Quote:
Originally Posted by GeeTwo
I cannot tell where the start or end points are; I thought this was searching from one point (presumably 57, 56), not two. It would be helpful if the nodes were color-coded by the iteration number. If it would not be too cluttered, displaying the obstacles would be nice, as well.
Finally, I'm not sure if I misread the original problem statement, but I understood that the goal was to optimize the route, not the search for any route. In optimizing the route, you would probably want to use the corners and edges of the obstacles as search nodes, as well as random points away from the obstacles. This tactic would also help locate paths through narrow gaps that the unconstrained exploring tree would be likely to miss.
|
Huh. I never thought about using the obstacles to guide the path itself. I always viewed them as disallowing certain paths (I'm here, I can't go to the right because there is an obstacle there, go somewhere else).
I am working on making the visualizer run in real time as the tree is being constructed, but I'm on holiday so I don't know when that will get done. I think that would be cool. This little demo doesn't find paths and therefore doesn't take obstacles into account. I wrote the program first and am just now getting to the visualizer.
The final demo program will be the actual program I submitted to @faust1706 with the case n = 2 (maybe another visualizer where n = 3), and it will be of the two trees growing with obstacles being displayed, with every piece of what is happening.
I wish I had the ability to actually use this on a small arduino bot or something, but I don't know how to do vision processing to find obstacles or make the robot go from (x1, y1) to (x2, y2). If I do could do that, then I'd love to see them on a quadcopter for some aerial maneuvers. I have lots to learn in terms of control theory and theoretical computer science!