3-wheeled robots

[Gearheads1]

I'm new to this and am still learning lots of things. One possible drive configuration I've been thinking about is two front wheels, each independently powered, and a third wheel on a caster at the hitch. I thnk that this could provide better traction and manuveurability than some othe configurations. Balance would be an issue to ensure more weight was on the front wheels. Could anyone please point out things I'm not considering well? Thanks!

[EricVanWyk]

Originally Posted by Gearheads1

Balance would be an issue to ensure more weight was on the front wheels.

Sounds like it could make for an interestingly shaped robot. I've highlighted a small portion of your post, and you are right. Every newton that the caster supports is mu newtons of traction lost.

[Daniel_LaFleur]

Originally Posted by Gearheads1

I'm new to this and am still learning lots of things. One possible drive configuration I've been thinking about is two front wheels, each independently powered, and a third wheel on a caster at the hitch. I thnk that this could provide better traction and manuveurability than some othe configurations. Balance would be an issue to ensure more weight was on the front wheels. Could anyone please point out things I'm not considering well? Thanks!

I would consider powering that caster wheel (you could keep it as a caster just put a motor on it) because if you don't you will be losing precious traction due to the weight that the caster wheel is supporting.

[taberski]

We actually considered a very similar design last night, but the team ultimately voted against this configuration.

Our thought was to think of it more as a 2 wheeled robot, with the center-line of the wheels just forward of the center-of-mass. Then, the caster would be very lightly loaded - the thought was that the tongue weight of the trailer, plus the moment provided by the hitch, would provided the required stability.

What we envisioned was basically an articulated design - similar to a scraper (earth mover), but with a supporting caster.

In the end, we too felt that with such limited traction, all wheels should be powered and felt that implementing a powered caster was more than we wanted to take on.

I'd like to see it done!

Kevin

[writchie]

You could configure this so that when the trailer is attached your "castering" wheel is just off the ground. But I can't think of any good reason why it needs to be castered or powered. Your design cg should be close to the pair of driven wheels. You will probably need another undriven wheel forward that also remains off the ground under normal conditions.

when you hitch the trailor, it's tongue weight will move your CG aft. The hitch will add to your stability. Nearly all of the normal force from your robot mass will be on the wheels and some from the trailer as well.

[cknight]

3 wheels? I don't think it will work.

http://www.chiefdelphi.com/media/photos/23651

http://www.chiefdelphi.com/media/photos/28166

http://www.chiefdelphi.com/media/photos/31681

http://www.baxterbombsquad.com

[EricH]

Originally Posted by cknight

3 wheels? I don't think it will work.

http://www.chiefdelphi.com/media/photos/23651

http://www.chiefdelphi.com/media/photos/28166

http://www.chiefdelphi.com/media/photos/31681

http://www.baxterbombsquad.com

You forgot Tumbleweed (148's 2008 robot)!

If you do use a caster and power it, you may want to power the turning too. Think of it as a rudder.

[Matt H.]

Here's a problem that I don't believe anyone has mentioned thus far.
The first and last robot we used caster wheels on (2006) was highly unstable. The problem was that after rounding a turn the rear caster wheels would keep right on going causing our robot to fishtail. It cause numerous headaches esp. during autonomous.

In a low friction environment the fishtailing due to a caster wheel will be exaggerated and might make your robot undrivable depending on weight distribution.

[David Brinza]

Originally Posted by cknight

3 wheels? I don't think it will work.

http://www.chiefdelphi.com/media/photos/23651

http://www.chiefdelphi.com/media/photos/28166

http://www.chiefdelphi.com/media/photos/31681

http://www.baxterbombsquad.com

I think some folks might not catch onto the sarcasm in the above post.
Team 16's "Two-minute warning" was an awesome robot last year.
They played in the finals on Einstein and were the #1 seed at IRI.

Oh yeah, I kinda liked their driver too !

I think robots with less than 4 wheels might do very well this year...

[shannieliz]

My thinking is that if you have three wheels, wouldn't it tip?
If you have the two wheels in the front, and one on the back?

Think of it as a two wheeled robot. It would sway from side to side because of all the weight.



Don't you think?

[EricH]

Originally Posted by shannieliz

My thinking is that if you have three wheels, wouldn't it tip?
If you have the two wheels in the front, and one on the back?

Think of it as a two wheeled robot. It would sway from side to side because of all the weight.



Don't you think?

See above-referenced robots.

[Meredith Novak]

Originally Posted by shannieliz

My thinking is that if you have three wheels, wouldn't it tip?
If you have the two wheels in the front, and one on the back?

Think of it as a two wheeled robot. It would sway from side to side because of all the weight.



Don't you think?

We used some variation of a 3-wheel drivetrain in 2003, 2006, 2007, and 2008. No tips.

Edit: Actually, that is not exactly true. In 2003, our robot screamed around and up the ramp on two wheels to unstack the totes in about 6 seconds during automous. But that was tipping on purpose!

[patrickrd]

Regarding stability (propensity for robot to tip over), you should see no increased tipping in cases where there is no lateral (sideways) acceleration.

You can think of this geometrically by drawing lines between your three wheels (a triangle). Next, draw an arrow in the opposite direction as you would like the robot will accelerate. You will see in the forward and backwards direction the arrow can be equally long before it exits the triangle (and thus it can accelerate in these directions with equal magnitudes before it begins to tip). But once you also have a lateral component of acceleration, there are directions where the arrow cannot be the same length. For these shortened vectors (again indicating acceleration in the opposite direction), you cannot accelerate as quickly before your robot tips.

Practically, this means that whenever you are making purely longitudinal (forwards or backwards) accelerations, you will not see any difference in how fast you can go before you tip. However, when you are making lateral accelerations as well (such as turning while simultaneously moving forward or backwards), you will lose stability in certain directions.

Further, I noticed you want to put the driven/powered wheels in the front. Keep in mind that when you accelerate in the forward direction, you are simultaneously taking weight off the front wheels, thus reducing your traction. It would make sense to put your powered wheels on: (a) the heavier side of the robot, and (b) the "rear" side behind the direction you will most often accelerate.

Hope this helps.

[Lesman]

And having that trailer in the back will help with stability, your base might actually look something like this:

|.................................|


.................|


............|..........|

if you have a wide front configuration.

[Ziaholic]

But the hitch-swivel in the tongue of the Trailer can rotate, so it does not provide stability for tippiness to the robot. It does provide some types of stability and/or weight to the bot, but you cannot depend on the trailer's wheel-stance to reduce tippiness.

(i.e. if it tips over, it does not induce a torque on the robot, the hitch-swivel simply rotates.)


For reference, see the 0900 (Trailer) and the 0936 (Hitch Swivel) drawings