Shop OBEX P1 Docs P2 Docs Learn Events
Erco's Figure Eight Challenge - Page 5 — Parallax Forums

Erco's Figure Eight Challenge

1235724

Comments

  • Martin_HMartin_H Posts: 4,051
    edited 2012-10-15 06:38
    Nice job. I can't wait for my motors to come at the end of the month. The girls are getting big too!
  • ercoerco Posts: 20,244
    edited 2012-10-15 08:38
    idbruce wrote: »
    Alright Martin_H and erco

    I think you guys are having just a bit too much fun.

    I have not really spent any time just having fun with microcontrollers, but I like a challenge and this figure eight challenge seems interesting. Are there any rules that apply to hardware?

    Bruce

    idbruce: Initially it looked like you were nearly seduced by the figure 8 side, but something stopped you from joining the fun. If all goes well with your patent, you may soon be a man of leisure (pronounced leh-zher, of course) and hopefully you will add the figure 8 challenge to your list of newfound dalliances!

    As Matt pointed out, it's been too long since we had a contest. If the figure 8 has run out of steam, I will spec out another challenge.

    Hmmm. Steam. What's crabfu's email addy?

    Personally, I think Cornell's walking Ranger robot covering 40 miles is a worthy challenge, but I've learned that the key to getting lots of people involved is to keep things simple. That includes a simple robot and a simple goal. And a small course as easy as setting two plates on the ground. Might be fun to include a simple demo at Parallax Expos. Another giant wooden table for Matt to build! Maybe just barely larger than a standard 4x8 sheet of plywood, for maximum inconvenience. :)

    http://www.engadget.com/2011/05/11/cornells-ranger-robot-walks-40-5-miles-on-a-single-charge-does/
  • ercoerco Posts: 20,244
    edited 2012-10-15 22:18
    Faster run, 38 seconds versus original 56 seconds. Eliminated 1-second stop at center. Shorter timing loop on the stepper pulses increases motor speed at the risk of stalling the motor by skipping pulses. Slight acceleration (velocity ramping) used here at start, finish, and center transition from left to right turns. Might be as fast as these can go at this voltage (7.4v) with a light load on them.
  • Martin_HMartin_H Posts: 4,051
    edited 2012-10-16 06:23
    Nice. I've found that steppers get hot, particularly when you pulse them slowly or go over their voltage rating. I believe these motors are rated for 5 volts, did you notice them getting hot?


    BTW once rwgast_logicdesign‎'s robot is ready we need to put some peer pressure on him to try the figure eight challenge.
  • ercoerco Posts: 20,244
    edited 2012-10-16 07:39
    Peer pressure? Moi? Isn't that against forum guidelines? :):):)
  • ercoerco Posts: 20,244
    edited 2012-11-26 13:47
    Just had a neat thought after reading http://en.wikipedia.org/wiki/Braitenberg_vehicles , which mentions circling around a point. That's a pretty basic maneuver for a private pilot. A robot could do a figure 8 around two IR beacons with side-looking sensors. Vehicle dynamics would affect how circular the path is. Spiraling in, oscillations add a whole new challenge. I'll add this to my to-do list with my Magician chassis, I'm nearly there already with its ability to track to IR beacons.

    http://www.youtube.com/watch?v=3rHuFFe6HRc

    No new or recent takers on the Figure 8 challenge. Has it run its course, literally? Do we need a new challenge? According to Duane, hardware is easy, but programming is hard. Is that true? Is that the problem?

    Forthers had a Thanksgiving challenge. Anyone interested in a new SIMPLE robot maneuvering challenge over the Christmas holiday?
  • Martin_HMartin_H Posts: 4,051
    edited 2012-11-26 13:54
    erco wrote: »
    No new or recent takers on the Figure 8 challenge. Has it run its course, literally? Do we need a new challenge? According to Duane, hardware is easy, but programming is hard. Is that true? Is that the problem?

    Forthers had a Thanksgiving challenge. Anyone interested in a new SIMPLE robot maneuvering challenge over the Christmas holiday?

    I'm in. I kind liked my challenges of Peril and Doom, but never got any takers.

    http://forums.parallax.com/showthread.php?138711-Martin-s-Figure-One-Challenge-of-Doom-and-Peril!-(Video)

    So perhaps something easier for anyone to replicate would be just the ticket. I'm almost ready to unveil my latest robot too.
  • mindrobotsmindrobots Posts: 6,506
    edited 2012-11-26 14:06
    A Christmas Challenge.....sounds like a plan!

    If I enter, my bot will be driven by Forth! In the spirit of "programs shouldn't just be hard to write, they should be hard to read too!!"
  • ercoerco Posts: 20,244
    edited 2012-11-26 14:47
    Two entrants already! Two more and we'll outdo the total number of participants in the Forther's Challenge! Whit? Amanda? Duane?

    @Rick: Way to combine challenges. Two birds, one GIANT stone.

