Cuda-based Sar Simulation System Design

Since Nvidia released the Compute Unified Device Architecture (CUDA), designing the high-performance program with GPU has currently become a hotspot. On the other hand, Synthetic Aperture Radar (SAR) simulation system has also become a necessary method and tool in research and development, functional testing and performance evaluation of SAR system. Nowadays, most of SAR simulation system are time-consuming,low efficiency of implement, and high cost of simulation. SAR simulation system based on the high-performance computing power of GPU can effectively improve the facing difficulties of the system. Currently, papers of research in the field are very rare at home and overseas. Therefore, the study of SAR simulation system based on CUDA is important and meaningful.The article do some deep research about the application of SAR simulation system designed based on GPU. Implement the algorithm based on GPU including SAR raw echo signal, range compression, range cell migration correction, and azimuth compression. Designing the Code Generator based on CodeSmith and Windows Wizard can improve the efficiency of developing the simulation system. The Code Generator can generate the dependable code instantly. Design the user's interface adopting the MFC, and it makes the system easy to use. With the floating-point computing power and high-parallel execution of GPU, the system greatly improves the real-time performance, effectively reduces the cost of the simulation.Main contents of the paper are as follows:(1) Discussing the development of CUDA, the architecture of GPU supporting CUDA technology, the basic concepts in CUDA programming, and how to use CUDA for development.(2) Establishing the simulation system model based on airborne stripmap SAR. Through airborne stripmap SAR signal imaging processing, it explains the imaging theory and the Range-Doppler algorithm (RDA).(3) Implementing the algorithm of SAR simulation system, how to generate SAR raw echo signal, range compression, range cell migration correction, and azimuth compression. Expound the part of the algorithm on the GPU implementation and optimization.(4) Describing how to design the SAR simulation system platform. This part contains how to design the interactive interface based on MFC, the innovative engineering code generator, application program design of agile software development and engineering software design skills.(5) The SAR simulation system is implemented on GPU and CPU platforms respectively. Through imaging the scene whose data come from Google Earth after building model, Rendering and adopting the classical RD algorithm, it achieves the functional verification and performance analysis of SAR simulation system. Comparing GPU and CPU platforms, the results of simulation show that the SAR simulation system based on CUDA is effective and real-time likely.