DEWBOT VI Finger Lakes Regional

From DEW Robotics
Jump to: navigation, search
Downingtown Area Robotics Team 1640
Team after crating the robot
participated in the 2010 Fingerlakes Regional held at the Rochester Institute of Technology on March 4-6, 2010.

Twenty-five roboteers, mentors and parents made the 6 hour drive on to Rochester NY to watch Sab-BOT-age compete. After getting off to a great start of being the first robot inspected, we set off to the practice field.

During practice we uncovered some code problems that we worked on, but never quite got 100% fixed. In a first for us we made all of our practice matches. Our new "design before building" practice has created robots that are mechanically sound. The only work on the robot on Thursday was to replace some bolts on the front of the robot that was interfering with the ball. This work was planned before we left and Clem's step by step instructions made the work go fast.

Thursday team dinner was at Dinosaur BBQ, a downtown Rochester Landmark.

On Friday the DEWBOT VI Chairmans Award presentation team did a final practice run and then did their "Ignite Sab-BOT-age" presentation for the judges. It went well and we got lots of great comments back. Only 11 teams out of 44 made the effort to compete for FIRST's highest award. We did not win, the winning team was a two time Chairman's winner.

Friday morning also saw the installation of a new chain tension device, a PVC cap from Home Robot Depot. Friday qualifications saw us move up and down many times in the rankings, topping out near 15 and getting down around 40. Control system problems were an annoyance and the team worked on them non-stop.

Friday night we celebrated Jen's 18th birthday with cake and pizza in the hotel. Some of the pit crew went to China Buffet of Rochester beforehand to decompress and review the day's events. Rizzo, Foster and Siri also met later to review and share perspectives (Rizzo acted as a referee at the event).

Saturday was a repeat of Friday on the field. We were not chosen for the final alliances, but we did use the time to allow the programming team to work on the robot.

This was not such a great performance for us. We finished the competition ranked 37 of 44, winning only 2 matches, losing 7, and getting a tie (zero-zero) in 1. We found several control bugs that we're addressing, though we ran into very few major mechanical problems. We hope to fix many of these by the Philadelphia Regional at Drexel University on March 25-27, 2010.

On a positive note our pivot drive got lots of attention from other teams, judges and spectators. Another team had a very similar setup and was impressed with ours. We also received kudos from the Safety Judges for our constant use of gloves to move the robot around, safety glasses always being worn, keeping our pit (and the one next to us) cleaned and all loose material picked up. By Saturday we had received 35 Safety Tokens, a new team record.

Our new "mascot" screwdriver also got lots of attention. It was in motion a lot, dancing with the teams and serving as a limbo pole during one of the breaks. It was visible from everywhere in the arena and even appeared on the local RIT television channel. Even though we were a small group it helped make us visible to other teams.

Improvement Plans

Mechanical

  • To do for Drexel - We need to readdress the possessor. On-field performance and pit research demonstrated a major correlation high-scoring robots and effective possessors--usually those with high tangential velocity . This is possible, but will need some work.
  • Fixed on-site - The right front steering chain was loose (we knew this), causing steering problems. This chain needed be tensioned - Clem & Siri made a quick stop at Home Depot to pick up a PVC end cap and installed this as a chain tensioner before our first competition Friday morning. The fix was effective and shows very little wear. It ought to last through Drexel (it did indeed), though we'll bring spares and plan to make a permanent nylon fixture for the post-season.
  • The latex tubing wears out occasionally. The pit checks routinely for signs of wear and replaces the tubing before any failure occurs. The 4-pass through tubes were replaced Friday morning before our first competition match. The 1-pass back tube was replaced Saturday...
  • ...but late on Saturday all (5) latex tube passes failed simultaneously (or nearly so). Analysis of the sheared tube ends show apparent shock damage. Observation of the kicker cycle shows that cocking starts before a kick is completed. Review of the code confirms that the delay between firing the kicker and re-cocking is zero. We believe that the damage was caused by a code change which eliminated the delay between kicking and re-cocking (can we confirm this by comparing with older code?). The high-energy collision between kicker and cocking arms severely shocks the entire kicker/cocking system. The tubing has been replaced, but there is concern this abuse may have caused other yet-to-be-observed damage.
  • Fixed on-site - We seem to occasionally bend the side-guard eye-bolts, probably from incorrect placement on the cart. They still worked and are easy enough to replace, but we should be careful. We replaced the bent eye-bolts.
  • To do for Drexel/temporarily fixed on-site - The battery mount broke late in the qualification matches on Saturday. This did not impact the robot's performance during the match and a solid field repair was made using Aluminum angle before we needed to queue for the next match. Failure originated at an unradiused inside corner (a stress concentrator). We will re-form two (one spare) replacement mounts for Drexel with radiused corners (to avoid stress concentrators) and slightly looser fit.
  • Possibly to do for Drexel - There's some driver concern about bungees as side-guard material. We may experiment with this for Drexel, though not to the determent of going over the bump. We only had two "ball infiltration" penalties at FLR, one on the side during our first competition due to human miscalculation, and the second at the robot front due to lack of robot control (St. Vitus' dance - a.k.a. the "happy dance" - continued from autonomous throughout teleop).
  • Fixed on-site - We occasionally lose our patented kicker latch wire-ties. This is an easy fix, but they need to be check regularly, and we should probably expand who has this replacement skill.
  • Fixed on-site - The centeror velcro needed replacement after match 62(?) in which we ran into some problems doing over the bump. This isn't difficult, though we didn't quite do it correctly (the back isn't on). We can fix that, though hopefully readdressing the posessor will make it unnecessary.

