基于四阶累积量的测深侧扫声纳波达方向 估计方法研究

针对高频测深侧扫声纳工作时会受到高斯色噪声影响的问题,提出基于四阶累积量及幅相误差自校正的波达方向估计方 法。 首先,计算声纳回波信号的四阶累积量,得到协方差矩阵;其次,采用加权平均算法对协方差矩阵进行降维处理,得到构造矩 阵;再次,通过恢复构造矩阵的 Toeplitz 特性完成幅相校正,并增加相位线性拟合算法进一步提升校正效果;最后,采用求根多重信 号分类(Root-MUSIC)算法求解得到波达方向估计。 仿真结果表明,本文算法的角度估计均方误差相对于基于自校正的 RootMUSIC 算法有约 0. 07°的提升,改善了四阶累积量算法在阵元幅相误差条件下估计精度下降的问题。 海试数据分析表明,本文算 法的相对测深精度相对于基于自校正 Root-MUSIC 算法有约 0. 03% ~0. 08%的提升,且在水平距离 90 m 范围内的满足 0. 75% 的测 深精度标准。 仿真和海试数据均证明了本文算法在高斯色噪声及阵元幅相误差条件下的 DOA 估计性能优于传统算法。

Research on the DOA estimation method of bathymetric side scan sonar based on fourth-order cumulant

The high-frequency bathymetric side-scan sonar is affected by Gaussian color noise. To address this issue, a direction of arrival estimation method based on the fourth-order cumulant and amplitude-phase errors self-correction is proposed. First, the fourthorder cumulant of sonar echo signal is calculated to obtain the covariance matrix. Secondly, the weighted average algorithm is used to reduce the dimension, and obtain the construction matrix after dimension reduction. Thirdly, the amplitude and phase correction are accomplished by recovering the Toeplitz properties of the construction matrix. The phase linear fitting algorithm is added to further improve the correction effect. Finally, the Root-MUSIC algorithm is used to obtain the DOA estimation. Compared with the selfcorrection Root-MUSIC algorithm, simulation results show that mean square error of the proposed algorithm angle estimation is improved by about 0. 07°, which improves the problem that the estimation accuracy of the fourth-order cumulant algorithm decreases under the condition of the amplitude-phase errors of the array elements. Compared with the self-correcting Root-MUSIC algorithm, the sea test data analysis shows that the relative bathymetric accuracy of the algorithm in this article is improved by about 0. 03% ~ 0. 08% and the bathymetric accuracy standard of 0. 75% is satisfied within the horizontal distance of 90 m. Both simulation and sea trail data prove that the DOA estimation performance of the proposed algorithm is better than the traditional algorithm under the conditions of Gaussian color noise and array elements amplitude-phase error.