EI / SCOPUS / CSCD 收录

中文核心期刊

基于频域–阶数域映射的深海声影区声源定位

Source localization in the deep-water shadow zone based on frequency-order domain mapping

  • 摘要: 针对深海近海底声影区内远距离、近海面声源条件下声源定位精度低且稳健性缺失的问题, 提出一种基于频域–阶数域映射的单水听器声源定位方法。将经多次海底反射后到达时间邻近的多途信号视为一个多途组, 通过匹配多途组间的时延差估计声源距离; 随后构建频域–阶数域映射消除多途组间频域干涉结构的差异, 从映射后的信号中提取出受声源深度影响的特征时延差, 进而实现深度估计。不同于仅利用前四条声线的传统方法, 该方法通过充分挖掘所有观测到的多途组包含的声源位置信息, 有效解决了远距离、小深度声源场景下定位性能大幅下降的问题。海试数据验证表明, 在60~80 km的近海底声影区内, 所提方法对50 m和200 m深度爆炸声源的水平距离估计均方根误差分别为2.65 km和2.26 km, 声源深度估计的均方根误差分别为4.10 m和5.85 m。

     

    Abstract: To address the issues of low localization accuracy and lack of robustness for long-range and shallow sources within the deep-water near-seafloor shadow zone, a single-hydrophone source localization method based on the frequency-mode order domain mapping is proposed. Multipath signals with proximal arrival times after multiple bottom reflections are treated as a multipath group, and the source range is estimated by matching the time delay differences between these groups. Subsequently, a frequency-order domain mapping is constructed to eliminate the differences in the frequency-domain interference structures among different multipath groups. Characteristic time delay differences influenced by the source depth are extracted from the mapped signals, thereby achieving depth estimation. Unlike conventional methods that utilize only the first four rays, the source position information contained in all observed multipath groups is fully exploited in this method. Consequently, the problem of significant degradation in localization performance for long-range and shallow-depth source scenarios is effectively addressed. As validated by sea trial data, within the near-seafloor shadow zone at range of 60–80 km, the root mean square error (RMSE) of the horizontal range estimation for explosive sources at depths of 50 m and 200 m are 2.65 km and 2.26 km, respectively. The RMSE for source depth estimation are 4.10 m and 5.85 m, respectively.

     

/

返回文章
返回