Study On Multi-mode SAR Echo Fast Simulation Method Under CPU+GPU Mixed Architecture

Due to the high manufacturing cost of the actual synthetic aperture radar,the development cycle is long,in order to ensure that the radar enters space,it can be accurately observed to observe the ground or sea scene target,and the echo signal simulation simulation system is required to reduce the research risk and cost of the actual system.However,with the increase in the domain width of the scene area observed by the radar system,the calculation amount of the simulation of the emulation system simulated echo signal has also increased significantly,and the existing echo simulation method simulates the scene returns to consume a lot of time,so how to implement the spaceborne SAR scene Fast simulation of echo signals is particularly important.This thesis combines the specific working model of the star-loaded multi-mode SAR,based on the various error factors,the full-link star-loaded SAR echo signal simulation model is established,and the rapid simulation method of the scene echo signal is studied,and will be improved.The post-rapid algorithm transplant the CUDA platform,complete the algorithm of the algorithm under the CPU + GPU hybrid architecture,and further improve the simulation rate.The main research content is as follows:Firstly,the system modeling of the full-link spaceborne SAR echo simulation is carried out.The satellite-to-Earth geometric observation model and satellite orbit model are established,and the definitions of six spatial coordinate systems involved in the model establishment and their conversion relationships are given.Then,based on the finite element division,the observed ground scene is divided into a limited point target,determine the position coordinates of each scattered point,and introduces a typical object scattering model.On this basis,the actual operation of the radar system is further combined,and the full-link echo signal model under the factors such as errors,antenna direction map points,and propagation errors are given.Secondly,on the basis of comparative analysis of the applicability of typical echo simulation methods,a fast simulation method of spaceborne SAR echo with a high-order expansion of the slope distance history relative to the reference point is studied.Although the traditional time-domain pulse-by-pulse simulation method is accurate,it is computationally intensive and time-consuming.The 2D frequency domain method is a fast simulation method when the antenna beam is not scanned,such as the stripe mode,and the pattern scanning in the sliding beam mode and the TOPS mode can cause this method to fail.In this case,the time-frequency hybrid method is usually used to simulate the SAR echo,but the amount of calculation is still very large.Therefore,according to the working mode of spaceborne SAR beam scanning,this thesis proposes a fast echo simulation method from the perspective of the high-order development of reference point slope distance history by taking advantage of the fact that the relative positions of each scattering point in the scene remain unchanged.This method uses the slant distance history of the center reference point of the divided scene and the compensation function obtained by the high-order expansion to obtain the slant distance of each scattering point in the scene block,thereby reducing the calculation amount of the slant distance and effectively improving the echo simulation.speed.Then,by simulating several sets of scene echo signals of different sizes,the experimental results show that the simulation rate of the proposed method is improved compared with the original method.Finally,the spaceborne SAR echo simulation method is transplanted to the CUDA platform to realize the parallel simulation under the CPU+GPU hybrid architecture.First,the GPU hardware and CUDA software programming architecture are briefly introduced.Secondly,the parallelization analysis of the echo fast simulation method is carried out,and the overall parallelization scheme flow under the CPU+GPU hybrid architecture is given by designing the kernel function call and memory allocation.According to the implementation steps of the simulation scheme,several sets of simulation experiments are carried out to compare the simulation results of the original method and the proposed method under the two platforms of multi-core CPU and CPU+GPU parallel.and consumption.