基于GPU加速的空间平滑多重信号分类算法实现  

Implementation ofSpatial Smoothing MUSIC Algorithm Based on GPU Accelerated

作　　者：周邦建 徐飞[1] 王谦[1] ZHOU Bangjian;XU Fei;WANG Qian(Xi'an Electronic Engineering Research Institute,Xi'an 710100)

机构地区：[1]西安电子工程研究所,西安710100

出　　处：《火控雷达技术》2025年第1期109-112,共4页Fire Control Radar Technology

摘　　要：空间平滑MUSIC算法作为一种超分辨测角算法,可以极大的提高雷达的角度分辨能力,对解决低空目标的多径反射及密集型目标分辨探测等问题具有重要的实用价值[1]。然而该算法中涉及大量的矩阵运算,计算复杂度高,数据运算量大,在传统DSP、FPGA等硬件平台开发中存在计算资源紧张、程序设计困难等问题。本文基于CPU+GPU的异构平台,采用GPU并行处理的特点,完成对空间平滑MUSIC算法的实现,对大规模阵元数据测角处理具有较高的可靠性和实时性能力。The spatially smoothed MUSIC algorithm,as a super-resolution angle measurement algorithm,can greatly improve the angle measurement resolution capability of radar,and has important practical value in solving problems such as multipath reflection of low altitude targets and dense target resolution detection.However,this algorithm involved a large number of matrix operations,high computational complexity,and a large amount of data computation.In the design of traditional hardware platforms such as DSP and FPGA,there were problems such as tight computing resources and difficult program design.Based on a heterogeneous platform of CPU+GPU,using the characteristics of GPU parallel processing,the implementation of spatial smoothing MUSIC algorithm was achieve,which had high reliability and real-time capability for large-scale array data angle measurement processing.

关 键 词：空间平滑MUSIC GPU 工程实现 

分 类 号：TN95[电子电信—信号与信息处理]