View Full Version : [Question(s)] Iv'e got a question about quad tracks and turning

08-13-2013, 11:14 AM
How must the 4 tracks on a quad tracked chassis be positioned to be able to rotate? And what if the front and rear tracks are different sizes? For instance, is this robot (https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=8&cad=rja&ved=0CFkQtwIwBw&url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DgLj wfUh1_1w&ei=7FoKUoHGKvTC2gWcyICYCw&usg=AFQjCNE-UepmwwnZJsp3wT0HgfsFndJj7w&bvm=bv.50500085,d.b2I) (link to youtube video) capable of spinning on itself in one spot? or must it turn gradualy because of the ratio between it's length and width?

08-13-2013, 11:51 AM
There is no particular rule. Pretty much all vehicles with tank steering CAN turn on the spot, assuming the necessary power and traction is there. They may tear up the ground while doing it, though.

Tank steering, in general, is somewhat inelegant IMO, for four reasons:

1) It wastes power.
2) It is not very precise, as it depends on slipping.
3) It tears up the ground, and/or the wheels/tracks.
4) On treaded vehicles, it may actually slip the treads off the idlers!

08-14-2013, 06:39 AM
See also: http://www.beam-wiki.org/wiki/Steering_Techniques :veryhappy:

08-19-2013, 07:05 AM
They may tear up the ground while doing it, though.

Like this? http://www.youtube.com/watch?v=7jWOOO0vgu0

08-19-2013, 11:10 AM
I see this issue come up in my classroom as well as in FIRST Robotics. I think of it kind of strange so I hope I can explain it in a way that makes sense without a diagram. If you can find it, there is a book called Robot Sumo I believe that can totally help you understand this.

Think of a robot with two wheels, or even draw a square and put two wheels on the center of two opposing sides. Draw a line coming streight out the front of one of the wheels and the back of the other. Make a circle connecting the two wheels. The center of this circle is where the robot would want to turn.

Now think of a robot with four wheels. Do the same thing as above, only draw two circles. One for each pair of wheels. Each pair of wheels has its own center of rotation... The difference between the centers of the two circles is adds to the difficulty of your robot turning because each pair of wheels wants to rotate around its own center. Tracks or treads would work the same way, only you have (depending on the track design) an infinate number of centers of rotation each fighting against all the others preventing your robot from turning.

Now in the case of a tank or bulldozer, they use power to overcome this tearing up the ground underneith the vehicle...The give in the fighting between all of the centers of rotation is the ground. With a hobby robot it may be the treads coming off, or if the robot is light enough and the ground hard enough the treads will slip.

One way to fix this is to slightly lower the center of each tread creating a pressure point in the tread and making it so that that center of rotation will more likely "win" the fight with all the others. (you are actually shortening the length of the tread that is on the ground at any one point in time so the center of rotation will be infront of or behind your lowered section depending on the way in which the robot is tilted at any given point in time.) You still have the benefits of having lots of grippy places that can help you move around, but you also help to alleviate the problems of having long tracks on either side of your robot.

If I have time I'll make up a diagram to help illustrate what I am talking about.

Hope this helps.


08-19-2013, 01:28 PM
each pair of wheels wants to rotate around its own center

That is not how it works. The wheels don't have some magic knowledge about which other wheel they're "paired" with. (Unless they are rigidly connected with an axle, but then you'd have no differential steering at all.)
It may seem like a simplifying assumption to explain it like this, but I feel that doing so is just a dis-service. A more correct explanation is not any harder IMO.

The easiest model of the object is that each wheel patch is a point contact with the ground, and each point contact exerts force around the center-of-gravity of the robot, which turns into torque.
The wheels want to spin forward; yet the sideways rotation around the center-of-gravity is not aligned with the wheel movement direction for the case of 4 wheels. See below.


The wheel contact patches are not actually point-shaped, which will matter once you want to do real-world simulation of various drive conditions (or write an accurate car racing game ;-) But the math of the point-contact model is surprisingly simple, and gives surprisingly good results for well behaved inputs (flat ground, well inflated tires, etc.)

And I realize now that the blue arrows on the left side should point down, not up, for maximum clarity...

08-19-2013, 01:52 PM
Jwatte...Thanks for clarifying and including a diagram.