DOI:10.11991/yykj.202302006网络出版地址:https://link.cnki.net/urlid/23.1191.U.20231214.1136.002基于矩阵重构和酉变换的DOA估计李齐衡,孙溶辰,孙志国哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001摘要:为了提高重构相干信号测向算法的估计性能,降低算法运算量,提出了一种基于矩阵重构和酉变换方法的酉矩阵重构算法。该算法首先通过酉变换将阵列接收数据从复值计算转换为实值计算,使计算量大大降低;然后计算阵列协方差矩阵并进行特征值分解得到信号子空间,再将信号子空间重构为Toeplitz矩阵实现解相干并再次进行酉变换;最后通过特征值分解得到信号子空间并使用最小二乘法实现波达方向(directionofarrival,DOA)估计。相比于改进的旋转不变性的信号参数(estimationofsignalparametersviarotationalinvariancetechniques-like,ESPRIT-Like)算法和空间平滑处理算法,由于消除了噪声影响、构造了Toeplitz矩阵以及充分利用了数据的共轭信息,该算法的估计精度更高、具有更高的运算效率且在ESPRIT-Like算法失效的条件下新算法仍能有效估计DOA。本文算法的运行时间是ESPRIT-Like算法的71.2%,实验结果证明了该方法的有效性和真实性。关键词:Toeplitz矩阵;解相干;酉变换;波达方向估计;特征值分解;空间平滑;信号子空间;ESPRIT-Like中图分类号:TN911.72文献标志码:A文章编号:1009−671X(2024)01−0120−10DOAestimationbasedonmatrixreconstructionandunitarytransformationLIQiheng,SUNRongchen,SUNZhiguoCollegeofInformationandCommunicationEngineering,HarbinEngineeringUniversity,Harbin150001,ChinaAbstract:Toimprovetheestimationperformanceofcoherentsignaldirectionfindingalgorithmandreducethecomputationalcomplexity,aunitarymatrixreconstructionalgorithmbasedonmatrixreconstructionandunitarytransformationisproposed.First,thealgorithmconvertsthereceiveddataofthearrayfromthecomplexdomaintotherealdomainbyunitarytransformation,whichgreatlyreducestheamountofcalculation,thenthecovariancematrixofthearrayiscalculatedandtheeigenvaluedecompositionisperformedtoobtainthesignalsubspace,andthenthesignalsubspaceisreconstructedintoaToeplitzmatrixtoachievedecoherenceandperformunitarytransformationagain.Finally,thesignalsubspaceisobtainedbyeigenvaluedecompositionandthedire...