New wheels / floor and encoders for position.
 

It is great that First included encoders and the ability to mount them to the gearbox. It's too bad that combination will be effectively useless this year. With the limited traction, there will be so much wheel spinning, as well as lateral sliding, that gearbox encoders will not give any useful information about where the bot is, and what it's doing.

<R06> talks a great deal about what wheels can be used for traction, and that no devices may be used to increase traction. What about a high traction wheel that touches the ground, but does not affect traction? Say, a wheel with an encoder attached? The way I interpret the rule, this would be acceptable.


Any disagreement?

Re: New wheels / floor and encoders for position.
 

The thing you would have to ensure is that the wheel can offer no resistance when pushed sideways (which is pretty hard). Even a caster design would offer some form of resistance until its direction was changed.

Re: New wheels / floor and encoders for position.
 

Actually, a caster was what I had in mind. I believe that any resistance from the caster will be at least an order of magnitude less than wheel friction, so in engineering terms is "negligible", therefore can be ignored.

Re: New wheels / floor and encoders for position.
 

If the wheel was mounted so the axis of rotation was fixed, it would provide lateral traction. Even if the wheel was mounted so it could swivel, it would provide an instantaneous amount of traction while the wheel swiveled to its new position. The only way I can see this feasible is if you mounted a sphere in the middle, like the ball for an old mouse. Even then you would have to prove to FIRST that it provided no additional traction. Unfortunately, I don't think they would accept this.

You could mount a wheel that was not powered. As long as you have good bearings that provide negligible friction in the axle, it won't slip. An encoder could be attached to the axle. The issue with this is that you loose precious traction with each wheel that touches the ground that is not powered. If your willing to go this route however, you could mount a wheel in a perpendicular direction to receive encoder values for both the x and y axis. The wheels have such little traction, a wheel in the perpendicular direction won't resist motion provided by the other wheels.

Re: New wheels / floor and encoders for position.
 

Why couldn't you use encoders with slipping wheels? The sensors will certainly tell you actually how fast the wheel is spinning. Maybe you can use that information with an expected PWM signal vs. speed curve for traction control?

For example, if you observe that the wheels suddenly start spinning from a low speed to a higher speed, assume that slippage is occurring and reduce the drive signal.

Cars are really starting to take advantage of traction control. Maybe FIRST robots can learn from them?

Russ

Re: New wheels / floor and encoders for position.
 

Not quite correct
<R06>
ROBOTs must use ROVER WHEELS (as supplied in the 2009 Kit Of Parts and/or their equivalent as provided by the supplying vendor) to provide traction between the ROBOT and the ARENA. Any number of ROVER WHEELS may be used. The ROVER WHEELS must be used in a “normal” orientation (i.e. with the tread of the wheel in contact with the ground, with the axis of rotation parallel to the ground and penetrating the wheel hub). No other forms of traction devices (wheels, tracks, legs, or other devices intended to provide traction) are permitted. The surface tread of the ROVER WHEELS may not be modified except through normal wear-and-tear. Specifically, the addition of cleats, studs, carved treads, alterations to the wheel profile, high-traction surface treatments, adhesive coatings, abrasive materials, and/or other attachments are prohibited. The intent of this rule is that the ROVER WHEELS be used in as close to their “out of the box” condition as possible, to provide the intended low-friction dynamic performance during the game.

R06 does not stop you from increasing traction, it just prevents you from using anything else to provide traction

Negligible traction is still traction.

Traction control and accelleration control will be key this year to winning alliances. IMHO

Re: New wheels / floor and encoders for position.
 

It looks like now is a great time to start programming! Traction control and robot position/velocity/acceleration information could really help a team control their robot.

Those of us who don't have to drive on icy roads are probably at a disadvantage....oh well....

Re: New wheels / floor and encoders for position.
 

I think this is the 1st time I've heard that driving in icy conditions is an advantage :P

Re: New wheels / floor and encoders for position.
 

We've been exploring intelligent slip-prevention systems, and that would be quite the challege even without trying to use the camera to track and shoot balls at a target... we're probably going to try but it will be tough.
As to the idea of providing another wheel which would be measured with the encoders, perhaps a trans-wheel? it would have extremely low friction when moving sideways, and trans-wheels are not "wheels, tracks, legs, or other devices intended to provide traction"

Re: New wheels / floor and encoders for position.
 

How about using and old ball type computer mouse? it's a sensor...so it's not intended to provide traction. I wonder if they're still available COTS?

Re: New wheels / floor and encoders for position.
 

My guess is (for official rules wait for the FIRST Q&A forum) that if it contacts the floor it will be illegal.

How about a IR mouse? It doesn't need to contact the surface.

Re: New wheels / floor and encoders for position.
 

I don't know how optically "interesting" this floor will be. namely, is there going to be a visible difference between one spot on the floor and another? it might need something as precise as a laser mouse. it's an interesting idea.

Re: New wheels / floor and encoders for position.
 

I was part of a group that used optical mice as encoders. It was a bit of a pain in the neck to start, but worked well once it worked at all.

One idea I'm fiddling with is measuring the current to the motors to approximate torque. A 20% difference between slip and grip should be measurable. I wish we had CAN this year, since the Jag's have that feature built in.

Re: New wheels / floor and encoders for position.
 

Your encoder wheel doesn't have to rob (much) traction from the rest of your robot. If you mount a fifth rover wheel on a swing arm or something similar, it will use only its own weight for traction. Then it's just a question of how good of a bearing do you need. ABEC 7 anyone?

Re: New wheels / floor and encoders for position.
 

Anti-lock brakes in cars use accelerometers plus wheel encoders to detect wheel slip.

