I have an area that is represented by triangular delaunay triangulations. I play with the problem of finding a path between two points. I am using a paper submitted by Marcelo Callmann as a guideline for solving this problem. However, instead of using the Chazelle funnel algorithm proposed by Kallmann on the paper link , I plan to use the A * search algorithm, which plans the path in the grid quite efficiently.
For the heuristic cost function, I use the Euclidean distance from the current node to the target node, and to select neighboring nodes I draw a line segment from the current point p to the middle from the edge of the triangle. [This idea was proposed by Kallman]. The cost of a neighborhood node is also determined by the Euclidean distance between them.
Here's a figure from Kallmann showing how to use the middle of the edge method to generate various paths to the triangle containing the target node.
Now this method works fine when the density of the triangle is not large enough in the region. But let's say if the generated triangulation for a set of points looks like this:
, , , IDA * ID-DFS? A * (TRA *), , . TRA * 5 thesis Demyen.
. - , .