| Radar simulation system is an indispensable role in the research and development of radar technology.With the increasing complexity of radar systems and increasing data processing volume,traditional serial computer simulation techniques are sometimes difficult to meet some of the stringent requirements in radar simulation systems.And people's demand for high-speed processing power brings parallel computing to this new way of computing.Parallel computing uses multiple computing units to work together to solve a complex problem,and its development provides great convenience for scientific computing.Based on the parallel computing method and its theory,this thesis analyzes some of the key algorithms in the radar simulation system and optimizes the design of the algorithm,which improves the efficiency of the simulation.The main work of this thesis is as follows:1.Two parallel processing methods based on CPU and GPU are studied.On the one hand,the parallel part of CPU mainly studies Open MP technology and the parallel principle of CPU based on Open MP,and analyzes three types of Open MP API components and Open MP memory model.On the other hand,the GPU parallel part mainly studies the parallel principle of CUDA technology,and analyzes the CUDA technology programming model and its storage structure.Then,this thesis introduces graph theory into the parallel analysis of radar simulation system,and focuses on the parallel algorithm decomposition problem based on directed acyclic graph.2.Based on the standardized modeling and simulation platform,the basic workflow and standardized modeling tools of the platform were analyzed,and a radar simulation system was designed under the platform using a simulation model.At the same time,some important algorithms in the signal processing of the system such as FFT,pulse compression,CFAR detection and so on are studied.Their basic principles are analyzed,and they are simulated on the MATLAB platform.3.According to the topological sorting knowledge of directed acyclic graphs in graph theory,a topological sorting algorithm for the simulation system is designed.According to the parallel sequences obtained,the parallel system processing of the CPU of the radar simulation system is performed using Open MP parallel technology.Then,according to the previous research on the principles of FFT,pulse compression,and CFAR detection,their parallel processing methods are analyzed.At the same time,a parallel algorithm based on CUDA is designed and each module is based on CUDA-based GPU modular parallel processing.The parallel efficiency of each module after processing was analyzed.4.To complete the overall parallel simulation experiment,the parallelized module applied to a well-designed radar simulation system.The experimental results verify that the proposed parallel computing method can effectively improve the simulation efficiency of the radar simulation system,and the parallel processing based on GPU has more obvious acceleration effect than the CPU in the large amount of data. |
PDF Full Text Request