With the increasing demand for the autonomous ship, a fast planning method for long-distance ship routes is needed. In this paper, a multi-scale Visibility Graph(VG) method is proposed for long-voyage route planning, as a solution to the problems of the slow planning and poor route accuracy. First, polygon data of obstacles are extracted from an electronic chart. In order to reduce the number of Visibility Points(VPs), the VPs are expanded from the convex points of these polygons. The small-scale, medium-scale, and large-scale VG models are established respectively. Second, this paper proposes the Local Planning Window(LPW) method, which greatly reduces the complexity of the VG models. The great circle route method is used to decompose the longer route, which further shorten the search time of the VG. The route planning process is designed for the multi-scale VG method. Finally, a long-voyage route planning example is carried out, in which, the utilization rate of the number of obstacle polygons and the number of VPs are analyzed. The data results show that: the complexity of VG models can be reduced greatly, and the search time of the VG will be shortened, by using the multi-scale VG method.

中文翻译：

---

基于多尺度能见度图的自主船舶远航航线规划方法

随着对自主船舶的需求不断增加，需要一种快速规划长距离航线的方法。本文针对长航次航线规划提出了一种多尺度可见度图(VG)方法，以解决规划速度慢、航线精度差的问题。首先，从电子图表中提取障碍物的多边形数据。为了减少可见点（VPs）的数量，VPs是从这些多边形的凸点扩展而来的。分别建立了小型、中型和大型VG模型。其次，本文提出了局部规划窗口（LPW）方法，大大降低了VG模型的复杂度。采用大圆路径法对较长的路径进行分解，进一步缩短了VG的搜索时间。路线规划过程是为多尺度 VG 方法设计的。最后，进行了长航路规划实例，分析了障碍多边形数量和VP数量的利用率。数据结果表明：采用多尺度VG方法可以大大降低VG模型的复杂度，缩短VG的搜索时间。