Here’s our traction control demo that we performed a few weeks ago.
http://www.youtube.com/watch?v=3kszGKFmgBI&feature=channel_page
Nice side by side comparison. That just goes to show the importance of having traction control this year.
Very cool. Nice implementation, well documented results.
Kudos, programmer… whomever you are.
-q
thanks.
Good job, RAWC! I look forward to scouting you in San Diego.
Effective traction control coupled with a good ball dumping system will make your team “devastating”.
The only reason I personally do not like traction control is because of pushing other robots, or your robot getting pushed, then you are kind of screwed. I think the best way of performing this year will be a team who has traction control, but can still turn it off if they need to.
and Great job! We tried doing traction control but we decided against it.
why turn it off?
So it is easier to defend from other robots pushing you, and to push other robots.
If you completely stop moving your robot you’re still going to have those 5 seconds where the robot will start moving slowly and gradually pick up speed.
A well implemented traction control system will decelerate and accelerate substantially faster. Turning it off would make it worse.
Traction control isn’t just slowing your acceleration. When effectively implemented, like it was by 968, it will increase your acceleration. Just look at the video 968 posted.
If you implemented a system where the 'bot increases the power output in relation to time, you’d get that effect.
If you have a traction algorithm that increases/decreases power output in relation to the amount of wheel slip, you would actually be able to accelerate and decelerate much faster; which, in turn, makes it easier to “defend from other robots pushing you”.
Also, with TCS, you’d have a better chance at pushing objects on the field.
If you grab a stack of books and have your robot push against the stack while running your wheels at full speed, you’ll move no where in a hurry. On the other hand, with TCS, you’d push the stack away at the quickest rate your wheels can spin without slipping.
Ah, thanks I understand now I was thinking of just a programmed in “increase speed for 5 seconds then steady the speed” so without sensors.
The Vid is like down or not available, could someone please re-upload this.
EDIT EDIT: IT WORKS NOW, I DUNNO WHY?
A technical question: do you have any sensors to see if your wheels are slipping, or did you just hard-code in empirical values for max acceleration and deceleration?
Very nice, I like it. Saw that bot in blue yesterday, should be a real threat in either color.
We’re using a follower wheel setup with US Digital S4 256 count encoders.
That video’s great, and really shows why traction control is essential.
Here’s a test I’d like to see (done by you, or any team with 2 identical robots and traction control): Take one robot and use Victors on the drive motors, and use Jaguars on the drive motors of the other robot. Keep the software the same (except for swapping the Victor/Jaguar vi/object). (a) test the two robots as in this video. (b) put the robots facing each other, touching, and have a pushing match.
If I remember correctly Youtube went down for updates at about 7pm CMT
it is quite nice… but for our team, we just have a really good driver and he manages without the traction control… Seems to actually work better for us.
It shows why you don’t want to just jam the sticks forward when you’re trying to get somewhere in a hurry.
It’ll be fun to watch how driver skill vs. traction control plays out this year.
Did your team use Labview or Windriver C++ for your software development environment this year?