基于GPU的宽带信号干涉仪测向算法实现  

Implementation of broadband signal interferometer direction-finding algorithm based on GPU

作　　者：苗峻 窦修全[1,2] 郭海召 MIAO Jun;DOU Xiu-quan;GUO Hai-zhao(The 54th Research Institute of CETC,Shijiazhuang 050081,China;Hebei Key Laboratory of Electromagnetic Spectrum Cognition and Control,Shijiazhuang 050081,China)

机构地区：[1]中国电子科技集团公司第五十四研究所,石家庄050081 [2]河北省电磁频谱认知与管控重点实验室,石家庄050081

出　　处：《信息技术》2023年第8期112-118,124,共8页Information Technology

摘　　要：在应用干涉仪算法测向时,需要对多个接收信号进行逐一处理,而且测向精度往往与本地样本库数据量正相关,因此当需对大量信号进行高精度测向时,常规的计算平台难以保证算法的实时性。针对以上问题,提出一种优化的干涉仪算法,能够在降低样本库数据量的同时保持较高的测向精度,并探索了一种基于GPU并行计算的算法实现方式。经计算机仿真验证,基于GPU实现的优化算法,可将执行速度提升至CPU平台的十余倍。When the interferometer algorithm is used to carry out direction-finding,the interferometer algorithm need to process multiple signals one by one and its accuracy of direction-finding is always positively correlated with the amount of local calibration data.Therefore,when high-precision direction-finding of a large number of signals is required,the normal computing platforms are difficult to ensure the real-time of the algorithm.Aiming at providing a solution for those circumstances,we propose a optimized method of interference direction-finding algorithm,which could obtain higher direction accuracy while reducing the amount of calculation.Besides,a parallel computing implementation for interferometer algorithm based on GPU has also been explored in this article,which has been proved through computer simulation experiments that the optimized algorithm based on GPU could increase the execution speed to more than ten times compared with CPU.

关 键 词：信号处理 宽带测向 干涉仪 图形处理器 并行计算 

分 类 号：TN911.7[电子电信—通信与信息系统]