Help explain how this robot works!

[Mossen]

I found this image of a robot on the web, and I can't find an email address for the person who built it. I was hoping someone on this site could try to explain why the builder would choose to use those belts on top to steer the wheels rather than just drive them directly with a servo mounted on top of the platform? Is there an obvious advantage to this configuration? The website lists them as "timing" belts. What is their purpose?



Thanks!
mossen

[photomark]

There can be many reasons why a builder would choose to do anything a certain way , maybe it was easier to mount the servos like this , maybe there was a height restriction , maybe the builder just wanted to use belts .

[Mossen]

OK, I just thought there might be a good reason for it that I don't understand and might be obvious to someone else. It seems to complicate the design so I doubt the builder "just wanted to use belts".

[VegaObscura]

What's more, why did he use two servos for the front and two servos for the back instead of using a front belt that goes all the way to both wheels with a single servo in the middle, and the same setup in the back?  Did his servos not have enough torque to turn both wheels?

The only solutions I can think of are: (1) he wanted to do what I said above and discovered that his servos weren't up to spec or (2) he thought belts would look cool.

[macdad-]

This is basically how an omni-wheel drive works, except he used the belts to keep the height of the unit to a minimum(Makes it more stable, especially on a omni-wheel drive, with low COG)

[Soeren]

Hi,

I found this image of a robot on the web, and I can't find an email address for the person who built it. I was hoping someone on this site could try to explain why the builder would choose to use those belts on top to steer the wheels rather than just drive them directly with a servo mounted on top of the platform? Is there an obvious advantage to this configuration? The website lists them as "timing" belts. What is their purpose?

Christine (the 'bot) is built for 4 wheel independent steering (and 4 wheel drive of course) to be able to make very sharp turns or eg. drive sideways or diagonally.
You won't find any serious (Danish at least) robot builder using cheap standard R/C servos for competition robots - they're just too darned slow, weak and unreliable.

Timing belts are so named, since the toothed belts are similar to those used in car engines to control the cam shaft timing. The teeth keeps it from sliding, which V-grove belts have a tendency to do, so they're very precise.
Christines brain is a 200MHz ARM9 running Linux and proof of concept lies in the fact that it won the Danish RoboCup 2010 half a year ago.

This is definitely not a beginners 'bot!

[blackbeard]

i can see some pretty compelling ups and downs of all of these things. i'm going to list them to make it easier to think about

Use of timing belts over direct drive from servos

-less strain on the shaft of the servo
-sturdier connection between servo and wheel mount
-less hight

use of belts instead of gears

- less precision needed in construction
-less chance of "skipping"
-more torque on a 1:1 ratio

use of 2 servos instead of 1 for front and back

-less strain on difficult terrain
- ability to turn more accurately
- reduces hight (belts would have to be on top of each other)
-more symmetrical

[Soeren]

Hi,

What's more, why did he use two servos for the front and two servos for the back instead of using a front belt that goes all the way to both wheels with a single servo in the middle, and the same setup in the back?  Did his servos not have enough torque to turn both wheels?

The only solutions I can think of are: (1) he wanted to do what I said above and discovered that his servos weren't up to spec or (2) he thought belts would look cool.

Oh how little faith you have in other peoples abilities. Let me assure you that you're pretty wrong though.
First of all it's not "he", but "she and he" and both Nana and Torsten are numerous levels above hobby robot builders, both ee's and researchers in automation, so they do know a thing or two about this stuff (and don't have to pay for their stuff, so if they want a larger motor/servo/whatever, they get it). There are hundreds of hours of research going into this version of their 'bot and they have been competing at the DTU RoboCup since 2004, both graduated from DTU (Technical University of Denmark).

This concept is what I refer to as Double Ackermann and when you turn, the inside wheels have to turn to a larger degree than the outside wheels and if you want to go eg. diagonally forward left, all 4 wheels have to turn the same number of degrees, but all in the same direction, hence independant steering. This is an advanced 'bot (as mentioned, it's a 200MHz ARM9 calling the shots), and a far cry from an AVR controlled $50 robot.

[Soeren]

Hi,

use of 2 servos instead of 1 for front and back
[...]
- reduces hight (belts would have to be on top of each other)

Well, no. If the idea was to run from a single drive wheel, a single belt would be quite enough.
But then you'd have much more wheel friction and instability when turning.

