基于GPU的软件雷达信号处理  

Signal Processing of Software Radar Based on GPU

作　　者：田乾元 徐朝阳[1] 赵泉[1] TIAN Qian-yuan;XU Chao-yang;ZHAO Quan(The 723 Institute of CSIC,Yangzhou 225101,China)

机构地区：[1]中国船舶重工集团公司第七二三研究所,江苏扬州225101

出　　处：《舰船电子对抗》2020年第1期58-63,100,共7页Shipboard Electronic Countermeasure

摘　　要：在基于图形处理器(GPU)的软件雷达终端架构下,采用统一计算设备架构(CUDA)技术,实现了常规的雷达信号处理,并与基于CPU平台的雷达信号处理过程进行对比。结果表明:在实现脉冲压缩、动目标显示(MTI)、动目标检测(MTD)、恒虚警率(CFAR)处理上,2种方法的最大相对误差在5×10^-4以下,平均计算速率比CPU平台提升了100倍以上,但这远没有达到GPU的计算极限。对于实际情况来说,随着回波采样点等数据量的增加,信号处理的吞吐量也会增大,更好地满足了雷达信号处理的大吞吐量和高实时性要求。In the software radar terminal architecture based on graphics processing unit(GPU),this paper adopts compute unified dvice architecture(CUDA)technology to realize conventional radar signal processing,and compares it with the radar signal processing based on CPU.The result shows that the maximum relative error of the two methods is under 5×10^-4 in realizing pulse compression,moving target indication(MTI),moving target detection(MTD)and constant false alarm rate(CFAR)processing,and the average calculation velocity is increased by more than 100 times than that of CPU,but it is far from reaching the calculation limit of GPU.In practical application,the throughput of signal processing increases with the increase of data amount(such as sampling points of echo),which better meets the real-time and high throughput requirements of radar signal processing.

关 键 词：图形处理器 软件雷达 信号处理 

分 类 号：TN957.52[电子电信—信号与信息处理]