    @ Martin: Yes, keeping the "course" simple to replicate is key.

    In fact I may have it...

    Just one juice can or oatmeal box. Whatever's big enough for a PING or a $2 ultrasonic sensor to pick up from 2-3 feet away. One side-looking PING or equivalent should enable a robot to circle around it at a 2-3 foot radius. Probably not difficult to to herky-jerky style, but it may take some careful tuning or a PID loop to make the path fairly circular.

    What say you boys? You know me, I'll have an old blue Scribbler 1 on the case ASAP! Been looking for a reason to hack a PING into it. BTW, I found something interesting in "another" forum about using the $2 HC-SRF04 sensors operate off of one pin. Have not had a chance to try it myself: http://www.picaxeforum.co.uk/showthread.php?22639-HC-SR04&p=223675&viewfull=1#post223675
  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-11-26 14:59
    erco wrote: »
    No new or recent takers on the Figure 8 challenge. Has it run its course, literally? Do we need a new challenge? According to Duane, hardware is easy, but programming is hard. Is that true? Is that the problem?

    Actually, I've been thinking about the Figure 8 Challenge a lot lately. I've refined some of the software for my Mecanum wheeled robot. I had a bug in the program that limited the travel angles to increments of 45 degrees. I think it would perform better now that it can move an any direct.

    I just finished building the hexapod Paul K. gave me. I don't think we've seen any walking robot perform the challenge so I want my bot to run the course (possibly under my control).

    I also want to try this with my three wheeled omni-bot.

    Those beacon are sweet. Are thinking of incorporating beacons into a challenge?
  • ercoerco Posts: 20,244
    edited 2012-11-26 15:13
    Hey, I love the IR beacons, they're cheap and easy to make. But no one else will build them. Maybe you & Martin, if I'm lucky. Maybe TommyT, if I start promoting casters heavily. :)

    I don't know much, but I do know human nature. The ONLY way to get lots of people to participate is if it doesn't require any apparent skill or effort, and the "course" is effortless to create. A juice can is about as simple as it gets...

    mindrobots (Rick) apparently likes to combine challenges. Maybe he'll use Forth to do a figure 8 around two juicecans.

    Actually, Martin_H may be close to done already. He used his wall-following robot in the "oval challenge" to good effect. Martin_H, I expect a video from you circling around a juice can this week.

    Stepper motor robots are exempt from this contest, as they can carve perfect circles all day. Same with wheel encoders. I'll say you can use either one ultrasonic sensor or one Sharp IR distance sensor.
  • mindrobotsmindrobots Posts: 6,506
    edited 2012-11-26 15:26
    I think the robot should approach the obstacle on a tangent line a t some distance (2-3 feet) and then begin a circle. That way you acquire the target and then maintain the distance as you circle it,
  • ercoerco Posts: 20,244
    edited 2012-11-26 15:37
    XLNT idea. Advanced challenge could use two cans, running an oval around them...

    Keep those cards & letters coming!
  • Martin_HMartin_H Posts: 4,051
    edited 2012-11-26 16:15
    I'm in, and yes my wall following bot could probably be adapted to this challenge fairly readily.
  • TtailspinTtailspin Posts: 1,326
    edited 2012-11-26 21:11
    Uncanny... Just this very morning I dug four pages back to find this very thread...
    When I left for work, I was thinking of clever responses to peer pressure.
    and now, when I get home from work, I need some kind of IR soup can just to keep up with you kids...sheesh..

    I did want so badly to have some awesome video of my stock Stingray running a perfect figure eight around a pair
    of conveniently placed manhole covers on the street out front of my house.
    Only thing is, I can not seem to gain control of this beast, I think encoders will help some, but the hard plastic of
    the Omni wheel is not ideal for asphalt pavement running, Especially at the speeds I was hoping for.

    I do believe the STOCK Stingray is perfect for a Gymnasium floor, or any large SMOOTH surface.
    But when I try precise maneuvers on black top, the Omni wheel bounces off of every irregularity and small pebble.
    Which means, Unless one can keep track of the height,direction and the number of times the Tail wheel leaves the ground,
    One will never be able stop the stock stingray in the exact spot it started, Sooo...

    Some changes will be needed...

    I have started making the necessary "corrections" to the Tail wheel assembly, and in fact have it working, however,
    for cosmetic reasons, I am starting a search for a smaller, sleeker, more "Stingray like" rudder servo...
    I want a smaller servo body then the standard size parallax carries. Just slightly smaller, about 1/4" shorter in length.
    of course, it needs to also have a similar torque rating to the standard servo... does such a servo exist?

