浅海环境中的匹配模态空间检测器

Matched mode space detector in shallow-water environments

将水下声传播规律融入到算法设计中可以有效提高被动声呐目标检测性能。当声源位置未知时,广义似然比检测器和贝叶斯检测器分别通过搜索和积分的方式来消除声源位置不确定性的影响。但是,基于有限个信号波前实现的广义似然比检测器和贝叶斯检测器在某些声源位置上存在性能大幅下降的问题。为此,利用水下声传播的物理特性,提出了一种稳健的子空间检测器——匹配模态空间检测器,稳健的意义在于：当阵列获取到的辐射声信号能量给定时,检测器可以在不同声源位置情况下提供相同的检测性能。该检测器通过模态空间一定程度上利用了海洋环境知识,获得了比具有相同稳健性的能量检测器更好的检测性能。典型浅海环境中的仿真实验对比结果表明：匹配模态空间检测器相比广义似然比检测器和贝叶斯检测器的峰值性能下降较小、所需的计算量更少、对环境失配的宽容性更好。

Incorporating the law of underwater sound propagation into algorithm design can improve passive sonar detection performance. When the source position is unknown, the Generalized Likelihood Ratio （GLR） and Bayesian detectors handle this uncertainty via searching and integration. However, both detectors suffer from great performance degradation in some source-position cases when they are implemented by only a finite number of signal wavefronts. Exploiting the physics of underwater sound propagation this paper proposes a robust subspace detector the Matched Mode Space Detector （MMSD）, robust in the sense that it can provide a consistent performance in different sourceposition cases given the radiated sound signal energy impinging on the array. The MMSD utilizes the environmental knowledge from the mode space and thus can obtain a better robust performance than the ED, which also has the stone performance robustness. The simulation results in a typical shallow-water channel suggest that compared with the traditional GLR and Bayesian detectors, the MMSD is just slightly inferior in terms of the performance peak, requires much fewer computational loads and enjoys a better tolerance to environmental mismatch.