A bottom-up approach to the systematic design of LNAs using evolutionary optimization

Abstract

A systematic design methodology for low-noise amplifiers (LNAs) is introduced. This methodology follows a bottom-up approach that employs a multi-objective evolutionary optimization algorithm, which is used at two levels. First, it is used to generate Pareto-based performance models for integrated planar inductors. To do so, an electromagnetic simulator that takes into account the inductor's layout, thus providing highly accurate performance evaluations, is coupled to the optimizer. Unlike foundry-provided inductor libraries, these Pareto-based models offer a detailed insight of the trade-offs between inductance, quality factor and area. Afterwards the Pareto-based models for the inductors are used as design variables to generate the LNA Pareto surface, thus providing highly accurate performance trade-offs of the LNA.