基于张量分解的子空间频率估计及应用

为了提高短时序列的频率估计精度,提出一种基于张量分解的适用于多通道接收数据的子空间高分辨率的频率估计算法。该方法将接收到的多通道采样信号按照一定的形式构建三维数据结构,利用子空间的移不变性,采用张量的高阶奇异值分解直接对数据进行频率估计。仿真表明在低信噪比的情况下,基于张量分解的频率估计算法性能要优于基于矩阵奇异值分解的频率估计算法,将此算法应用于目标轨迹测量中,可以获得高精度的目标轨迹参数。

Subspace Frequency Estimation and Application Based on Tensor Decomposition

Subspace high-resolution algorithm of frequency estimation suitable for multi-channel receiving data is proposed to improve the accuracy of short-sequence frequency estimation based on the tensor decomposition. The sampling signal received by multi-channel are structured into three-dimensional data structure according to certain form. The frequency may be estimated by using the character of subspace invariance and high-order singular value decomposition. The simulation results show that the performance of the frequency estimation algorithm based on tensor decomposition is better than the one based on matrix singular value decomposition algorithm in the case of the low SNR. The algorithm may obtain high accurate target trajectory parameters, which is applied to target trajectory measurement.