DEWBOT VI IRI Post-Mortem
From DEW Robotics
DEWBOT VI shutting down opposition scoring 1-on-2
Summary
- The team & robot did very well
- 5 wins - 3 losses
- Finished qualification matches 31st out of 78 teams
- We were not selected for a play-off alliance
- DEWBOT VI played phenomenal defense, virtually shutting down scoring 1-on-2 during two matches
- DEWBOT VI also scored very well in the near field
What went right
- The drive team brought their Mojo
- Siri coached 6 of the 8 matches for 4 wins and 2 losses. Great job!
- The crazy wheel problem did not occur. Replacing steering Jaguars with Victors seems to have done the trick.
- No fatal field failures. We played every scheduled match; through the match.
- Excellent performance in defense
- Very good scoring from field 1
What went wrong
- Communications - we had a communications problem between the Classmate and robot when tethered. No significant field failures. Most times, tethering through the router worked. Most times, tethering directly from the Classmate did not. We know we have a problem with the ethernet port on the Classmate, but this problem appears different.
- Our possessor eats balls. It will not eat the team's practice balls, so there appears to be a significant difference between the two. We will correct this.
- Possessor actuation failed. We destroyed two cylinders on the left side. Issue is poor alignment. We'll fix.
- Possessor grip became less grippy. Need to explore this & resolve.
- The mirror sagged. We removed (to avoid penalty). Need to fix.
- Autonomous kicking often failed. When it worked, it was too late & missed the goal. We need to make kicking reliable and set angles empirically (calculated angles are a starting point, but the assumption for calculated angles is that the robot kicks at the original ball position. this is not the case. If we know speed & delay, we can refine the model)
- The robot appeared to kick by itself in teleop. Kicking was mis-timed and ineffective. This may be a control issue, but more likely a user interface problem.
- Performance in the mid-field is lack-luster. We are not so good at gaining control of balls and scoring from here (and, I think, have low confidence of our abilities as well).
- The user interface is complex and not understood by the drive team (esp. the driver).
- The battery box broke (again). The field fix (from Finger Lakes) is quite effective.
Action Plan
- Resolve comm issues. Purchase back-up Classmate if advisable.
- Purchase accurate game piece. Develop / rebuild possessor. Fix geometry / reliability problems.
- Make sure positioning of optical sensors is correct for autonomous kicking
- Autonomous testing and setting of Auto angles to score
- Resolve user interface issues between drive & programming teams - simplify controls and make these more intuitive.
- Remount mirror securely
- Design / procure / build a battery charging caddy (Design Team)
- Get camera-automatic or camera-assisted aiming working
- Shop smart for some workshop furniture
- Buy friction tape (done)
Other pages about the DEWBOT VI Indiana Robotics Invitational
