宽带高分辨DOA估计的CPU-CUDA协同实现  

A Cooperative Implementation of CPU-CUDA for Wideband High Resolution DOA Estimation

作　　者：李学敏 黄海宁[1] 李宇[1] 叶青华[1] 张扬帆[1] 韦琳哲 LI Xuemin;HUANG Haining;LI Yu;YE Qinghua;ZHANG Yangfan;WEI Linzhe(Key Laboratory of Science and Technology on Advanced Underwater Acoustic Signal Processing,Institute of Acoustic,Chinese Academy of Sciences,Beijing,100190,China;University of Chinese Academy of Sciences,Beijing,100190,China)

机构地区：[1]中国科学院声学研究所先进水下信息技术重点实验室,北京100190 [2]中国科学院大学,北京100190

出　　处：《网络新媒体技术》2020年第5期9-14,共6页Network New Media Technology

基　　金：国家自然基金(11504402)。

摘　　要：针对DOA估计中高分辨方位谱估计算法需要进行多次矩阵求逆运算,整个计算过程时间复杂度高这一问题,采用CPU和CUDA协同并行化的方式提高运算效率,实现了2种实时高分辨宽带信号处理系统。仿真实验结果表明,该协同架构的信号处理系统,能够大幅提升处理速度,满足阵列信号处理的实时性需求,同时克服了DSP阵列信号处理系统开发效率低、成本高和可移植性差的缺点。Two real-time high-resolution signal processing systems based on CPU and CUDA synergy architecture are realized in this paper,aiming at solving the problem that the high time complexity of inverse of correlation matrix in the high-resolution DOA estimation.Computer simulations indicate that,it can greatly increase the processing speed,meet the real-time requirements of array signal processing,and overcome that disadvantages of low development efficiency,high cost and poor portability of the DSP array signal processing system.

关 键 词：高分辨 宽带信号处理 实时系统 GPU 

分 类 号：TN929.3[电子电信—通信与信息系统] TP332[电子电信—信息与通信工程]