Cartesian to Diamond coordinate translation for joystick to differential drives

Question

To anyone interested,

I have a simple telepresent robot that I built.  The user drives the differential drive robot using a joystick, specifically an XBox controller.  In code I had to deal with translating the X and Y Cartesian coordinates from the joystick to values that can be applied to the left and right drive motors.  This isn't a trivial task if you want a smooth translation.

After coming up with a unique approach and making it work in my robot, I thought others might be interested in my solution.  I created a Dot Net DLL containing the basic functionality so that it could be implemented in any Dot Net application.  I also created a document describing the algorithm for anyone who wished to convert it to a different environment as well as a description of the DLL and how to use it.  I provide both of these for free with the usual "use at your own risk" provision.

I invite any interested people to look at my document and/or the kit and give me feedback.

http://www.rchordata.com/files/Using%20Diamond%20Coordinates%20to%20Power%20a%20Differential%20Drive.pdf

Thanks,
Dogulas

Answer 1

file not found. BTW I am kind of interested in this since I'm doing it myslef in Java at the moment.

Answer 2

Spiked3,

Sorry for the bad link.  I changed around my website and forgot about this post.  Try this.

http://www.rchordata.com/DownloadDiamondCoordinatePg.aspx

If you still have problems, let me know.  You can also contact me through the website.

Thanks,

Dogulas

Answer 3

yeah, actually I had run across that in my searches - but it gets a 404 there as well when you go to get the files.

for the time being I snagged an algorithm out of MRDS. it looks like this;

static void tankDrive(float x, float y) {
double left = 0, right = 0;
// lifted from mrds
left = y + x / 4;
right = y - x / 4;
System.out.printf("tankDrive %f %f / %f %f\n", x, y, left, right);
}

I know its not perfect, but dang its close and simple and does the job.

Answer 4

Spiked3,

Ok.  I think I have it fixed now.  I had someone outside of my network to try it and they had no problem.  If you're still interested, give it another shot.

Thanks,

Dogulas

Answer 5

By the way, an alternative location is here:

http://www.accipitersoftware.com/#/Downloads

Thanks,

Dogulas

Answer 6

By the way, an alternative location is here:

http://www.accipitersoftware.com/#/Downloads

Thanks,

Dogulas

I know this is an older post, but I wanted to comment and say that I recently used this algorithm in a project and it was way better than any other that I found. It handles full forward, reverse, and full left and right (turn on a dime) much better than others I tried. If anyone is interested let me know. Douglas has given me permission to re-distribute his algorithm. He provided the c# version. I have a Java converted version I used on an Android.

Answer 7

I know this thread is a bit old (~5 years), however I am looking to get a copy of the PDF that Douglas rchordata.com mentioned. None of the links provided have worked.

Does anyone have a copy to share or can you point me in the right direction? In the interim, I will take a look at the MRDS that Spike3 mentioned.

Thank you all.

Regards,

BDSI