Shop OBEX P1 Docs P2 Docs Learn Events
BS2 with infrared line follow kit (280324) — Parallax Forums

BS2 with infrared line follow kit (280324)

krswinneykrswinney Posts: 3
edited 2012-05-16 10:30 in Robotics
We have built a bot around the BS2 microproprocessor and the infrared line follow kit (280324). We are using HB25 motor controllers with a couple of 12 volt motors. We have a time constraint so we are following a curve as fast as we can push it. we have a code similar to the one in the literature that came with the line follow kit (case statements on bit pattern). It works about 75% of the time because it is fish tailing. Sometimes loses the line if it hits the curve at the wrong angle. We can slow it down to fix the problem, but then it does not make it around track fast enough.


The chasis is larger than the boebot, with the perpindicular distance from sensors to wheels (line connecting centers of wheels) about 2.5 times that of the boebot.

I assume that the fishtailing is a common problem and am wondering if anyone has figured out how to correct it.

Comments

  • TtailspinTtailspin Posts: 1,326
    edited 2012-05-16 08:55
    Without looking at your chassis it is hard to say exactly what you would need to do.

    But, sounds like you have to much weight on the back end, You should put all the weight you can over the drive wheels.
    However, that leads to balance problems when you start and stop. and that is why you put the caster there in the first place.
    So... My personal solution for the Fish Tailing caster, is to ditch the Caster,
    and replace it with a Rudder system..:thumb:
    LightShow.jpg
    BackCornerPerspective.JPG

    A Rudder Wheel will steer the weight around a corner instead of trying to continue going straight.

    You get to carry lots more payload with a Rudder, and they just look cool too... But maybe thats just me...:cool:


    -Tommy

    P.S. Welcome to the forum krswinney :smile:
    1024 x 768 - 137K
    800 x 600 - 146K
  • krswinneykrswinney Posts: 3
    edited 2012-05-16 09:06
    Thanks for the welcome.
    :smile:
    Good point about the castors.
    We are still working on adjusting code.
    IF cann't get that too work we plan on moving sensors back toward drive wheels. Seems intuitve that a large distance from sensor to wheels could be a problem. any thoughts on that?
    You can see pictures of bot on facebook, search FEMB.
  • TtailspinTtailspin Posts: 1,326
    edited 2012-05-16 09:25
    Ok, Call it a Tail Wheel, just add a Servo, and you have yourself a genuine "Rudder System", or "Tail Wheel System" if you like...

    Calling it a Rudder, brings visions of the assembly being on the back end of the Robot, But it will also work on the front of the Robot.
    Yes, a large distance between the drive wheels and the sensor will create a large turning arc, and the center of the robot will have a hard time following the line...

    -Tommy

    P.S. I am not real big on FaceBook, could you maybe drag one of those pictures over here?
  • krswinneykrswinney Posts: 3
    edited 2012-05-16 09:42
    I see, I didn't look close enough at the first pictures you posted.
    so you are actually talking about stearing with a single front wheel like a tricycle...

    Not a great picture, but the best we have at the moment. Bot must pass golf ball to another bot. Body and ball handeling mechanism were designed in AutoDesk Inventor and printed on rapid prototyping machine (3d printer). Sets very low to the ground so the rubber system may not work with out major redesign of base. Bot is about 7"x10.5".
    thanks for you help.

    bears-bot.jpg
    640 x 480 - 26K
  • TtailspinTtailspin Posts: 1,326
    edited 2012-05-16 10:08
    Actually, Those Robots have differential steering, AND rudder steering at the same time... Works great!

    You don't need to put a Rudder System on your bot,
    Just shorten the distance between the drive wheels and the sensor.
    Think of it this way...
    Imagine your sensor placed exactly between the drive wheels,
    Try to follow the line, just manually hold the bot over the line and "drive"
    the bot over the line, keeping the sensor centered over the line..
    See how small motor corrections would keep the sensor centered..

    Now, Imagine your sensor attached to the end of a yard stick, then try to keep your sensor centered over the line... Not so easy, eh?

    Hopefully I haven't confused the issue for you, Some one will be along shortly to help sort things out...:)

    -Tommy
  • ercoerco Posts: 20,255
    edited 2012-05-16 10:30
    Per Tommy, you have many variables. Sensor distance from wheel centerline is just one. Line width & geometry (curves or square intersections), chassis size, weight & inertia, motor gear ratio, acceleration and dynamic braking ability, tire width (scrub factor) and traction, and certainly how you are using the sensors. You should be using all 8 sensors for maximum effect.

    BTW, a tiny bot called 3pi is optimized for FAST line following. Check these videos:

    http://www.youtube.com/watch?v=mJV-KDqHgDQ

    http://www.youtube.com/watch?v=RVBYWHOarRE
Sign In or Register to comment.