信源空间散布条件下的声呐目标滤波方法

A target filtering method under the distributed noise background for active sonar

基于密集多波束接收空间观测数据向量,定义了一个与信源空间分布参数相关的特征量-空间散布指数,并利用蒙特卡洛积分对不同信源散布角宽度、不同信噪比和不同噪声空间分布特性条件下的空间散布指数进行了建模与计算。在此基础上,形成了在分布式混响/噪声背景下检测目标回波信号的空间散布指数滤波方法。利用轻型AUV(Autonomous UnderwaterVehicle)水下航行器声呐获得的实验数据进行了验证,表明该方法可显著抑制了混响及流噪声干扰。该方法无需假设信源间彼此独立,对混响干扰具有良好的适应性;由于混响及流噪声成分被大部分滤除,显著提高了信号处理增益,在后续检测过程中避免了宽波束接收、旁瓣泄漏、主动声呐平台运动等因素引起的背景噪声的显著增加。

A measure of target＇s angular spread is introduced, which is defined as Angular Distribution Index （ADI）. A target echo filtering method is developed based on different ADIs of target, reverberation and noises in active sonar. The ADI is computed from the observation data vector formed by a compact bank of receiving beams which are overlapped in receiving azimuth domain. Properties of ADI under different conditions of SNR, angular spread of target, reverberation and noise sources are analyzed in the formation of Monte Carlo Integration. According to the properties of ADI, the target can be separated from reverberation and noise if an appropriate threshold of ADI is adopted. This method does not require neither the independence of reverberation with echo nor the independence between reverberation/noise sources. And therefore it is advantageous over conventional processors for active sonar where reverberation is always hard to deal with due to its correlation with echo and distribution in space. It is also true for noise-limited target detection which may suffer from much higher noise level due to the beam sidelobe leakage, wide beam receiving and sonar platform maneuvering unless the proposed target echo filtering is applied. Using experimental data from Autonomous Underwater Vehicle （AUV）, it is shown the proposed method can suppress the reverberation and flow noise effectively. As a result, the SNR of target detection can be significantly improved.