    As soon I can find one, I will be one step closer to precision six foot per second figure eights...:thumb:

    -Tommy
  • ercoerco Posts: 20,244
    edited 2012-11-26 21:31
    Yep, I'm on record as anticipating that the Stingray's nifty looking tailwheel would be problematic, even on smooth surfaces. When rolling sideways, it's essentially a 1/2" diameter plastic roller at best. A nice 2" or bigger rubber caster would be smoother, but of course the caster pivot must be installed vertical to the ground, not simply perpendicular (normal) to the robot's bottom plate, which might not be parallel to the ground.

    Same for a rudder servo axis, Boss!
  • rwgast_logicdesignrwgast_logicdesign Posts: 1,464
    edited 2012-11-28 14:12
    @Martin lol I acually was just browsing through here to see some of your guys robots doing this thinking this is probably one of the first things I should get my robot doing, Im going to have to read through the whole thread first to get the rules down.
  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-11-28 15:41
    Im going to have to read through the whole thread first to get the rules down.

    I think the only rule is "no apologizing".
    erco wrote: »
    stop apologizing already!

    So even if your robot is under remote control we want to see it. (Just tell use if it is remote controlled.)

    I hope to have another video to add soon.
  • ercoerco Posts: 20,244
    edited 2012-11-28 17:21
    Duane Degn wrote: »
    I hope to have another video to add soon.

    Def want to see your mecanum robot "doin' da 8".

    Hmmm. Never seen W9GFO's mecanum trying it either. Rich?
  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-11-28 17:57
    erco wrote: »
    Def want to see your mecanum robot "doin' da 8".

    Hmmm. Never seen W9GFO's mecanum trying it either. Rich?

    You do remember my earlier attempt?

    Now that I've fixed the 45 degree only bug, it should be able to complete a more rounded "8".
  • rwgast_logicdesignrwgast_logicdesign Posts: 1,464
    edited 2012-11-28 18:22
    Duane i originally looked into mecanum wheels and now that I watched your video along with another using steel version im pretty sold on using them on this or another robot at some point, that video was really cool. Im just wondering as an owner of the wheels do you think that dirt would mess them up, would they be good off road wheels?
  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-11-28 19:23
    Duane i originally looked into mecanum wheels and now that I watched your video along with another using steel version im pretty sold on using them on this or another robot at some point, that video was really cool. Im just wondering as an owner of the wheels do you think that dirt would mess them up, would they be good off road wheels?

    The Vex wheels are great. I haven't used my Mecanum robot outside but I think I saw of video of Rich driving his in the parking lot at Parallax. I doubt they'd do well in dirt.
  • W9GFOW9GFO Posts: 4,009
    edited 2012-11-28 19:33
    The Mecanums will go through the grass just fine. Think of them as regular wheels with big knobs. Same with dirt. In challenging terrain the problem will be strafing. Straight ahead or regular differential steering type turns work just like with any other wheel.
  • W9GFOW9GFO Posts: 4,009
    edited 2012-11-28 19:59
    erco wrote: »

    Hmmm. Never seen W9GFO's mecanum trying it either. Rich?

    Kepler was built with the intention of performing complex path following. Now only if someone would write some code so that it could do that...
  • rwgast_logicdesignrwgast_logicdesign Posts: 1,464
    edited 2012-11-28 20:00
    Well I was looking at the vex ones, unfourtantely i live in the deseret no grass all sand. I was afraid of sand getting in the crevices and running the wheels does that seem like it could be an issue?
  • W9GFOW9GFO Posts: 4,009
    edited 2012-11-28 20:09
    I would avoid soft sand, the rollers may get jammed up. Hard sandy surfaces should be okay. You can always wash them out.
  • rjo__rjo__ Posts: 2,114
    edited 2012-12-03 20:22
    What is it with you guys... I finally get my peanut butter wheels to stay on... start thinking about figure 8's and you go and change the rules.

    So... my next question is ... can I cheat?

    Rich
  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-12-03 20:43
    rjo__ wrote: »
    What is it with you guys... I finally get my peanut butter wheels to stay on... start thinking about figure 8's and you go and change the rules.

    So... my next question is ... can I cheat?

    Rich

    I think the rules stated in post #139 still stand.

    And yes you may cheat as long as you tell us what the cheat is (i.e. remote control).

    I think the entry in post #106 was a bit of a cheat since the figure 8 wasn't programmed.

    As I mentioned earlier, I'm still planning on making several more entries to this challenge.
  • rjo__rjo__ Posts: 2,114
    edited 2012-12-04 09:18
    fair enough:)
  • GarethGareth Posts: 278
    edited 2012-12-13 14:53
    Here is my first attempt to take the title.... much training was required ..... not the result i wanted ... however robot had other things on its mind .
    The tighter the clearance the better (mine was 2mm from wheel to bottle)
    In my case the robot started out on course and completed the first ring.....
    .... however at the start the robots voltage regulator seems to have clipped the bottle and thus threw it ever so slightly off course .........
    .... end result is the robot decides to "Toast" me instead.......good robot.......
    It just goes to show how powerful the motors are.
Sign In or Register to comment.