Robot Course and Video

Here is a link to a file of our course that we use in this videos below.

FinalCourse.rvl

Picture of Course

IMPORTANT:

The program we are using test the robots, ROBOTC, has a very inaccurate testing platform. In other words one can get the perfect speed and time needed for a robot to make a 90 degree turn, but due to its inaccuracy there is no assurance that it wont do a 80 or 100 degree turn instead. This was a very big problem, as we would run the same code with the same robot on the same course, and most every time its supposed 90 turn would be different each time. This means that it took multiple attempts to get a near perfect run. This problem has nothing to do with the written code, but rather with the programing of the robot and course, which seems to have an inherent and constant flaw that causes inaccuracy, and seeing as this robot is a natural disaster robot, accuracy is essential.


Another important fact that one should know is that in many of the given maps, walls and lines on the ground appear near identical, but our sonar sensor can only detect actual obstructions. The fact that it can not detect lines we believe is of little consequence, as in a real natural disaster situation lines on the ground do not mean failure. Therefore one should use either our course or create one yourself that uses actual walls, and not lines.

Our best run!!!

We were unable to get a perfect run no matter how much we tried for the reasons outlined above (application inaccuracy), but this is the best one we did get. We also decided to only use the sonar sensor, as it is basically the same and better than the bumper and limit switches, and the light sensor would not be useful in real natural disaster situations.

Video showing application inaccuracy

This video is using the same code, robot and course as in the video before, yet it was not even able to complete the first corner due to reasons outside our control.