【摘要】 ESPRIT算法是一种用信号旋转不变性来估计信号参数的方法,属于子空间分解技术。应用于矢量阵扩展孔径的方位估计时,则需要一定的算法去除其本征模糊。文中提出了一种适用于矢量水听器线阵的ESPRIT去模糊算法——MUSIC,并与另一种粗略估计去模糊的方法在性能上作出比较。仿真的结果证明:使用MUSIC方法不仅可以去除ESPRIT算法中的本征模糊,还可以克服线阵中的左右舷模糊问题,就不必进行归一化、不需要考虑声压传感器的存在。不论是单目标方位估计还是双目标分辨都具有更大的方位角估计适用范围和更强的稳定性。更多还原

【Abstract】 ESPRIT is a signal parameter estimation method using rotational invariance approaches,which is a subspace decomposing technique.When applied to extended-aperture DOA estimation,some way to eliminate cyclic ambiguity in the phases of ESPRIT eigenvalues is needed.In this paper,a MUSIC-based disambiguation method suitable for ESPRIT in a linear vector-hydrophone array is proposed.Compared with the disambiguation method that derives unambiguous coarse reference estimates,MUSIC can not only eliminate ambiguity in ESPRIT eigenvalues but also solve the problem of port/starboard ambiguity in customary linear hydrophone arrays.Normalization estimator is not required.MUSIC is effective and stable in most azimuth angles in the case of one or two sources.These have been confirmed in simulation.更多还原