教師資料查詢 | 類別: 研究報告 | 教師: 丘建青 Chiu Chien-ching (瀏覽個人網頁)

標題:演化計算於微波成像之應用
學年99
學期1
出版(發表)日期2010/10/01
作品名稱演化計算於微波成像之應用
作品名稱(其他語言)Evolutionary Computing in Microwave Imaging
著者丘建青
單位淡江大學電機工程學系
描述計畫編號NSC98-2221-E032-009;研究期間200908~201007;研究經費673,000
委託單位行政院國家科學委員會
摘要本計劃擬研究利用演化計算的技巧來解決微波成像問題。我們將以各種的不同演算法應用於逆散射,如差異形演化策略法則、粒子群聚最佳化法則與基因法則,分別在時域與頻域探討重建自由空間物體與掩埋物體的影像。 第一年擬於時域與頻域中使用差異形演化策略法則(Differential Evolution)模擬研究二維物體的電磁影像重建問題。吾人擬將分別探討於時域中以時域有限差分法(Finite Difference Time Domain, FDTD)與頻域上利用等效源法(Equivalent Source)、格林函數(Green’s Function)配合動差法(Method of Moment)求解自由空間中的二維完全導體與介電質物體、散射場,並將以上所使用之電磁散射理論配合差異形演化策略法則進行逆推散射物體形狀與電磁特性探討,進而找出適當的差異形演化策略法則改良方法藉以增加收斂速度與全域最佳化能力。 第二年擬模擬研究掩埋物體的電磁成像。吾人擬考慮模擬物體為一個未知物體存在空間中,以 TM極化波入射,利用在不同介質的邊界條件,可以導出兩個積分方程式,利用差分化可以化為矩陣形式,再經由簡單的矩陣運算,就可以克服積分運算上的困擾,進而重建掩埋物體物理特性。在理論部份,主要是根據邊界條件導出兩組線性的積分方程組,接著利用以動差法計算其散射場,再利用粒子群聚最佳化法則(Particle 表 C011 第 2 頁,共 2 頁 Swarm Optimization, PSO),重建出物體的形狀及介電常數,並改良演算法增加其強健性與搜尋能力。 第三年擬研究各種全域最佳化法與區域搜尋法應用在微波成像問題上。吾人於非均勻介電物體與完全導體周圍適當安排不同位置的天線發射極短脈衝波與極化波並分別量測非均勻介電物質周圍之時域與頻域散射場,經由適當的處理以反求物體的物理特性。吾人將利用接收到的散射場及適當的邊界條件,導出一組非線性微分方程式,將電磁成像問題化為一求極小值的最佳化問題,然後引入全域最佳化法與區域搜尋法將逆散射問題轉化為求解最佳化的問題。藉以重建物體的位置、形狀和介電常數分佈。最後將電磁成像所得結果與原先假設者比較,藉以驗證並改進電磁成像理論,進而做出針對不同空間中的物體,利用不同演算法的重建效果何者為佳的嘗試性結論。 In this project, evolutionary computing techniques are used to solve the microwave imaging problems. Many algorithms, such as differential evolution algorithm, particle swarm optimization algorithm, genetic algorithm are investigated in time domain and frequency domain, respectively. In the first year, a powerful optimal algorithm for solving the inverse scattering problem of two-dimensional homogeneous dielectric objects and perfectly conducting cylinders will be investigated. Scattered fields of objects are calculated in time domain and frequency domain by the Finite Difference Time Domain method (FDTD) and Method of Moment (MoM), respectively. Electromagnetic scattering data combining with the differential evolution algorithm are employed to reconstruct the unknown electrical property, permittivity distribution, shape and position of the scatterer. Properly searching parameters and updating methods are discussed to increase the speed of convergence and the ability for global optimization. In the second year, dielectric cylinders of unknown permittivities and perfectly conducting cylinders are buried in one half-space and scatter the field incident from another half-space where the scattered fields are measured. Based on the boundary condition and the incident field, a set of nonlinear surface integral equation is derived. Moment methods and particle swarm optimization algorithm are used to solve a set of linear integral equations. By using 表 C011 第 2 頁,共 2 頁 received scattered fields, both the shape and the dielectric constant of homogeneous dielectric object are reconstructed. In addition, we will improve the robustness and the ability for global optimization. In the third year, global and local optimal algorithms are investigated in microwave imaging problems. Objects in different kind of space are computed in time and frequency domain, respectively. Based on the boundary condition and the measured scattered field, a set of nonlinear integral equations is derived and the imaging problem is reformulated into an optimization problem. The evolution algorithms are then employed to find out the global extreme solution of the objective function. Different algorithms are used to reconstruct the images for objects in different kind of space. The reconstructed images are compared with different algorithms and the best algorithm is chosen for the corresponding space. Moreover, we would like to make some experimental conclusions for inverse problems. We will also focus on how to increase convergence speed of the evolution algorithms.
關鍵字微波成像; 演化計算; 完全導體; 均勻介電質物體; 非均勻介電質物體; 半空間; 粒子群聚最佳化法則; 差異形演化策略法則; 基因法則; 全域搜尋法; 區域搜尋法; 時域有限差分法; Microwave Imaging; Perfectly Conductor; homogeneous Dielectric Object; Inhomogeneous Dielectric Object; Particle Swarm Optimization; Genetic algorithm; Differential Evolution algorithm; Global optimal algorithm; Local optimal algorithm; Finite difference Time Domain method; Inverse Scattering
語言中文
相關連結
Google+ 推薦功能,讓全世界都能看到您的推薦!