Shop OBEX P1 Docs P2 Docs Learn Events
One-Servo Robot Challenge — Parallax Forums

One-Servo Robot Challenge

ercoerco Posts: 20,244
edited 2015-06-15 20:50 in Robotics
How about a one servo (or one motor) only mobile robot? I have a concept I'll be trying soon. Inviting others to start thinking and building.

It's a no-brainer to do a wheeled robot doing J-turns: a continuous rotation servo or motor drives robot forward in one direction, and spins robot in the other direction. So the bot either goes straight or turns in one direction. But I'm thinking of a bipedal weight shifting walker. Maybe limper. Will see.

If I add a requirement that it's gotta do a figure 8, then the J-turn bot is mostly out, unless the mechanism only selects curve left and curve right. That's cheating.in my book. So it has to be able to drive straight too.

Anyone else game? Start thinking, Duane, Martin_H, Amanda, Gareth, Whit, Rick, Rich, Matt_G, Nikos and everyone else! ( I know I forgot some people...)

Let this 3-servo bot inspire you: https://www.youtube.com/watch?v=ej1_BWnmpgg

That music is second only to Cliff Richard's "Congratulations"!
«1345

Comments

  • Martin_HMartin_H Posts: 4,051
    edited 2015-01-12 13:14
    I'm interested, but completely befuddled how to do a figure eight with only one servo.
  • ercoerco Posts: 20,244
    edited 2015-01-12 13:25
    Genius starts with befuddlement, Mister First Responder!

    Please promise me that this challenge will not in any way become the object of your fixation. I expressly forbid it!
  • WBA ConsultingWBA Consulting Posts: 2,933
    edited 2015-01-12 15:16
    I may not have the adequate "walking mechanism" knowledge to explain this right, but a single servo robot that can do a figure 8 should be this easy:

    The servo needs to drive two actions for each side of the "walking mechanism". At dead center, 90 degrees, the servo is in "neutral". When the servo is driven to the "left" or 0 degrees, two actions need to occur. One is the action of shifting the center of gravity to the right foot and the second is to pull the left foot forward. Same goes for driving the servo to "right" or 180 degrees, but shifts weight to the left and pulls the right foot forward. My knowledge of these mechanical actions is limited to what I have seen on Youtube or project sites, so I can't explain it more nor draw it properly to show the lever and cam actions that would need to occur, but I can visualize it mostly in my head.
    The action of driving the servo from neutral to either side and back to neutral needs to engage the movement mechanisms through a complete cycle. In other words, going from 90 to 0 and back to 90 should mean the robot goes from a balanced stance, leans to the right, moves the left foot forward, and then balances again.

    attachment.php?attachmentid=112779&d=1421104327

    Once the mechanics of that nature are in place, the robot's movements are pretty simple to make up:

    Servo positioning for actions:
    Walk forward: 0, 180, 0, 180, etc (left, right, repeat)
    Walk slight left: 0, 90, 0, 180, 0, 90, 0, 180 (left, left, right, repeat)
    Spin left: 0, 90, 0, 90, 0, 90, etc (left, left, etc)
    847 x 768 - 102K
  • W9GFOW9GFO Posts: 4,010
    edited 2015-01-12 16:43
    Think how a canoe is driven (paddled). Only one side at a time, for various stroke lengths/directions. One servo could do that pretty easily I think.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2015-01-12 16:51
    W9GFO wrote:
    Think how a canoe is driven (paddled). Only one side at a time, for various stroke lengths/directions. One servo could do that pretty easily I think.
    I agree. Lever-actuated ratcheted wheels would be one way to do it. The really tricky part would be how to make the bot go in reverse.

    -Phil
  • ercoerco Posts: 20,244
    edited 2015-01-12 17:44
    The really tricky part would be how to make the bot go in reverse.

    Reverse isn't a requirement. Straight, left and right turns are. Could be curving arcs or turning in place to the figure 8.
  • xanaduxanadu Posts: 3,347
    edited 2015-01-12 19:02
    Tri-copter with tail servo.

    What do I win?

    Edit: Oh dang it says one motor. I get it now...
  • ercoerco Posts: 20,244
    edited 2015-01-12 20:05
    Ah xanadu, I'm sorry I left you off of my rushed call out list, but you're certainly a capable builder too.

    Also Tommy Tailspin, Chris L8, zappman, Paul K, vanmunch. And so many others!

    @Tommy: One servo is JUST a rudder. What can you do?

