Wavelength prediction of laser incident on amorphous silicon detector by neural network

In this paper we present a method based on artificial neural networks (ANN) and the use of only one amorphous semiconductor detector to predict the wavelength of incident laser. Amorphous semiconductors and especially amorphous hydrogenated silicon, a-Si:H, are now widely used in many electronic devices, such as solar cells, many types of position sensitive detectors and X-ray imagers for medical applications. In order to study the electrical properties and detection characteristics of thin films of a-Si:H, n–i–p structures have been simulated by SILVACO software. The basic electronic properties of most of the materials used are known, but device modeling depends on a large number of parameters that are not all well known. In addition, the relationship between the shape of the induced anode current and the wavelength of the incident laser leads to complicated calculations. Soft data-based computational methods can model multidimensional non-linear processes and represent the complex input–output relation between the form of the output signal and the wavelength of incident laser.