Programming

Likely Priorities: 1) Make the happy dance stop--all the time. 2) Either reinstate a less sensitive snake mode or create a reliable roll forward-shoot-score autonomous mode(s), if the new possessor makes snake unnecessary (by the time this priority decision is made). 3) Create reliable camera aiming code for real conditions in teleop. This does not need to be active, but rather could resemble DEWBOT V's "locked on target" display (though it may or may not be heads-up). 3) Augment (especially autonomous) with camera and/or IR sensing. This is not a major priority, considering both how little scoring/ball movement can happen in autonomous and how well correctly aligned and coded robots (without sensing) can and did preform in this area.

Field Team

  • Practice counts. We didn't. It showed. We should fix this at Drexel on the practice day (Thursday).
  • The robot is not fragile. We can be more aggressive.
  • We have not developed clear tactics. Poorer opponents with clearer tactics have beaten us. More practice, thought, and research should help with this.
  • Decision to abandon snake - was this wise? (Note: we hope to address this before Drexel; the original snake mode was sensitive to the point of irrelevancy for the drivers in competition--who subsequently did not use it at all. The command became an on-the-spot wheel zeroing function, which was used to at least some good effect.)
  • How will we score via kicking if the vision assist can't be fixed for Drexel (so far, we do not)? Regardless, we'll need to address this at Drexel on Thursday.

What worked

  • We were the first robot passing inspection
  • Pivot drive operated well in Crab Mode after the chain tensioner was installed and control lag reduced, except for during "happy dance" incidents
  • We go through the tunnel - in fact, we were the first team to do so
  • We go over the bump
  • We don't tip over (thus far)
  • Breakdowns were quite limited. The broken battery box was the largest mechanical failure. Had to replace the passive centeror velcro. We wear out elastic tubes and kicker latch wire ties. Side-guard eye-bolts were bent. No electrical issues excluding the sudden loss of communications in match 41.
  • Bumper changes are fast & easy (but a little tough for big hands)
  • All of the maintenance we had to perform was easily and quickly accomplished. The robot is very serviceable.
  • We attracted a great deal of attention for our design. Unfortunately, our performance did not impress anyone. We could talk the talk, but we couldn't walk the walk.
  • Human Players - Andrew & Matt performed flawlessly.
  • We got very few penalties - just two, both "ball infiltration" violations, one early on from human miscalculation and one from loss of robot control during the "happy dance".

Match and Standing Results

Drive team lists are estimates.
Match 4: Rough start, we lost 0 to 9 after getting a "ball infiltration" penalty for allowing a soccer ball more than 3" into our frame. Drive Team Paul, Jen, Matt, Clem.
Match 14: 0 to 0 tie, partially for penalties on both sides -- this happened to a lot of people. We were, however, the first robot to go through the tunnel in the entire competition. Drive Team Carly, Paul, Matt, Siri.
Match 21: Won 2 to 1 with 610 (eventual finalist) and 3015, beating 217 (champion), 3181, and 316. Drive Team Carly, Paul, Matt, Clem.
Match 26: - Lost 3 to 8. Drive Team Carly, Paul, Andrew, Clem.
Match 31: - Lost in a painful 1 to 2. Drive Team Carly, Paul, Andrew, Clem.
Match 41: - We lost communication to the robot early in teleop, but our allies 217 (champion) and 1591 pulled out a 9 to 0 victory against 2053 (quarter-finalist), 2340, and 1405. Drive Team Carly, Paul, Matt, Clem.
Match 49: - Lost in a painful 2 to 3. Drive Team Carly, Paul, Matt, Clem.
Match 54: - Were killed 0 to 3. Drive Team Carly, Paul, Matt, Siri.
Match 62: - Lost with a close 1 to 2. Drive Team Carly, Paul, Matt, Clem.
Match 69: - Lost 1 to 3. Drive Team Carly, Paul, Matt, Clem.

Seeding Score: 51.00 (highest was 118.00)
Coopertition Bonus: 6.00 (highest 52.00)
Hanging Bonus: unsurprisingly, 0.00 (highest 22.00)
See Seeding Points Description for details.

We scored twice, both times by herding balls into the goal. We kicked a number of times (autonomously and teleop), but did not score by kicking.

Chairman's Award Presentation

Carly, Sasha and Doug were the chosen representatives to do the presentation. There was a flurry of minor editing that went on Thursday afternoon to complete the slides. Foster took them to Kinkos after dinner and spent 1.5 hours getting handouts, presentation copies and poster boards completed.

The team reviewed the materials a few times in the demo area before they were called in. It went well and we were given comments and notes to work on for next time.

Out of 44 teams, only 11 submitted Chairman's Awards. 8 of the 11 were prior Chairman's award winners, so we were in good company. We didn't win, but we came away with great ideas on what to work on for next year.

Thanks to all that helped work on the document and the presentation (Andrew, Doug, Sasha, Nicole) and to Nicole with a last minute assist from Siri on creating our musical slide show.

Crew

Julie Bachman - Mentor
Clem McKown - Mentor
Siri Maley - Mentor
Faith McKown - Mentor
David Moyer - Mentor
Mike Rizzo - Mentor
Elizabeth Murphy - Mentor
Foster Schucker - Mentor
Rita Wall - Mentor
Carly M
Amanda F
Brianna E
Jennifer S
Sasha W
Nicole W
Paul K
Matt K
Cole R
Douglas M
Ben W
Michael B
Andrew W
Ben R
Dhananjay

Finger Lakes Regional Photo Gallery