    Who knows, this might lead to another viable robot for Parallax to sell. Start writing your "Robot of the Year" acceptance speech now.
  • xanaduxanadu Posts: 3,347
    edited 2015-01-12 20:46
    Thanks for the mention! There was a small robot made a few months ago that used the acceleration of the motor timed with flexible legs and rigid legs to create something like what you're talking about. Memorable article. http://spectrum.ieee.org/automaton/robotics/robotics-hardware/how-to-make-a-steerable-robot-with-just-one-single-motor

    I guess timing and acceleration are one very good approach. For me it would be easier to develop some kind of clutch between the wheels and motor. I have some ideas, but would the clutch be considered extra motors?

    There's a magnet in development where an electrical pulse can change its poles. It will be neat to see if that is used as a clutch. You could have one motor, and pulse the magnetic clutches to control their friction. Sounds almost as complicated as encoders and timing though.

    hmmm...
  • ercoerco Posts: 20,244
    edited 2015-01-12 21:03
    Clutches, gears, mechanism, all doable but overkill IMO and certainly no fun to prototype. I'm thinking simple, all momentum and weight shift. Tilted axis servo shifts battery weight L/R, CW/CCW. Feet might be Hexbug or bristlebot-style "clutches" that grip better one way. Control is by how far and how fast the servo moves & shifts the CG.

    IIRC, didn't PhiPi have something like this with a tilted axis servo a while back?
  • Tim-MTim-M Posts: 522
    edited 2015-01-12 21:10
    Brings back lots of memories Beau...

    I used to repair consumer electronics for a living and saw hundreds and hundreds of VCR's come across my bench, mostly needing rubber parts like the idler wheels and belts central in your photo. That is a clever mechanism providing a lot of function from a simple single motor.

    Now back to your original thread programming and topic...
  • W9GFOW9GFO Posts: 4,010
    edited 2015-01-12 21:32
    Have you seen those walker toys that will walk down a slightly inclined ramp? You could use one servo for one of those to gently sway a weight side to side. Then to turn you would quickly accelerate the weight in a particular direction.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2015-01-12 22:05
    erco wrote:
    IIRC, didn't PhiPi have something like this with a tilted axis servo a while back?
    Two servos, Mark Tilden-style:

    [video=vimeo;71207671]

    -Phil
  • ercoerco Posts: 20,244
    edited 2015-01-12 22:31
    Now we're getting somewhere, fellas. Keep those cards & letters & ideas coming.
  • TtailspinTtailspin Posts: 1,326
    edited 2015-01-12 23:02
    Thanks erco, that means alot, Just wish I was in the same league as yourself, xanadu, vanmunch, Et al...(forget naming them all, You're all my hero's...)
    The only thing I've got that is even a little close to what you are describing, is this thing I built for a "white Elephant" gift exchange for Xmas this year...
    This Unfinished Prototype is for a "State of the Art, Table Top Flight Simulator." is just screaming for some Electronics to be installed. Just waiting for some free time to get back into the fun...
    I have been swamped with "Dream Jobs"(one offs I can't believe I have been commissioned to complete.) Some of my best work too..

    Sorry, gonna have to sit this one out for a few more weeks. :(

    Thanks again erco. You Da Man!.



    -Tommy (Casters are for Suckers) Tailspin.
    1024 x 1365 - 191K
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2015-01-12 23:11
    Tommy,

    Wait a minute! You're not You weren't [I see the business has been purchased] this guy are were you?

    whose Cessna 152 Aerobat I once had the "pleasure" of almost throwing up in? (You kept ashes in the ashtray so you could tell when the rental pilot pulled negative G's!)

    -Phil
  • NikosGNikosG Posts: 705
    edited 2015-01-13 05:57
    I would say a robot similar to the "Penguin Robot ". But looking at Penguin Features: I realized that penguin has two servos and not one!
  • TtailspinTtailspin Posts: 1,326
    edited 2015-01-13 07:54
    No sir, that is not my Airplane.

    But, Uhm, errr... So you got into an Airplane with someone that uses the name Tailspin in the company name?
    That's kinda like playing cards with a guy named "Doc" isn't it? :)



    -Tommy
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2015-01-13 08:12
    Martin_H wrote: »
    I'm interested, but completely befuddled how to do a figure eight with only one servo.

    HINT -- I suspect the car you drive everyday is the equivalent to a one-servo robot. After all, it has merely one engine.

    Of course, a clever fellow might come up with several other solutions. I think that something similar to rowing a boat might work out. To turn, you just do more strokes on one side.

