Content area
Full Text
1. Introduction
One of the primary objectives in the design of compact wireless communication systems is the multi-band operation. In this regard, the Wilkinson power divider (WPD) is one of the fundamental microwave passive components that has been widely investigated in the literature (Jwaied et al., 2007; Younesiraad et al., 2017; Alshamaileh et al., 2015; Chen et al., 2013; Al Shamaileh et al., 2011). Multi-band WPDs depend on replacing the conventional quarter-wave section at each arm of the divider by multi-section transmission line transformer (TLT), which has been widely reported in the literature (Khodier et al., 2008). A dual-band WPD was proposed in Wu et al. (2006), which consisted of two-section TLT with a parallel resistor-inductor-capacitor (RLC) combination connecting the output ports. A tri-band WPD using three isolation resistors, instead of the RLC combination, is presented in Khodier et al. (2008) and Chongcheawchamnan et al. (2006). Numerical optimization was used to find the divider parameters.
In Dib and Khodier (2008), an equal-power split multi-band WPD, similar to that in Chongcheawchamnan et al. (2006), is designed by using particle swarm optimization (PSO) method. It is a symmetrical WPD that achieves equal-power split at N arbitrary frequencies. Each quarter-wave branch in the conventional Wilkinson divider is replaced by N sections of transmission lines, and the isolation between the output ports is achieved by using N resistors. The design parameters are the characteristic impedances and lengths of the N transmission line sections, and the N isolation resistors. The even–odd mode analysis is used to derive the required design equations. Closed-form expressions are derived for the dual-band divider. For N ≥ 3, closed-form expressions are not available, and therefore, the PSO method was used to obtain the design parameters. However, the associated time-consuming procedure and computational burden in realizing WPD through PSO optimization are major disadvantages; needless to mention the substantial increase in optimization time because of the multi-band design.
Neural networks (NNs), in this context, are one of the best candidates in addressing the abovementioned challenges, owing to their ability to process the interrelation between electrical and geometrical/physical characteristics of the WPD in a superfast manner. The basis of NN modeling is to capture the inherent nonlinear input–output functional relationship, and model...