Bug navigation
Lab #2 write-up
Michael J. Chan
Ben Hulse
Peter
Kasting
2/27/01
Results
Bug Algorithm 2
The bug algorithm in this section is a little
sketchy. It makes a starting rotation toward its goal at the outset. Then, it
runs the standard code from Project A (go straight and wall-follow). Once it
encounters a point where the s-line intersects (and provided it was
wall-following and a slope tolerance of 5 degrees), it calls the rotate function
again to reorient itself. In order to try to avoid missing the turns and
stopping point, the bug slows down within 5 inches of the targets. Some tweaks
were made to make this code work with this project, so it is not as robust as in
its previous form.
On
the second map, the results ended up being the same, though the bug refused to
heed the sline in the middle of the traversal. Instead, because of the
fortuitous shape of the map, it wrapped around and finished. I believe this is
because this algorithm does not check the angles and relative location while
doing a hard corner turn. It is a hard-coded turning function that seems to
ignore most of its surroundings.
All in all, the algorithm works even though it is a little loopy, and it
misses a few cues, but at least in these cases, it turned out alright. All it
needs is better checking, both in turns and while racing at break-neck speed.
Sometimes, if conditions were just a little unfavorable, the simulator would
skip, and the bug would go blithely onward. Thus, some greater tolerances should
also be made here.
For the code, click here.