Design And Implementation Of A Real-time Sorting System For Large-scale Distributed Full-pulse Radar Signals

In recent years,from being used only for military purposes,such as electronic warfare,to civilian use,such as intelligent traffic speed measurement,the use of radar is becoming more and more widespread.However,with the wide application of radar,the radar electromagnetic signal environment has become more and more complex.The number of radar source increases dramatically,signal density becomes higher,and signal becomes more complex,resulting in the traditional radar pulse signal sorting method being more difficult to meet the actual application requirements in the real time and throughput.Aiming at the dense radar pulse signal scene,we draws on the popular batch processing technology and streaming processing technology in big data processing parallel computing and the data clustering algorithm in data processing method,designing and implementing a system suitable for large-scale data processing scene,and providing technical support for the large-scale radar pulse signal sorting scenario.The main work of this thesis is as follows:1)Due to the increasing intensity of the radar pulse signal,which leads to the problem of real-time sorting and throughput getting harder and harder to meet the requirements.This paper draws on the design ideas of storm and flink which are the popular streaming computing framework,designed and implemented a lightweight parallel computing framework in the C ++ programming language.The core of the computing framework draws on the thinking of the sliding window and the adjustable buffer,the user can build business operations through the system programming interface,and run in this framework.By adjusting the size of the sliding window and the adjustable buffer,users can get the balance between real-time and throughput requirements.2)Due to the increasing complexity of the radar pulse signal,which leads to the problem of the traditional PRI-based radar signal sorting algorithm being difficult to meet the sorting effect,this paper draws on the existing density grid clustering algorithm,and parallel computing batch processing mode,designed and proposed a distributed density grid clustering algorithm.The core of the algorithm is to realize the model of the flow clustering algorithm by sliding the window.For intensive radar pulse signals,the streaming clustering model can adapt to the micro-batch calculation mode,which makes the sorting result get better real-time performance.In addition,through the segmentation of data volume and data space dimension,the algorithm achieves a good parallel computing effect,which makes the sorting result obtain greater throughput.Combined with the above lightweight parallel computing framework and distributed density grid clustering algorithm,this paper designs and implements a radar signal sorting system applied to large-scale data scenarios.The test results show that the large-scale radar pulse signal sorting system designed in this paper has no loss in radar pulse sorting accuracy compared with stand-alone algorithm,but the delay of sorting can reach the millisecond level,and the throughput also obviously is due to the existing single radar signal sorting algorithm.