Re: New wheels / floor and encoders for position.
 

How about simply using two omni wheels with encoders mounted at a 90 degree angle from each other. These would not generate resistance when changing direction and could easily be the x and y for movement.

Re: New wheels / floor and encoders for position.
 

Omniwheels do, indeed, add traction (albiet minimal) and thus would be illegal

Re: New wheels / floor and encoders for position.
 

you could just use a kit wheel and make it pivot that way your not breaking any rules because its not being powered it would not slip so the encoder would know where it was

Re: New wheels / floor and encoders for position.
 

years ago the railroads used quantity four 3,600 hp locomotives to pull a large unit train.

today they use at most three. the way they did that is by automated traction control.

a doppler radar is under the locomotive measuring the actual speed. the traction computers optimize the power delivered to the wheels so that there is optimal slippage driving wheel, about 3%

they got rid of a whole locomotive, plus they get greatly improved fuel economy and operating performance.

anyone game for trying this ? we certainly have the computer to do the job !!

put a dedicated speed controller, drive motor combo on each wheel, figure out where you are actually going, versus where you want to go and off you go.

Re: New wheels / floor and encoders for position.
 

You're proposing to make a castor out of the rover wheels? that sounds legit but difficult.

Re: New wheels / floor and encoders for position.
 

not really it be really easy with a piece like this



attache the caster plate to the top and throw some encoders on it along with a giro

dimension are not exact i gust threw it together in cad

Re: New wheels / floor and encoders for position.
 

I just tried placing my optical mouse on top of a round marble ball and scooting it around. Seems to work, and provides no traction. COTS - Came Out of The Scrap. I'll bet something like that would be legal.

Re: New wheels / floor and encoders for position.
 

Look at a rear caster and you'll see the axle is located significantly away from the vertical axis of rotation, meaning the wheel 'trails' the vertical axle. Look up the automotive term "Caster" to see where the name came from...
No, only wheel speed is needed for ABS - the derivative of the wheel speed is monitored and any decelerations in excess of a 'known' amount (slope) means the wheel is locking up. Accelerometers are not used for traction control either, where the undriven wheel speeds are compared to the driven wheel speeds to address the effects of excess acceleration. ESP control uses accelerometers.
I think that an optical mouse, 1/16 - 1/32" off the surface, might work well. Doppler radar or some similar method would be pretty cool - can that be done with light or ultrasonics? Or just get a radar gun (Ramsey Electronics sells a kit for $60) and point it at an angle to the floor.

Re: New wheels / floor and encoders for position.
 

Use an inertial mouse. Or refocus an optical mouse to measure movements...

Re: New wheels / floor and encoders for position.
 

The hard part of using the mouse is interfacing to it. The new controller doesn't have any USB inputs on the Robot side, so you would either have to add a microcontroller or learn the PS/2 standard.

Alternatively, you can poke directly into the chip inside the mouse. There is no guarantee that this is a standard protocol - each vendor can use their own. This is the route we chose.

Re: New wheels / floor and encoders for position.
 

The kit also came with a nice accelerometer. If the wheel speed does not match the expected value from the accelerometer, then .....

I think the encoders are going to be very useful this year.

And my gut feeling is that anything touching the floor other than the approved wheels is going to not pass inspection.

Re: New wheels / floor and encoders for position.
 

Back in 2004 when we worked with Technokats to rework an optical mouse to detect robot movement, we interfaced directly into the chip in the optical mouse and bit-banged the protocol. It was pretty easy with the IFI RC. At that time, there was only 1 manufacturer of the optical interface chips (probably due to patents etc). Not sure if that has changed since then.

Re: New wheels / floor and encoders for position.
 

You could use a combination of encoders and an accelerometer to create a traction control system. Traction control works by transferring power to the wheel(s) that is not slipping (ex. if you're driving a car and one of your powered wheels is on ice and the other is on pavement). This wont work when all your drive wheels are on the same surface.

Preventing your driver from spinning out the wheels is the best that you can do (this will provide rolling friction rather than sliding friction). Anti-lock braking wouldn't be a bad idea either.

As far as telling where you are at, a system using an accelerometer and gyro should do the job.

Re: New wheels / floor and encoders for position.
 

Using the accelerometer and gyro for traction control are a bit more difficult that people have been saying without a pretty decent bit of math.

The accelerometer gives linear acceleration.

The gyro can give angular acceleration.

You will need to combine the two mathmatically and integrate to get the expected wheel speed.

Most quadrature encoders that are reasonably expensive are 256 counts a turn (or thereabouts).

When you start to compare the wheel speeds to the expected speeds, the error introduced by the integration of the acceleration and the inaccuracies due to the counts / turn calculation of wheel speed create a very non-trivial problem to keep your wheels from slipping all the time.

Most people I know who have traction control turn it off in the snow. Many times you NEED to floor it, and the traction control simply won't let you. Plus, you can't accelerate at all.

This isn't ice - it's simply low friction. Desensitized joysticks and careful drivers will likely be a better idea than traction control for 'most' teams.

Re: New wheels / floor and encoders for position.
 

Has anyone considered using an omni-wheel to act as an encoder wheel? I agree with the last post that other than an encoder, an accelerometer can be used to correlate the bots acceleration to what the wheel's angualar acceleration is. The wheel's angular acceleration can be figured out using the kits's encoders. If there is a deviation then you would know that the wheel is slipping and will need to slow down the drive wheels to regain traction. It is important to note however...for this method to work it is assumed that the bot's acceleration is due only to the wheels. If another bot is pushing your's or you hit something while trying to maintain traction, it will be harder to sort out (using the accelerometer and wheel encoder method) wether or not your bot has lost traction due to wheel slippage.