Shop OBEX P1 Docs P2 Docs Learn Events
Segway type balancer... — Parallax Forums

Segway type balancer...

quick questionquick question Posts: 50
edited 2005-07-31 15:21 in General Discussion
Has anyone been able to use a MEMSIC sensor to make a two wheeled "balancer" with a BS2 or BOE?
I saw it was done with a Lego mindstorm kit.
·

Comments

  • Paul BakerPaul Baker Posts: 6,351
    edited 2005-07-14 18:28
    David Goodsell submitted a self balancing robot for the SX design project, his submittal is availible here. It is based around the SX, and uses an optical angle sensor instead of the accelerometer. I think the BASIC stamp should be able to do something similar using the accelerometer, but I couldn't say for certain.
  • Chris SavageChris Savage Parallax Engineering Posts: 14,406
    edited 2005-07-14 19:55
    Hello,

    ·· The height (Well, center of gravity) of the bot as well as the type of drive system will also affect the balancing system.· So far steppers have been the most widely used motors I have seen for this kind of thing, but that doesn't mean a DC motor won't work equally as well with proper design consideration.· Servos may be too slow unless you have a well-balanced bot.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Chris Savage
    Parallax Tech Support
    csavage@parallax.com
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2005-07-29 17:10
    Chris' robot balances a stick on four wheels. Is that what you had in mind.

    You say 'two-wheeled'. So, I imagine that you have seen the two-wheel contraption that used two IR distance measuring devices to keep upright. It seems to be able to adapt to radio control of movement.

    The problem with the pure optically balanced design is that it cannot handle sloped ground.

    Assuming you combine both the IR sensors from the optical design and the Memsic, you could have the Memsic benchmark vertical within a given limit [noparse][[/noparse]one that would allow the robot to keep moving] and have the IR sensors handle the swing [noparse][[/noparse]or plugoid motion] by adjusting one to the high side and one to the low side.

    I would really love to build one of this and have it r/c controlled.

    Above all put the batteries low as they will give you the lowest center of gravity and allow you to build for more height. I don't think steppers are as good as geared-reduction DC motors with big wheels. They have some special problems, such as pulsing at certain speeds. Also, I think they use more power.

    Bicycle trainer wheels are very cheap and light. I have my eyes on a 18 long rectangular wire basket for a light weight chassis. It just need to get the wheels adapted from 1/2" holes to a 6mm motor shaft.

    Anyway, both are quite eye catching as they seem to defy being mundane machines.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    G. Herzog in Taiwan
  • quick questionquick question Posts: 50
    edited 2005-07-29 17:43
    looks like it has been done!

    From the great customer applications page .....

    http://www.parallax.com/html_pages/resources/custapps/app_bway.asp
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2005-07-30 05:59
    Actually it only uses the IR and it seems to use the cheaper Ir phototransistors rather than the Sharp Ir sensors.

    Can it handle hills and slopes?

    That is why you may need the Memsic or more than one Memsic - one for vertical, possibly another for acceleration control.

    The article actually states that they want to do more research about this.

    Additionally, the photos show the batteries at the top end which makes the programing and design more of challenge as the center of gravity is higher.· It also demands more power to maintain.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    G. Herzog in Taiwan
  • Stanley FredericksonStanley Frederickson Posts: 19
    edited 2005-07-30 06:46
    A lot of people place the weight as high as possible on balancing robots so that they don't need as tight of a control loop. Adding more mass to the top of the robot will help dampen out the oscillations. Try balancing a hammer on one finger under the small end of the handle and then under the more massive head of the hammer. You will have to move your hand around a lot more when the bulk of the mass is down by your finger rather than when it is at the top. I also notice this on the Segway. If you don't have the bag on the handle bars, the handle bars tend to shake when you try to ride with no hands on the handle bars. Adding the bag, dampens this oscillation...so does riding the Segway properly with your hands on the handle bars [noparse]:)[/noparse]
  • ManuelManuel Posts: 105
    edited 2005-07-31 00:26
    That project looks cool, the legomindstorms and IR. The guy who made it is my friend. It is a good project, but sometimes it didnt sense well and it required a white floor. Its cheap, but IR is not an option to make a segway.
    BEst Regards
    Manuel C. Reinhard
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2005-07-31 15:21
    I guess I just assumed that all robots are better off with a low center of gravity. But now that you mention it, the weight at the top is coupled with the mass of the wheels and the motors. So, as long as you don't go to far from plumb, it does make sense that it can dampen.

    I was thinking that the balancing was centering around the wheel axle. But if the two masses are coupled and equal, the center is halfway inbetween.

    Very interesting. I'll have to find the time to build my own version. I have the Memsic and I just got 6 Sharp IR sensors.

    There is a lot of mechanics that should be discussed more in robot design. It can't all be done with big motors and faster processors.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    G. Herzog in Taiwan
Sign In or Register to comment.