基于主分量反转-递归最小二乘的双基地直达波抑制算法(英文)

Principal component inverse-recursive least-square algorithm for direct wave suppression in bistatic sonar

由于目标回波信号与主动声纳直达波干扰能量比值小,且目标回波信号与直达波干扰相关性强,因此双基地声纳目标回波的检测受直达波影响严重.为了克服双基地声纳系统中主动声纳直达波干扰,达到检测弱目标回波信号的目的,本文提出了一种时域分块自适应主动声纳直达波干扰抑制方法,该方法利用主动声纳直达波干扰能量远大于目标回波信号的特点,将主分量反转算法(PCI)与最小二乘算法(RLS)相结合.该方法首先对接收信号分块构造前向矩阵,然后利用主动声纳发射信号形式与参数已知的条件,通过主分量反转算法对构造的各前向分块矩阵中的直达波干扰分量进行提取,最后将提取得到的直达波干扰作为参考信号输入到递归最小二乘算法的自适应滤波器以剔除直达波.其中,主分量反转算法可以通过对构造的前向矩阵进行奇异值分解(SVD)实现,递归最小二乘算法收敛速度快,能够动态的跟踪直达波干扰信号的局部畸变.文中还对构造的前向矩阵的秩与主动声纳发射信号形式与参数的关系进行了分析.最后,进行了计算机仿真和湖试实验对本文方法进行验证.计算机仿真和湖试实验结果表明,本文方法能够较好的跟踪直达波信号的局部畸变,有效地抑制直达波干扰,降低直达波干扰对目标回波信号主动检测的影响.

The bistatic sonar which contains separated active and passive sonar has both the advantages of them.It has great potential advantages in anti-stealth and anti-underwater acoustic countermeasure.But detection of the target echo is often heavily influenced by the active sonar direct wave interferences,due to that the signal-to-interference ratio（SIR） is very low and the direct wave interference and the target echo are highly correlated.To improve detection of the target echo,it is of crucial importance to suppress the active sonar direct wave interference in bistatic sonar system.This study focuses on a vector bistatic sonar model in which the passive sonar is vector.The following assumptions are made： 1） the active sonar direct wave interference and the target echo differs from each other only by time of arrival and amplitude;2） the active sonar direct wave and the target echo are partly overlapped in time-domain;3） there is only one vector hydrophone in the passive sonar.For suppressing the direct wave interference and achieving the purpose of the detection of weak target echo under these assumptions,a method based on time-domain blocked adaptive active sonar direct wave interference suppression is proposed.This method combines the principal component inverse（PCI） and recursive least-square（RLS） algorithm.To distinguish the direct wave interference from the target echo plus marine environmental noise,it is important to choose a metric firstly.The direct wave interference and the target echo are highly correlated and they are partly overlapped in time-domain,so it is hard to distinguish them in frequency or time domain.Another important difference between the direct wave interference and the target echo is power.As the power of the direct wave interference is much greater than the sum of the target echo and the marine environmental noise,it can be chosen as a metric.In bistatic sonar,the signal type and parameters of the active sonar transmitted signals are all known in advance.So with the priori knowledge of the transmitted signal,the direct wave interference is firstly extracted from the received signal by PCI algorithm which can be realized via singular value decomposition（SVD） of the generating forward matrix.The forward matrix is generated with the receipt signal blocked in time-domain.The rank of the generating forward matrix to different transmitted signals is analyzed.Having extracted the estimated direct wave interference,it is then taken as a reference signal for RLS algorithm to remove the direct wave interference.As the adaptive wiener filter of RLS algorithm is used,it can track the distortion of amplitude dynamically over time.Both computer simulations and lake experiments carried out in a lake in eastern China are given to verify the validity and feasibility of the algorithm.The results show that this method is efficient for suppressing the direct wave and tolerable of local distortion of direct wave.