High Precision and Realtime Sensing of Underground Target by Ultrafast Analysis on GPU Cluster

2014 Fiscal Year Final Research Report

• Project/Area Number: 24510259
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Natural disaster science
• Research Institution: Sendai National College of Technology
• Principal Investigator: SONODA Jun 仙台高等専門学校, 知能エレクトロニクス工学科, 教授 (30290696)
• Project Period (FY): 2012-04-01 – 2015-03-31
• Keywords: 地中レーダ / 地下変位検出 / 不均質大地 / FDTD法 / GPUクラスタ

Outline of Final Research Achievements

In this study, we have analyzed detection characteristics of a subsurface displacement by ground penetrating　radar (GPR) using the finite-difference time-domain (FDTD) simulation with graphic processing units　(GPUs) cluster. Inhomogeneous subsurface with many randomly placed scatters which are different in size and　electric permittivity have been modeled, we have computed detection characteristics of subsurface displacement in　the inhomogeneous subsurface models using the FDTD method. We have estimated the detection of subsurface　displacement by cross correlation value between homogeneous and inhomogeneous GPR profiles. Moreover, we have　shown that the use of GPU cluster is much more faster and practical than conventional method.

Free Research Field

電磁波工学