Identification of complex Bragg gratings (Apodized and chirped) using artificial neural networks (ANN) (inverse problem and ANN)

A new method based on artificial neural networks (ANN) for solution of the inverse problem for reconstruction of the complex Bragg gratings, precisely, is proposed. The Runge-Kutta method for calculation of spectrum of the reflection coefficient based on the Riccati equation in a fiber Bragg gratings is used and the application of the multilayer perceptron neural network (MLPNN) in inverse scattering problem is considered. The training of the MLPNN is based on the back propagation algorithm. The simulated results of the complex Bragg gratings for given non-uniformity are used as training data. Finally, after training the simulated results of the output of ANN shows effectiveness of the proposed methodology. In this paper the proposed idea is examined for some examples.