基于Hausdorff距离的海底地形匹配算法仿真研究

地形匹配算法是海底地形辅助导航系统的关键技术之一。根据海底地形的特殊特性，传统地形匹配算法直接应用于STAN时，易出现定位精度大幅降低和算法不稳定性问题。该文提出一种基于地形轮廓匹配原理的新算法，该算法利用改进的Hausdorff距离作为水深实时测量曲线与海底地形图水深曲线之间的度量函数。改进的Hausdorff距离对匹配对象存在的小误差不敏感，在一定程度上还可反映空间两条曲线的方位关系，该算法定位精度高、稳定性好。利用某海域地形图进行的仿真实验表明：该算法是可行的，定位精度比现有算法提高一倍左右。

Research on Seabed Terrain Match Algorithm Simulation Based on Hausdorff Distance

Terrain match algorithm is one of the key technologies in seabed terrain aided navigation system (STAN). Most of terrain match algorithms are derived from aided navigation system in the airplane or the missile. When they are directly applied to STAN, the positioning precision have fallen off sharply and become unstable because of the seabed terrain particularity. Here, a new approach based on the principles of terrain contour match(TERCOM) is presented. It uses the modified Hausdorff distance as a cost function for the two strips of spatial water depth curve in the survey and the map. As the modified Hausdorff distance is quite tolerant of the small perturbations of the curves and inclusive of their direction partly, the new approach has high positioning precision and robustness. Moreover, experiments with pedestrian simulation confirm the effectiveness of the proposed approach. The system achieves higher positioning precision than that of the conventional TERCOM algorithm.