability for complex nanostructures based on their far-field optical responses. Our potential readers are scientists who are interested in working on spectral responses of light

Technical Details of Final Deliverable

MEEP Library: While giving Radius, Lattice Constant, Refractive Index and Density as input to MEEP class. MEEP (MIT Electromagnetic Equation Propagation) can generate data against these inputs and saved them in flux0 and flux1 in csv format. MEEP is a free and open-source software package for electromagnetics simulation via the finite difference time domain method spanning a broad range of application.

Peak Detection:  Peak Detection class takes the generated csv files as input to detect peaks in the graph automatically and shows the value of wavelength. This wavelength value is recorded against each parameter and lately used for model training.

Trained Model: This class takes the generated wavelength for training the model. It can take Radius as input and predict the wavelength.

The project can be operated in many fields such as Optical Communication, Optical Transistors for optical and Quantum Computing, Optical Filters for Fiber Optic Communication, Light Filters in Light Emitting Diodes (LEDs), Lasers, TV Screens, Fluid Sensing for Impurities, Photonic Circuits etc.

• Visual studio code is the environment in which the software will operate through MiniConda.
•  LINUX is the only operating system which can support all the useful libraries of software.

Final Deliverable of the Project

Hardware System

Core Industry

IT

Other Industries

Core Technology

Artificial Intelligence(AI)

Other Technologies

Sustainable Development Goals

Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com