基于单个矢量水听器和子空间旋转的宽容宽带双目标分辨与跟踪方法

A robust separating and tracking method on two wideband sources by subspace rotation with one vector hydrophone

理论证明线性信号处理可以分辨两个不完全相关的目标。针对宽带双目标问题,根据矢量水听器信号模型,设计了并行的两路结构相同的处理通道,每个通道先将接收信号做子空间旋转,分离出单个目标,然后用空间匹配滤波方法估计该目标方位。仿真研究表明,该方法适用于任意频谱结构的信号,对目标强度差和初值误差的宽容性好,可以全方位跟踪目标而性能不受目标具体方位影响,并具有时间带宽增益,在相关目标或者低信噪比情况下的性能优于宽带聚焦子空间方法。海试数据验证了该方法的可行性。该方法是一种比较实用的水下双目标分辨与跟踪方法。

It has been proved theoretically that two targets of incompletely correlated could be identified by linear signal processing methods. Aiming at difficult problems upon two wideband targets the vector hydrophone signal model is deduced, then two parallel processing channels of the same structure are designed each of which has two steps, first subspace rotating on the received signals to separate a single target, then estimating the target＇s azimuth with spatial matched filter. Simulation researches indicate that this method were acceptable for every kind of bandwidth and for two targets even much different in strength. The method is insensitive to initial azimuth error, robust the same efficient in all the azimuths and it possesses time-bandwidth-gain. In the circumstance of correlated targets or low SNR, this method outperforms wideband focusing subspace methods. The method is verified with sea trial data. So it is a very practical separating and tracking method on two targets in underwater.