pic: 1024's Fix-It/Practice
 

Re: pic: 1024's Fix-It/Practice
 

Here's a video of our robot running in non-lane-change race mode for autonomous: http://www.youtube.com/v/-q4NTcanElE ...amazing the difference a spiffy new gyro and a few hours to work with the robot can do.

While the robot just crosses 6 lines in autonomous in this video, our competition robot is considerably faster and should be able to make seven lines on a good battery.

Unfortunately, we have no video of the robot lane-changing, but if you have any comments/questions about it let me know and I'll see what I can do.

-q

p.s. A little later in the evening some people got to pushing the robot around: http://www.youtube.com/v/7IszxvWliPA The improved collision recovery software still kept the robot on track. No video of recovery during arc interpolation, but yesterday we also tested that the robot can continue after being completely blocked in a turn. GO TEAMS, and see you all in Atlanta!!!

Re: pic: 1024's Fix-It/Practice
 

I would be worried about driving that close to the lane divider. For one thing, your bumpers are probably going to rub on the sides slowing you down slightly, which will create a discrepency between how far your robot thinks it has traveled and how far it has actually traveled.

There's also the problem of running into other robots that are stuck on the lane divider (seems to happen a lot).

Re: pic: 1024's Fix-It/Practice
 

The turn radii can be adjusted easily if need be... they were much wider at midwest and other regionals, true, and if there is a problem with rubbing then we'll make them wider again. The arc interpolation on the turns is not a lookup-table based operation, it's done formulaicly in real time, so changing the radii just means changing the radus for the turn in question in the autonomous script.

Why would the robot have a discrepancy between how far it thinks it has gone and how far it actually has gone if it runs into a robot/rubs the lane divider (not that that's the plan)?

Thanks for the comments!

-q

Re: pic: 1024's Fix-It/Practice
 

Well I know on our robot we're using encoders to measure distance. Whenever we started on the side closest to the divider, and rubbed it, the robot would turn too soon and miss the turn.

But when we started from the middle position we never had any problems.

I guess if you get really clever with the gyro you could use it to sense deceleration.

Re: pic: 1024's Fix-It/Practice
 

Thats looking good. Kicking the robot in auto is fun isent it. (its one of my favorite activities, right behind starting it in funny positions to see if it can recover :P)

hope we play each other in Atlanta so we can race...
(we did the same thing with our practice bot (aka Swampthing 07))
http://www.twzdesigns.com/MOV02281.MPG

Re: pic: 1024's Fix-It/Practice
 

Looks good... knew you guys still had some tricks up your arm. ;)

Re: pic: 1024's Fix-It/Practice
 

Looking good! Is this your 2007 robot in the video? 2008?

If it's your 2008... it would be fun to watch a race! :D

-q

Re: pic: 1024's Fix-It/Practice
 

its our '07 bot modified and weighted to match our '08 bot so all speeds and performance are the same.