[blackbeard]

Hi,

use of 2 servos instead of 1 for front and back
[...]
- reduces hight (belts would have to be on top of each other)

Well, no. If the idea was to run from a single drive wheel, a single belt would be quite enough.
But then you'd have much more wheel friction and instability when turning.

not really. you'd need some sort of wrack and pinion  or something to give it car like steering. this would be more complicated and i believe the poster was concerned at how complex it was to begin with. you could have it steer like a kid's wagon but that would be opening up a whole other can of worms and would make it virtually impossible to control accurately with all the torque on the servo.

[Soeren]

Hi,

Hi,

use of 2 servos instead of 1 for front and back
[...]
- reduces hight (belts would have to be on top of each other)

Well, no. If the idea was to run from a single drive wheel, a single belt would be quite enough.
But then you'd have much more wheel friction and instability when turning.

not really. you'd need some sort of wrack and pinion  or something to give it car like steering.

Oh, you lost me there. I thought you were talking about double height with belts over each other, since that's what you wrote and now you're saying it really meant Rack 'n' Pinion?? 

I assume that by "car like steering" you're referring to Ackermann steering where the inner wheel in a curve is turned more (to describe a curve with a smaller radii) and Rack 'n' Pinion haven't got any influence on that at all. Rack 'n' Pinion is just the way that cars usually (some are going electric) transfers the rotational movement of the steering wheel into a linear movement of the steering arm.
Here is an illustrative article on Ackerman steering geometry that should help you understand the concept.

Without Ackermann (like when turning the wheels the same amount like with the belts over each other), you'd get, as I wrote, "much more wheel friction and instability when turning" for this exact reason.

this would be more complicated and i believe the poster was concerned at how complex it was to begin with.

As I stated, this 'bot is not for beginners, so he shouldn't even consider it.
No point in making any more soup on that stone.

you could have it steer like a kid's wagon but that would be opening up a whole other can of worms

I assume you're talking about what the OP could do or not here?
Not that I see any worms in a cart wheel steering - unless it's using a worm drive since it's the oldest and simplest steering in existence, dating back to when horse carriages went from two to four wheels. It takes a bit more space, but it doesn't come any simpler, the wheels will always be pointing in the right angles and when drawn (by horse or man) from a rod perpendicular on the wheel axle, will always follow perfectly (in regards to the steering angles).

and would make it virtually impossible to control accurately with all the torque on the servo.

How's that??
If you'd argue that torque has anything at all to do with accurate, it would be that it increased it proportionally.
Torque is just a measure of the power available to do the job - it's the feedback that control the accuracy of a servo (as long as the motor has enough torque to do what its controller tells it to).

[blackbeard]

i'm not going to quote little bits to make addressing it easier since i'm tired and was slightly less tired when i wrote my last post. here is my explanations for your points. try to decipher whatever nonsense i'm about to write rather then poking holes in the terminoligylessness of it all.

first off i DID mean stacking belts one on top of one another originally. you said you wouldn't have to use 2 belts and i explained why even though you technically wouldn't have to it would be more complicated and you might as well just use a second servo. unless you have a servo in the middle of some sort of rod with the wheels fixed in position like a kids cart (which i'll get to in a second) you would need something to tilt both wheels in the front and possibly the back in order to allow it to turn. i don't know the technical term but i didn't mean whatever you were talking about with Ackermann steering. i meant that the pivot point would be centered on that wheel as apposed to on the end of a rod somewhere around the middle of the vehicle.

as for it's complicatedness i was merely comenting on the fact that the whole point of this thread was to find out why that bot was as complicated as it was rather then being much simpler. i'm not talking about what the op could or couldn't do, i'm talking about what the builder of that bot did and why.

i lost my train of thought and just filled in whatever came to mind first when i went on about the torque thing. in reality it's not a very good system. i've tried it with go karts, i've tried it with bots and none of those ever turned out well (perticularly the go kart). it's a bad system and if you're going to do that then you might as well not build that bot since it completely changes the way it drives. money spent on getting a motor on each wheel for instance would be half wasted since you wouldn't get the maneuvering ability they offer in this configuration. in short i was commenting on the design not the ops ability to build it.