PSO-based evolutionary optimization for black-box modeling of arbitrary shaped on-chip RF inductors

In the present work a particle swarm optimization (PSO) based off-line system identification algorithm has been proposed for modeling of RF on-chip inductors. The proposed scheme has a distinctive feature that the determination of the system structure and the identification of parameters can be simultaneously obtained. The system identification algorithm used here is based upon a black-box modeling approach. Unlike the conventional equivalent circuit models, in the proposed modeling a priori information of the physical structure of an inductor is not required. Consequently this generic algorithm is capable of modeling any arbitrary shaped inductor more accurately than the conventional equivalent circuit models