Amazing Hemispherical Omnidirectional Gimbaled Wheel Robot


Bradley University grad student [Curtis Boirum] has built a robot which uses quite a unique drive system, one we’re guessing you have never seen before. The robot uses a single motor to drive its hemispherical omnidirectional gimbaled wheel, propelling it across the floor at amazing speeds with uncanny agility.

The robot uses a simple two axis gimbal for movement, which houses a small brushless RC airplane motor. The motor spins a rubber wheel at high speeds, which propels the robot in any direction at the flick of a switch, thanks to a pair of RC servos. When the servos tilt the gimbal, they change which side of the wheel is touching the ground as well as the gear reduction, eliminating the need for a mechanical transmission or traditional steering mechanism.

While he originally thought that he had invented the concept, [Curtis] found that this technology was nearly 100 years old, but that most people had forgotten about it. We’re pretty sure people will remember it this time around. How could you not, after watching the demo video we have embedded below?

We think it’s a great concept, and we can’t wait to see what other robot builders do with this technology.

[via Gizmodo]


33 thoughts on “Amazing Hemispherical Omnidirectional Gimbaled Wheel Robot

  1. darn, I experimented with this system when I was 6 and I made some spinning tops with some dc motors but at the time I never thought it was practical or could be utilized. That’s awesome

    1. The practicality of this is limited only by your imagination!

      I envision these as being very useful for factory or warehouse robots where two or more of these drives render a vehicle with true omnidirectional steering. Very nice indeed!

  2. At a glance it looks like the ball would wear quickly, and one point of contact would give little traction.

    Otherwise, it’s actually quite awesome, and with 2 or more of these ball-wheels the potential for omni-directional movement is akin tho the much more complicated omni-wheel design.

  3. A far out unintentional reinvention. 74 years ago didn’t have the tech to easy make a functioning model easily. The real world isn’t all smooth surface with the occasional runner on the floor that’s designed not to impede wheels. A whole lot of development needed. I’d think the majority of mobile robot applications require agility, and precision, with speed somewhat a luxury. Do robot events feature a gymkhana? Where their agility precision is testing first, and how quickly they can be agile, and precise second? Not really seeing this being used for cargo or passenger motor vehicles. AFAIK CVT that use a similar concept haven’t yet made inroads at all. The young man is an engineer, I doubt I mentioned anything he doesn’t already realize.

  4. Very nice model, however it appears that the model has little to no directional control of any degree of precision which would be critical for combat bot, well any true practical application for that matter.

    However for a college paper it is the perfect thing to get that credit and an A which I do believe the creator deserves dully!

  5. This probably isn’t the most practical drive for, say, a military bomb bot or a typical house bot. Then again, it looks pretty ideal for factory work.
    Smooth surfaces? Check
    Requires maneuverability? Check
    Industry that uses a lot of robots? Check

    @Munden The heavy load could be worked around by using 3 or more omnidirectional heavy-duty wheels and using 2-3 of these for the actual drive. The torque on the shaft would be only the force needed to change velocity, and would be perpendicular to the axis of rotation.

  6. Again @Hackerspacer:

    My Q: Why wouldn’t it perform on uneven terrain?

    Your A: Run the spinning wheel into an obstacle and you will quickly find out.

    Either i am too optimistic (i usually am) or you are pessimistic. I fail to see the difference between an ordinary wheel and this wheel?

    Hitting an object while moving in any direction the would perhaps increase the speed while moving over the object. This would probably interfere with steering but then again hitting a bump with ordinary wheels could interfere withe steering as well (decreasing the speed).

  7. Brilliant idea, but it seems like an inefficient method of travel. Not much surface area on the wheel, means lots of slippage, which in turn means premature wear on the ‘tire’ and reduced acceleration speeds, not to mention the motor practically being belted at full speed fairly consistently. My 2c. Having said all of that, I now want to go build one :) looks like fantastic fun

  8. Many years ago someone developed a vehicle that had 4 of these on each corner. they used an aluminium dome coated in rubber and was able to drive in any direction rotate and it performed well in water, in sand and on other hard surfaces.
    awsome system and I often wondered why nothing came of it.
    From memory the drive motors were hydralic.

  9. @WitchDoc a normal wheel when hitting an object won’t suddenly with the same force as before propel it in an arbitrary direction.

    This setup although nice does really not seem to be that good in going in a straight line anyway, it’s more a 2D fly (the insect) movement simulator :)

  10. @WitchDoc

    Maybe this will help:

    Visualize and ordinary wheel. The only part likely to touch the ground (regardless of how bumpy) is spinning on one direction. This “hemispherical thing” has surfaces spinning in all directions all very close to the ground. If you were unfortunate enough to hit a pot hole, there’s no telling which direction you might end up going. To say the least it would be a jarring experience. To say the worst it could be deadly.

  11. I am thinking this type of drive is best kept in a controlled situation. Where the surface contact is very predictable. Such as in this type of transmission (for a bike evidently – cool):

    It does occur to me this drive makes for a very responsive vehicle. So much so I doubt one could really control it well. Which leads me to suggest a much different control paradigm.

    Instead of steering it, why not tell it where to go. I know the author is working on a Masters of Mechanical Engineering, and is rather proud of the fact this is a totally mechanical design with no need for a processor. But thinking how fast a multiple drive platform like this could maneuver, it would almost be better to supply start and ending points along with obstacles to a processor.

    Think of a classic robot war game. With this type of drive you wouldn’t need armor. The robot would simple get out of the way. And with a processor involved, getting out of the way can also mean maneuvering into attack position.

  12. Here I was about to suggest that one way of solving the issues with uneven terrain would be to put this on top of a spherical ball (so that the drive only touches the smooth ball, and then the ball transmits the motion to the ground) — and I see that tech-no-pest has just suggested the same idea.

    Another advantage of that arrangement is that the ball can support the loads, so that they’re not being supported on the drive mechanism.

  13. I to came up with this when I was a kid, after thinking it through i was unable to come up with a practical application.
    After seeing it again here it seems to me it may be just the thing for single wheeled balancing robot, however that too seems to defy practical application.
    The problem of high zero speed friction can be dealt with by adding a button to the center of the wheel mounted in a thrust bearing, so that no friction occurs at zero speed. Flattening out the wheel into a smaller section of larger Sphere will increase load capacity and decreasing the speed the wheel turns will increase precision at the expense of top speed.
    However it remains obligated to flat, hard surfaces and in my opinion a solution in search of a problem.

  14. @Matt, “not to mention the motor practically being belted at full speed fairly consistently”

    It doesn’t have to be that way. Car engines run pretty fast all the time when on. You just speed up the engine and apply the rotation to the clutch to engage the wheels when you need motion. The same could be applied to this no?

  15. Perhaps this might have application in industry. Automated robots that transport things around a factory might benefit from fewer moving parts. Or cleaning robots?

    They’d clearly make awesome indoor RC racing cars. That thing can really move – plus you’ve got built in smooth acceleration, thanks to the variable effective gearing ratio.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.