Re: Tank Drive Car
 

To your knowledge, does this interface already have a unique name, and if not have you given it one that we could use here on the forums?

Re: Tank Drive Car
 

No, we haven't given it a name. We've been talking about using this interface since the 2008 game, but this is the first year we've actually done it.

Feel free to think of a good name. I'm not very good at coming up with names.

Re: Tank Drive Car
 

Chris,

Do you have any idea why using the knob works better than a joystick when the robot it facing the driver?

I have suggested a rotating control to our team several times, but none of the students seem interested. I just figured it was the gaming mentality coming out.

-Hugh

Re: Tank Drive Car
 

I know you addressed this question to Chris, but while you are waiting for him to answer...

It is because with the twisting action of the knob, the driver intuitively thinks of rotation as a clockwise/counterclockwise thing (which is independent of the orientation of the robot) rather than as a go_left/go_right thing.

Re: Tank Drive Car
 

Excellent point! I didn't think of that... Thanks Ether.

I will try that point with the students...

Any ideas for a good supplier of a centered spring return pot?

-Hugh

Re: Tank Drive Car
 

Thank you Chris for confirming my quick study!!

We call that type of interface, "RC Car steer", because they are independent sticks (used with your thumbs).

Chris's design is more like a modern RC car where you have a throttle and a wheel on the side of a gun like stick... sounds like he has it laid out flat though? mini steering wheel then?

Either way, they are independent up-down and left-right axes where you cannot inadvertently bump one when doing the other. I'd call all of them "RC car" steer.

We also used speed encoders and stiff proportional gain to enforce the selected turning angle... At that point it felt like a highend RC car (linear turn angle response), except that it really slowed down in the turns. We put neutral axis steer within a deadband of zero speed. The problem was that the stiff gain was a serrious battery killer and we ended up going open loop during tele-op.

One of my objectives in software for team 302, is to push the actual driver interface to the drive team to program, or at least become an itegral part of its development. If we have multiple drivers, they can have different "crew stations." One driver can have tank steer and another RC steer, and different buttons for "turbo" and "crawl", or this year's "spin throw". Different joysticks with different dead zones... the trick is to create a module which generates the same high level commands.

This architecture is shown a little bit in our VIRSYS example code... where among other things, we will be driving a virtual robot before it's built and able to practice anytime/anywhere. http://www.chiefdelphi.com/forums/sh...ghlight=VIRSYS

Re: Tank Drive Car
 

Ether said it exactly.

If a student has a hard time thinking of clockwise/counterclockwise, I say the following to them: "place your index finger on the front of the knob (part of the knob farthest from you), and your thumb on the back of the knob (part closest to you). Now, the front of the knob (your index finger) represents the part of the robot farthest from you - regardless of which direction the robot is facing. The back of the knob (your thumb) represents the part of the robot closest to you - regardless of robot orientation." That works better for some students. It's really depends on the person as to how they sort things out in their mind, but the knob keeps things much more consitent regardless of robot orientation.

We looked like crazy for an off-the-shelf spring-loaded knob. We couldn't find one. If anyone knows of one, please let me know.

We had to make our own spring-loaded knob out of a potentiometer. We did the spring loading by drilling a hole into the knob then running some surgical tubing from the hole to a screw in our control board about 10 inches away. It was a bit crude, but it worked. We wanted to create a better solution eventually, but you know how it goes when you have a million other things to do to get the robot shipped and the current solution works.

Re: Tank Drive Car
 

these ones

http://www3.towerhobbies.com/cgi-bin/wti0093p?&C=JAC


each have two spring center off loaded pots. one for steering, one for throttle.

Re: Tank Drive Car
 

Surprised you didn't just use the twist axis of a "side winder" (take the stick off and put a knob on there), that has a nice return. Or a "3D mouse" or "space ball" has a nice twist axis and would work too. It's easy if they are recognized as "gamepads" and not mice... not really sure about that though.

Re: Tank Drive Car
 

Here is one:

http://www.precisionsales.com/potent...ringreturn.htm

-Hugh

Re: Tank Drive Car
 

That won't work because it is is spring return to one end, not to center.

Re: Tank Drive Car
 

Anybody had a look at the Playstation Ultra Racer?

Re: Tank Drive Car
 

What is the meaning of this, then ?

Re: Tank Drive Car
 

Oops, i guess it will work.