LNA Design Optimization Using DNA Computing

Abstract

Abstract A noise model for heterojunction transistors using a new technique for prediction was introduced using a neural network model, and was applied to get higher accuracy for transistor noise parameters. The new model is employed in designing of a wideband Low-Noise Amplifier (LNA), which resulted higher accuracy for the four noise parameters required, using only one neural network for simulation of noise figure parameters. The accuracy of this model has been demonstrated by coordinating anticipated and estimated values of heterojunction transistors for a specific data set of noise parameters at various frequencies, temperatures and bias points. DNA computing was used to design a Low-Noise Amplifier (LNA). The DNA computing method demonstrates good and very accurate results and also shows a very high accurate results in prediction of the noise parameters by using it as FFNN to determine a threshold level value, which consequently increased the gain leading to higher bandwidth. Comparison of the new method (DNANN) to other classical optimization techniques shows that the DNA computing method results in optimized noise parameters, which consequently leads to higher LNA gain which consequently leads to improved bandwidth. Copyright © 2017 Penerbit Akademia Baru - All rights reserved.