Shop OBEX P1 Docs P2 Docs Learn Events
Wall Following Robots — Parallax Forums

Wall Following Robots

ercoerco Posts: 20,256
edited 2012-09-01 11:35 in Robotics
My first test was a little shaky:

But then I saw this related video of someone's kit and I felt pretty good! http://www.youtube.com/watch?feature=player_detailpage&v=abncTBjWpk8#t=72s

Comments

  • TtailspinTtailspin Posts: 1,326
    edited 2012-08-27 18:54
    Another easy to describe, yet, hard to master challenge... Sigh.

    I started pecking away at this seemingly simple task, thinking it would be easy with the S2 and Phils super cool Ping((( holder.

    Had to wrap a bit of tape to the bottom of the Ping((( holder, so as to raise it above the wheel when it's pointed to the side.
    No big deal, it only needed to raise about 3/8". So now I am able to make some progress with the code.(well, what I call code...)
    My S2 can follow a wall fairly well, as long as the wall doesn't end, or have a corner...

    I hope Whits crew can step and try this out, It is an excellent learning oppertunity...
    Anyways, thanks again for posting erco. :thumb::thumb:
  • ercoerco Posts: 20,256
    edited 2012-08-27 19:25
    Shall we make it an official challenge, like the Dinner Plate Figure 8?

    Let's see some video!

    Endless walls? Exactly how big is your house, Mr. Contractor?
  • WhitWhit Posts: 4,191
    edited 2012-08-28 11:26
    Thanks for keeping the projects coming...

    By the way, one of the first robots ever was a bit Shakey...http://en.wikipedia.org/wiki/Shakey_the_robot

    260px-Shakey.png
    260 x 600 - 234K
  • ercoerco Posts: 20,256
    edited 2012-08-28 14:14
    To follow walls or lines, chassis and sensor geometry are at least as important as the programming. AFAIK, Martin_H would prefer to "bounce" off the wall by arcing mildly right at all times, then blipping left briefly when the sensor sees the wall. Probably smoother driving overall, but here my primary goal was to maintain a fairly straight path parallel to the wall. The wall sensor is a Sharp IR unit, which works smoothly . But I used an SRF-04 ultrasonic unit for a front-facing sensor, which causes some servo jitter when I ping it.

    You know a PING sensor would interface without any issues to a Scribbler 1. :)

    I'll get around to it unless someone beats me to it. Long live old blue!
  • varnonvarnon Posts: 184
    edited 2012-08-28 16:04
    I saw that chassis recently. It seems pretty reasonably priced. How do you like it? Do the included motors and wheels seem to work okay with carpet?

    I think I like the your behavior, it does look very parallel. It just keeps a consistent distance to the wall with the IR, right?
  • ercoerco Posts: 20,256
    edited 2012-08-28 23:01
    Yessir! Details are in Robot magazine right now and also in the upcoming issue. More to come.
  • Martin_HMartin_H Posts: 4,051
    edited 2012-08-29 15:58
    Erco, I did wall following two different way. The single sensor bounce technique and the two sensor parallel technique.

    Bounce video

    [video=youtube_share;xkYpWYhpIis]

    Parallel video

    [video=youtube_share;AQK6b2CNTrg]

    Complete thread:

    http://forums.parallax.com/showthread.php?139096-Wall-Following-Examples-(video-and-code)
  • ercoerco Posts: 20,256
    edited 2012-08-31 10:20
    Great videos, Martin_H. Nice actions using both methods. What a difference an extra sensor makes, huh? :)
  • ercoerco Posts: 20,256
    edited 2012-08-31 10:23
    These wall racers are dandy, too!
    http://letsmakerobots.com/node/696
  • Martin_HMartin_H Posts: 4,051
    edited 2012-09-01 11:35
    erco wrote: »
    What a difference an extra sensor makes, huh? :)

    Indeed, I've been contemplating using the Ping))) turret to alternately read forwards and to the side. Basically multiplexing the sensor because you can estimate from the last reading the amount of forwards and sideways leeway you have. So you don't need to constantly read both directions.

    Those cars are neat and really move compared with servos.
Sign In or Register to comment.