In our regional we found our robots ability to hit its target points changed throughout the competition, and part of the issue seems to be tread wear. We were using blue nitrile treads which were severely worn by the end. Just 0.050 inches of wear on a 3.75 inch diameter wheel would cause 6.4 inches of error over 20 feet of travel, which is enough to cause problems.
We looked into making the paths able to accommodate a range of travel distance since most of the travel is along the length of the field, and we can pick up a ball on the move. However, the ball near the human player is the real problem because there’s not much additional distance for the robot to continue past.
How do other teams account for that in non-vision based auto? Monitor the actual wheel diameter and update that in their robot code throughout a competition, frequent tread changes, something else?
My team switched from blue nitrile to Buna-N and found it lasts almost 5x longer with about the same amount of grip on the carpet. You will have to cut to width and length but we have found it to be much more durable. https://www.mcmaster.com/5994K9/
Monitoring it and just adjusting a wheel diameter constant in the code as the autos start going short. You can push the robot while disabled and compare the measured vs actual and just multiply based on the ratio of the two. Changing the tread often also works.