    On the other hand, there is a simple one-servo robot that I built years ago... almost forgot. It requires for wire legs. It was a BEAM project.

    http://www.instructables.com/id/How-to-build-the-one-motor-walker/

    http://www.solarbotics.net/library/circuits/bot_walker_uCrawler.html

    Sorry if I spoiled someone's fun.
  • Duane DegnDuane Degn Posts: 10,588
    edited 2015-01-13 09:19
    Sorry but I'm sitting this one out. IMO, a robot really needs to have at least six servos to deserve the name "robot". (I am joking (mostly).)

    Let's Make Robots recently had a single motor robot challenge. This was the winning entry.
  • W9GFOW9GFO Posts: 4,010
    edited 2015-01-13 12:44
    HINT -- I suspect the car you drive everyday is the equivalent to a one-servo robot. After all, it has merely one engine.
    One motor AND one servo (the driver which turns the steering wheel).
  • kwinnkwinn Posts: 8,697
    edited 2015-01-13 13:17
    Duane Degn wrote: »
    Sorry but I'm sitting this one out. IMO, a robot really needs to have at least six servos to deserve the name "robot". (I am joking (mostly).)

    Let's Make Robots recently had a single motor robot challenge. This was the winning entry.

    Clever, but could be done better and with smoother motion using a hexapod.
  • Martin_HMartin_H Posts: 4,051
    edited 2015-01-13 17:57
    W9GFO wrote: »
    One motor AND one servo (the driver which turns the steering wheel).

    Bingo, that's why I find this so hard. Each degree of freedom usually requires a motor or servo.

    I can imagine a geartrain which turns the drive wheel side to side, as well as providing forward power. With proper tuning such a mechanism could probably do a figure eight. However actually building such a machine is not something I'm good at.
  • ercoerco Posts: 20,244
    edited 2015-01-13 19:21
    Heck, if it was easy, anybody could do it. Which is why I've recruited you cats, the best & brightest. Duane, I'm not letting you sit this one out. Nice try though.

    You CHICKEN, McFly?

    I hope to see some entries in a week or so, or I'll have to add a "solar powered" requirement too!
  • W9GFOW9GFO Posts: 4,010
    edited 2015-01-13 19:26
    You could also drive a robot like a fish. Have a bristlebot type thing mounted on the end of an arm attached to the servo. The servo would vibrate back and forth a couple of degrees to provide the propulsion. To steer move the servo to one side while still vibrating.

    With the right type of attachment you might even be able to go in reverse by varying the frequency of the vibration.
  • ercoerco Posts: 20,244
    edited 2015-01-13 19:48
    Great call, Rich. I was thinking along those lines with a skateboarder metaphor. Cycling through the servo's middle range provides straight ahead propulsion (perhaps like a railroad handcar, you might also get fwd/reverse with phasing), and to turn, you coast for a bit while the servo goes to either extreme, which tweaks the steering linkage L/R.
  • kwinnkwinn Posts: 8,697
    edited 2015-01-14 07:23
    Pretty hard to beat two wheels with sawtooth gears and a servo for simplicity. Rotate left to turn left wheel, right to turn right wheel. Alternate left and right to go forward, do more left than right to turn right, more right than left to turn left.
  • ercoerco Posts: 20,244
    edited 2015-01-14 07:56
    Love it, kwinn!

    Video below is for popsicle stick loving-Duane and Hexapod fans. No turning or control, but it's inspiring since it has such nice articulation and only uses one motor. Paper clip legs and linkages. Shorter video at https://www.youtube.com/watch?v=hO9pCjlSmso

    Those purple batteries look like the cheap Ebay NiMH cells I posted about long ago. :)
  • WBA ConsultingWBA Consulting Posts: 2,933
    edited 2015-01-14 09:10
    erco wrote: »
    Love it, kwinn!

    Where's my love Erco?? I posted the exact same idea 2 days ago in post #4.. now, where's that crying emoticon.............
  • ercoerco Posts: 20,244
    edited 2015-01-14 09:43
    Where's my love Erco?? I posted the exact same idea 2 days ago in post #4.. now, where's that crying emoticon.............

    OMG Andrew I did completely miss your post. Spot on, my bad, you nailed the left/right wiggle walk early on in great detail. Apologies, I defer to your venerable expertise and considerable body of work.

    So it should be a simple matter for you to build your creation and post the first video...? :)
Sign In or Register to comment.