| Due to the rapid development of wireless communication, there is increasing demand for small and wideband antenna to the application of portable wireless communication devices with multi-systems. Micro-strip antennas are widely required for features such as low profile, light weight, low cost and ease of integration into circuits or arrays. However, the micro-strip antenna is limited by its narrow operating bandwidth.This thesis presents an introduction of well-known methods to increase the bandwidth of antennas, but the bandwidth and the size of an antenna are generally mutually conflicting properties, that is, improvement of one of the characteristics normally results in degradation of the other. To overcome this problem, antenna design evolved from a simple dipole antenna into antenna with all shapes and dimensions, which means there are more parameters had to be controlled.Genetic Algorithms(GA) has become an ideal way for antenna optimization on the advantages of simplicity, independence on the initial conditions and so on. This paper discusses a method which combines the advantages of both Genetic Algorithms and High Frequency Structure Simulator(HFSS). This method can optimize several parameters at a time. The mixed optimization program is investigated under the environment of Matlab. GA is applied to optimize the antenna basing on S11, which needs calling HFSS to simulate Su parameter of antennas.Furthermore, this paper proposes a novel method for optimizing slots on patch. The traditional means using GA to optimize slots on the patch is modeling the planar antenna configuration expressed by meshes with a binary string. The relationship can be simply mapping between each mesh and sub-patch. The bit "1" in the binary string represents the presence of metal, and "0" means absence. Then GA can optimize the figure of slots, but there are two bugs with this method:first, this method couldn't converge, when adjacent cells only connect at a vertex after some sub-patches being cut off. Besides, the diversification of slots considered by this method is not enough. This paper proposes a novel program without dividing the patch into meshes, but given a'mold'which serves as a basic cell. GA optimizes the location of the'mold', that is, the chromosome of each individual gives the coordinate where the'mold'will be cutoff from the patch one at a time. After one cut, the optimizer calls HFSS to simulate the antenna and gets S11 for GA to calculate the fitness. By several cuts, antenna will get an optimal slot figure. An example which enhances the antenna bandwidth by 50% is also given to confirm the effectiveness of the proposed method. Finally, this paper designs two antennas including a monopole antenna and a planar inverted F antenna (PIFA) assisted by the GA/HFSS optimizer. The monopole antenna consists of grating patches, a rectangular ground plane and a substrate with concave. The antenna is only 60×25×3.2mm3 and shows a wide operating S11≤-10dB bandwidth of 1.67-3.68GHz which covers the frequency bands of DCS1800, PCS1900, WCDMA, UMTS and WiBro/WLAN. Besides, the proposed antenna has a stable omni-directional radiation patterns in the H-plane. By using the substrate mixed by ferrite and air, as well as T-shaped ground plane, the PIFA is only 29×5×lmm3. S11≤-10dB bandwidth covers 620-820MHz. The PIFA characteristics measured by network analyzer are in accordance with the simulation result. |
PDF Full Text Request