Our main kick is looking better- not that it could have looked any worse easily. With one front leg coming back and one going forward, in addition to really getting a legitimate push off the back legs, we have a good amount more power. The back kick works sometimes thanks to the help Murch and Jesse provided, but is not easy to say the least. It doesn’t help that the dog keeps on dying mid-kick for no apparent reason.
Upgraded to WordPress 2.0.2. Fixed odd font size issue. Atlanta Three (!) Days away.
Robo dogs don’t exactly ‘bend it like Beckahm,’ but they get the job done – The Times Record 04/17/06.
My favorite line: “The thing is, robot dog soccer is like a regular soccer game in slo-mo. Soccer on Ambien.” We are the lead picture in the ‘news’ section also. Nice work, RoboCuppers!
Really great turn out, and our dogs scrimmaged super well. Blue team bested Red Team 5-2.
9:31 PM: It’s hour four and a half and we’ve split in two — four dudes have gone off to the lab to work on some things while the rest of us are in Daggett readying the Color Tables for Vision recognition. Spirits are high because hell, it’s just robots. Pictures Soon.
9:35 PM: We just added the Orange, Green, Blue, and Yellow thresholding–things are looking pretty good. Ball recognition not that bad. We ran into a bit of an issue when, after calibrating orange for a good hour, nothing remotely orange appeared on the screen. Turns out that our horizon detector was messing up orange because it smartly needs transitions from green to orange to identify.
9:41 PM: YAYY!!! Our WordPress Spam Filter Akismet has caught its first fish! Its content? ‘Compare Diet Pills’ ! And sometimes we wonder why we actually block spam!?
9:42 PM: I’m a blogging nut! And cooler/taller than Jesse Butterfield!
12:08 AM: Woah, forgot about blogging for a while. Prof. Chown came and just left. We had our first _ever_ 4×4 scrimmage this year. Pictures when we make it back to the lab and get the cord. Didn’t go so badly, and we are fixing a lot of simple bugs now.
12:34 AM: We’re going to cheat a little bit and get some walls for the edges of the field. This will hopefully solve some phantom ball issues.
12:58 AM: Slowing Down as we all take to our various corners and fix our various behaviors.
1:42 AM: We’ve just heard the word: we have to be out by 3:30 AM. Dining Services, even in this area, still rocks the house. Big improvements on Player so far–we’re working on some glitches in SEARCH mode. KILL mode later.
2:52 AM: Final Scrimmage of the Night.
Tonight I worked on trying to get the trap and kick better. I looked at Player, and together with Jesse and Henry tried to figure out a better way for the little guys to a)kick the ball from a trap and b) confirm that they actually have trapped the ball. For part (a) I also watched video footage to get some ideas– some of those kicks are weird.
Robot-enthusiasts! If you’re in the Brunswick, ME area and want to see Bowdoin’s RoboCup team in action, come to Daggett Lounge (Thorne Dining Hall) on Bowdoin College Campus this Friday, 2-4pm! At 2:30pm we’ll be holding a big scrimmage and at 3:30pm we’ll be showing off some demostrations of our robots’ abilities. The public is welcome to come between 2-4pm. The paper and TV press will be covering the event and many Bowdoin students, faculty, and staff will be attending. For the team, this scrimmage is the first test of its code base in preparation for the U.S. Open in Atlanta and the World Championships in Bremen, Germany.
Any questions about the event feel free to contact us!
For the past week, I’ve been trying to implement line recognition. The purpose of this is two-fold: at some point, we might be able to use a better version to help with localization and, more importantly, we can use it to try to keep our robots from crossing lines that they should not cross.
I’m using the green-white transition rules that I devised before spring break, and am hoping to be able to have the dogs discern which of the six line objects they might be looking at (a corner, a midline intersection, a goalie box intersection, a goalie box corner, the center circle or any other boundary line).
Just committed our 400th revision to our SVN repository code base for this year’s team. And it’s a goody: the dog really finds and moves to the ball well, and traps are up to about 75% accuracy.
I’ve set up the foundations. It is working whimsically as of now, but it has great promise. It is certainly necessary to perfect this since the full game test we made today caused all 3 dogs to get stuck into each other and couldn’t continue playing unless taken away from each other. In a real game they would have all gotten penalized for 30 seconds…Not a good thing!
Before the dog goes into approach state, it considers whether there is another dog that is closer to the ball. If there is, stay put. This is all based on communicated localization from other dogs on the field. Eventually, we could give idle standers-by some other work while a single robot approaches the ball.
