Study On The Raw Data Simulation Of The Arbitrary Trajectory Sar Based On GPU

Synthetic Aperture Radar(SAR)is an all-weather,all-day,long-range,high-resolution detection equipment,which is widely used in the defense and civilian areas.As an important part of SAR system demonstration,SAR echo simulation plays an important role in SAR system design,imaging algorithm verification,and system robustness test.When the largescene and high-resolution raw data simulation is needed,the calculation amount of the whole echo simulation process is especially huge,so that the time cost greatly increases,and the computing efficiency becomes an urgent problem to be solved.In recent decades,GPU highperformance computing has been developed rapidly,and its processing core and powerful floating point computing enable researchers to gain a way to improve the efficiency of scientific computing.A large number of experts and scholars have tried to apply the GPU to non-graphics and have achieved some success.Moreover,with the improvement of GPU hardware performance and the programmability of architecture platform CUDA,it is an inevitable trend to strengthen SAR system research on GPU-based general calculation in SAR research field.The available SAR system generally works in a variety of modes,and the trajectory of radar platform is uncertain.Thus,with the need to simulate the specific echo under different conditions,one often needs to re-design a set of specific parameters of the SAR raw data simulation system,which can greatly reduce the efficiency.And non-common simulator makes the work without inheritance.This also will bring some problems into the SAR raw data simulation.Therefore,this paper aim to present the study on the raw data simulation of the arbitrary trajectory SAR based on GPU.The main works are listed as follows:1.Based on the SAR basic theory,and the original SAR raw data simulation algorithm,a raw data simulation algorithm based on an improved concentric circular is proposed to deal with the problem of low SNR of imaging results due to the low precision of the conventional concentric circle algorithm.The improved algorithm take into account the influence of range sidelobe and azimuth antenna pattern.Thus,the improved algorithm makes the raw data simulation process more realistic,and thus improves the quality of raw data simulation.2.According to the improved concentric circular raw data simulation algorithm,this paper further analyzes the echo model,and proposes a general SAR raw data simulator,which is suitable for SAR echo simulation with arbitrary trajectory and arbitrary configuration.From system design point of view,we define the standard interface for simulator,so that the development of raw data simulation software can be modular and hierarchical.The design of the standard interface makes the SAR raw data simulator not limited to the plane amplitude image,but also includes the 3D topographic map,which makes the application of the raw data simulator wider.Meanwhile,the serial-parallel of the general simulator is analyzed,and the parallel design and depth optimization of the program are carried out from the aspects of threading,memory management,stream processing and so on,which improves the efficiency of SAR raw data simulation.Finally,this paper carries out the design of the simulator graphical interface,which improves the human-computer interaction experience.3.The raw data simulation accuracy of the proposed algorithm and the acceleration performance of the simulator are verified by point and scene simulations(including 3D topographic map).The effectiveness of the proposed method is validated by comparing the experimental results of the original concentric circle method and of the improved the concentric circle method.The simulation results show that the generality of the design simulator and the acceleration performance of the simulator are verified by echo simulations under the conditions of different trajectories and different configurations.Finally,by comparing the implementation time of GPU simulation program and of CPU simulation program,it truly reflects the performance advantage of GPU platform echo simulation...