基于矩阵束算法的极点提取分析

Analysis on Pole Extraction based on Matrix Pencil Algorithm

目标极点特征具有目标姿态不敏感性以及明晰的物理意义,是雷达目标识别研究的重点和热点方向之一。为了实现目标极点的准确提取,提出了目标极点数目已知及未知情况下,奇异值分解参数的确定方法,并以实例验证了方法的有效性;通过对不同信噪比的噪声信号进行极点、留数提取及误差分析,探究了矩阵束算法抑制噪声的能力。仿真结果表明,矩阵束算法对信噪比为10 dB的噪声信号进行极点提取时仍具有较高的精确度。

The pole characteristic of a target, with insensibility to target gesture and clear physical meaning, is now the focus and hotspot in the study of radar target recognition. An algorithm for determining the parameter of singular value decomposition is proposed, thus to achieve accurate pole characteristic of a target under the situations that the number of pole is known or unknown. The experimental example proves the validity of this algorithm. The noise suppression ability of matrix pencil algorithm is studied through the poles and residues as well as the numerical error extracted from noise signal with different signal-to-noise ratios. Simulation results show that the matrix pencil algorithm is still of a fairly high accuracy in pole extraction of noise signal with signal-to-noise ratio as below as 10 dB.