Synthetic Aperture Radar On The Graphics Chip To Generate A Simulation Study

For many years, with the advancement of Graphics Processing Unit (GPU) and the creation of its new feature of programmability, GPU for general purpose computations (GPGPU) become a hot topic for research. There are two reasons for using GPGPU. The first reason is performance. The floating-point calculations run much faster on GPU than on CPU, and the development of GPU is also faster than GPU. The second reason is load balance. If CPU restricts the performance of the application and GPU has an idle cycle, we can let GPU do the job to make the application run faster.Right now, many rearch have been done in GPGPU in other countries, but domestic research in this area was relatively small. This dissertation explores the applications of GPU in the field of signal processing based on the research of SAR (Synthetic Aperture Radar) Raw Data Generate on GPU. The computing complexity of SAR Raw Data Generate is high. CPU needs much time to compute, and GPU can improve the performance.The main jobs of this dissertation are as follows:1. Use improved DIF (Decimation in Frequency) algorithm to perform FFT (Fast Fourier Transform) on GPU and compare the performance of FFT on GPU to FFT on CPU.2. Implement SAR Raw Data Generate on GPU using the FFT based RTPC (Range Time Domain Pulse Coherence) algorithm and compare the results and performance of SAR Raw Data Generate on GPU to that on CPU. Discuss the advantages and disadvantages of the two methods.The experiment results are as follows:1. CPU has better performance in computing 1D-FFT and 2D-FFT when the quantity of input data is low. But when the quantity of input data grows bigger GPU has better performance.2. When applied in SAR point target, GPU could aquire effective result but poorer time performance. On the other hand, when applied in SAR surface target, GPU could aquire effective result and much better time performance.