Optimal distribution of RF receiver multistandard specification based on multiobjective optimization

Abstract

This article presents a refinement methodology for the global specification of RF receivers producing a detailed optimal specification of architectural components based on a genetic algorithm (GA) optimization. To demonstrate the reliability of our methodology, our approach has been applied to both the IEEE 802.15.4 and DVB‐SH standards for optimizing system constraints such as noise figure (NF) and total third order distortion. This article presents the generation of optimal and feasible solutions as a positive feature of the applied nonlinear multiobjective optimization procedure. Moreover, an algorithm is also proposed to obtain, from one NF of the possible solutions, the feasible design parameters of an inductive source degenerated cascode low noise amplifier (LNA) loaded with a tunable LC tank. By using the values generated by the proposed algorithm, this tunable LNA meets the requirements of both aforementioned standards. An LNA demonstrator was prototyped on a 150‐nm CMOS technology. The actual LNA parameters (NF, gain, and IIP3) can then be back‐annotated onto the top hierarchical instance to produce a last refinement to its remaining components, demonstrating how existing components can be reused using this design approach.