This is an example of a solution with less than
optimal parameters. For this planner, we only chose 5 random states, and also
we had a very low stuck ratio and stuck path size, which caused the algorithm
to think it was stuck when it was not.
The algorithm took only .12 sec to plan, but it got stuck 7 times. There were a
total of 505 nodes in the solution path.
This is an example of having the parameters
correctly tuned, but just getting unlucky in our random bounces. You can see
that each bounce carried the robot straight back, where it then just followed
the potential function back into the local minima.
The algorithm took .58 sec to plan. An additional 4.88 sec were spent eliminating
unnecessary nodes with our path smoothing function. It got stuck 7
times, and there were a total of 653 nodes in the graph upon completion.
This is an example of our function working in an
optimal manner. You can see that it recognized that it was stuck shortly after coming
near the obstacle. Also, we got 'luck' with our bounce, and we were
carried away from the local minima, where we were then able to find a path to
the goal.
This algorithm took only .15 sec to find a solution, and got stuck only once.
There were a total of only 342 nodes in the solution path.