Automated Power Amplifier Design Through Multiobjective Bottom-Up and Particle Swarm Optimizations Using Neural Network

Abstract

This study presents an automated power amplifier (PA) design process by optimizing the topology and values of design parameters, sequentially. The automated optimization environment is created with the combination of an electronic design automation tool and a numerical analyzer. As a first step, the configuration of the PA is generated using the bottom-up optimization (BUO) method, then the values of the components are optimized using the particle swarm optimization (PSO) algorithm that is employed with a shallow neural network. The PSO method is applied for optimizing PA in terms of output power, power gain, and efficiency leading to obtain optimal design parameters. The proposed optimization process is automatic and compact leading to reduce interruptions of designers during the process. In order to verify the effectiveness of the presented method, one lumped element PA including GaN HEMT transistor is designed and optimized. The optimized PA reveals higher than 45% power added efficiency with the linear gain performance between 10 ÷ 14.6 dB in the frequency band of 1 ÷ 2.3 GHz.