一种GTD模型参数估计的改进2D-TLS-ESPRIT算法

针对经典二维总体最小二乘法旋转不变子空间（2D-TLS-ESPRIT）算法估计二维几何绕射理论（GTD）模型参数精度不高、抗噪性能较差这一问题，提出了一种改进2D-TLS-ESPRIT算法。首先，改进算法通过将目标的极化散射矩阵加入到二维GTD散射中心模型，使得模型对目标极化散射特征的描述更加精准；其次，构建置换矩阵得到原始回波矩阵的共轭矩阵，并将两者结合起来，从而延长了目标电磁散射数据的长度；最后，仿真结果验证了改进算法的参数估计性能与噪声鲁棒性均要优于同类已有算法，雷达散射截面积（RCS）外推结果进一步验证了改进算法参数估计性能的先进性。 

An improved 2D-TLS-ESPRIT algorithm of GTD model parameter estimation

The parameter estimation and noise robustness ability of classical Two-Dimensional Total-Least-Square Estimating Signal Parameter via Rotational Invariance Techniques (2D-TLS-ESPRIT) algorithm are not effective when extracting parameters of the two-dimensional Geometric Theory of Diffraction (GTD) model. To solve this problem, an improved 2D-TLS-ESPRIT algorithm is proposed in this paper. Firstly, polarization scattering matrix is added into the two-dimensional GTD model and hence the full-polarization scattering center model can be obtained. Secondly, the covariance matrix of the original echo matrix can be achieved by constructing a permutation matrix. The length of electromagnetic scattering data can be added by combing these two matrices. Finally, the simulation results prove that parameter estimation performance and noise robustness ability of the improved algorithm are better than those of the same kind of algorithms. The Radar Cross Section (RCS) extrapolation results also validate the superiority of the improved algorithm in parameter estimation performance.