基于高阶累积量的酉变换DOA估计算法  

Unitary Transformation DOA Estimation Algorithm Based on High - order Cumulant

作　　者：韩勇[1] 乔晓林[1] 金铭[1] 曹丙霞[1] 

机构地区：[1]哈尔滨工业大学,黑龙江哈尔滨150001

出　　处：《计算机仿真》2010年第11期336-339,共4页Computer Simulation

摘　　要：目前将酉变换算法用于奇数阵元均匀圆阵时需要利用模式空间算法进行预处理,而模式空间算法需要阵元数过多,且与酉变换算法结合后只能针对特定俯仰角在方位向上进行一维搜索。针对问题提出利用高阶累积量算法对阵列接收数据进行预处理,为提高测向精度和估角性能,利用酉变换算法将预处理得到的高阶累积量矩阵实数化。与模式空间算法相比上述方法降低了对阵元个数的需要,并且能够直接进行二维谱峰搜索。与经典多重信号分类算法(MUSIC)相比,在不增加特征值分解及谱峰搜索计算量的情况下,提高了算法的分辨力。按照给出的算法原理,通过仿真试验验证了所得的结论的正确性。At present,pre-processing should be done by using the pattern space algorithm when the unitary transformation is applied for the odd array element of uniform circular array.However,too many elements are needed in the pattern space algorithm.And only 1-dimensional search can be operated according to special elevation angle under the combination with unitary transformation.Aiming at the problem,the high-order cumulant algorithm is pro-posed to do preprocessing based on the array receiving data.And then the high-order cumulant matrix is realvalued using the unitary transformation algorithm.Compared with the pattern space algorithm,this method can decrease the demand to element number and can do 2-dimensional spectrum peak search.Compared with the classical MUSIC algorithm,the resolution is improved without increasing computation cost of eigenvalue-decomposition and spectrum peak search.The algorithm principle is given and the conclusion is verified using simulation experiments.

关 键 词：波达角估计 圆阵 高阶累积量 酉变换 

分 类 号：TN911.7[电子电信—通信与信息系统]