Novel design of a planar UWB antenna with enhanced gain using a mixed-parameter CMA-ES algorithm

Summary form only given. Different planar monopole antenna configurations have been used in Ultrawideband (UWB) applications. They include rectangular, triangular, elliptical, and hexagonal configurations, among others. These antennas present a number of advantages such as a low profile, ease of fabrication, low cost, and compatibility with RF circuit boards. However, they can also suffer from an unstable boresight gain which is undesirable in most applications. Squinting of the main beam occurs at higher frequencies which causes a degradation of the boresight gain and instability of the antenna radiation pattern.A wideband planar stair-like monopole antenna was previously developed [R. Zaker and A. Abdipour, IEEE Antennas Wireless Propag. Lett. 2010]. Two geometry modifications were later introduced to the original design to enhance boresight gain performance of the antenna [W. Zhang, C. Thajudeen, and A. Hoorfar, IEEE AP-S Symp., 2012]. The modifications involved the addition of a gate-like structure to the ground plane of the antenna to reduce the squinting effects at higher frequencies, and the introduction of a stair-like parasitic patch which served to further enhance both bandwidth and boresight gain. In this work, the stair-like monopole antenna is further optimized and its geometry is significantly modified by introducing slots into its surface, in order to achieve enhanced bandwidth and boresight gain. A mixed-parameter Covariance Matrix Adaptation Evolution Strategy (CMA-ES) algorithm [E. Bou Daher and A. Hoorfar. IEEE AP-S Symp., 2013] was applied to perform this optimal design process. CMA-ES is one of the most efficient nature-based global optimization techniques and has been shown to provide a better performance than GAs, EP, ES, and PSO in many complex engineering problems. CMA-ES with mixed parameters is used when both continuous and discrete parameters are involved. In this work, the dimensions of the antenna are continuous parameters. The main pa- ch is then divided into a large number of rectangular segments where each segment is associated with a discrete (binary) variable. The discrete parameters are then used to determine whether each segment should be present in the patch. The parasitic patch remained a scaled copy of the main one. The geometry of the optimized design is shown in the figure at the right. Preliminary results are promising and show an improved boresight gain performance of better than 0.3 dBi over a bandwidth of 1-10 GHz. The simulated and measured results will be given in the presentation.