tank drive in this years game
 

how many teams will build 4wd tank drive anyway? even though it is probably not the best drivetrain for this years game? I am betting most rookies will anyways, as will quite a few veteran teams.

Re: tank drive in this years game
 

Our team, 744 Shark Attack, put the wheels on an old tank drive robot and found that our head driver was able to make turns, spin, and decelerate with out much struggle or delay. Although, this may not be the best, most efficient system, it is simple and is being considered by our team.

Re: tank drive in this years game
 

Tank is probably the easiest, fastest to build, and (if you use the kit parts) cheapest drive train available, so naturally it's going to see a lot of use by new or poor teams. Other drives that work better just take too long, are too complex, or are too expensive for many teams to build. My team is currently going through a money/mentor crunch, so we're going with tank drive (even though we really wish we could build something else).

Re: tank drive in this years game
 

So what kinds of drive are teams considering? Not looking to get anyone's secrets here, jst a discussion of what other alternatives there may be.

Several teams got some good experience with active steering in last year's game. Sadly we are not one of them. :(

Re: tank drive in this years game
 

if only we cold some how independently drive and steer each wheel......oh wait:ahh:

Re: tank drive in this years game
 

If you're going with tank drive, I'd recommend 6 wheels (all driven) with a wider base than its length to maximize turning and minimize hysteresis. This is probably what we'll be doing.

Re: tank drive in this years game
 

We're most likely looking at 4WD tank steering, also. It's the most inexpensive, and (as far as we currently know) will be the best option for us.

Siri, interesting suggestion, I'll bring it up at our meeting tomorrow.

Re: tank drive in this years game
 

I don't understand what you mean by this. But I find the idea interesting. Can you explain more?

Re: tank drive in this years game
 

My team will more than likely be using a wide chassis 4wd or 6wd, depending on how tests go, as they are cheep, and we're likewise low on money.

Re: tank drive in this years game
 

i think that the most common design will be 4wd tank which is the intent of the gdc. however i do think that the most effective design for a drivetrain will be vector drive because of the fact that you will not have a very hard time turning your wheals this year because of the lack of friction and being able to move in all directions will help keep your robot under control however vector drive is still very hard to make.

Re: tank drive in this years game
 

correct me if im wrong but my understanding is this...

surface area has NOTHING to do with friction. the coefficient of friction is the interaction between two surfaces on a molecular level. I think increasing the normal force over each wheel is the way to go....I also think i have said too much :p

Re: tank drive in this years game
 

I don't think vector drive will be as effective as it seems. According to the white paper on robot traction/turning, vector type drives may have problems even moving(IIRC).

Re: tank drive in this years game
 

What is vector drive? I feel very naive asking this, but up till now I felt pretty well versed in drive train types/names.

Re: tank drive in this years game
 

As a guess I believe jblay is referring to what is normally known as crab or swerve.

There is no reason this shouldn't be able to drive so logicalyrandom is likely referring to a different holonomic drive using perpendicular wheels (omnis in previous years)

Re: tank drive in this years game
 

2046 considered building a swerve drive, but the last two we've built used 4 and 6 wheel tank drives, respectively. We'll likely go with 6 since it will be easiest.

Besides, the rest of the team wants a manipulator that's...sketchy at best. No way we'll build our first swerve drive and the manipulator this year.

Re: tank drive in this years game
 

I believe we will be doing a 6wd tank drive with a slightly lowered center wheel. we will also try to encourage it to ride on the back 4 most of the time.

Re: tank drive in this years game
 

We have decided to do a 4WD -wide with the wheels shifted back. This will increase maneuverability and because of the trailer, we're not as concerned about tipping forwards.

Re: tank drive in this years game
 

Could you please provide a link to this paper?

Re: tank drive in this years game
 

Here you go:

http://www.chiefdelphi.com/media/papers/1917

Re: tank drive in this years game
 

Personally, i like 8 wheel 1 motor = 2 wheels l:cool:
tones of traction but a bit on the heavy side:eek:

Re: tank drive in this years game
 

Our mechanical team's basic idea was to use a 4WD system and then make getting it to actually work a programming issue. Although we've also got a "steering booster" system using a gyroscope to help us make sharp turns.

Re: tank drive in this years game
 

We are also thinking of 4 wheel tank drive on a wide chassis? We have also considered a long wheel base with the wheels moved toward the rear of the bot. Just a few questions, has anyone tried running the wheels directly from the toughbox gearbox? Does it work? Are their any modifications needed to make the wheels run directly from the toughbox? We are thinking that the weight from four toughboxes near the wheels will increase friction by increasing normal force near each wheel.

Re: tank drive in this years game
 

what makes a wide base better than a long one?

Re: tank drive in this years game
 

Generally having a shorter wheelbase makes it easier to turn. It's the same concept as having a 6wd with the center wheel dropped, you are basically creating to 4wd systems with shorter wheelbases.

Re: tank drive in this years game
 

4 wheeled tank drive is what i think my team will do *696* MAYBE because we finish designing our bot tomorrow or the next day *this is with complete design with sensors and exact dimensions etc.

Re: tank drive in this years game
 

"Better" is a relative one. The ease with which a base with tank steering and 4 wheels touching the ground turns is related to the ratio of wheel track (width between wheels) and wheel base (front-to-back distance between wheels).

A robot with a long wheel base and narrow wheel track (in FIRST, the 28" wide 38" long configuration) prefers to drive straight, whereas as robot with a short wheel base and long wheel track will turn easier, but have a harder time driving completely straight.

The reason for this is because a 4wd tank steering system requires all of the wheels to skid sideways slightly when turning. If you imagine the robot trying to turn around its centroid, the wheels ideally want to be rolling along the circle circumscribed by their radius to the center of rotation (aka the wheels would prefer to be rolling tangent to the circle described by that radius) But because the wheels are all oriented in the same direction (which is not tangent to the circle), it's easy to see that they will all be sliding sideways slightly in addition to rolling forwards or backwards.

Now go back to the long and narrow versus short and wide base. The wheels on the long and narrow base are pointed at a steeper angle away from the ideal tangency than those on the short and wide. The extreme case for the long and narrow is when your frame gets so long and narrow that it becomes a line, with one wheel at each end oriented to roll along the line. Trying to turn around the midpoint of this line will cause the wheels to only skid sideways and not to roll forwards or backwards. Thus we see that the higher the ratio of wheel base to wheel track, the more skidding is necessary and the harder it is to turn. In addition, because the wheels are laterally closer to the centroid, the torque they can apply to the center is less. In the extreme case, the wheels cannot apply any turning moment to the midpoint of the line, because the applied force is going straight through the point of rotation and thus the lever arm is zero. Conversely, if powered, these wheels will drive ONLY in a straight line, because they cannot turn.

In the extreme case of the short and wide base, it again becomes two wheels on the endpoints of a line, but this time the wheels are oriented to roll perpendicular to the line. In this case, the wheels ARE rolling along the line of tangency, and therefore are not slipping sideways at all. This is the theoretic best case for turning, but the wheels are much more susceptible to being rotated, since they do not resist the rotation at all.

In the context of Lunacy, a while wheel base provides a better turning ability. When you have a trailer attached, it becomes easier to drive a wide base in a straight line because the trailer is trying to straighten out the robot, and it also becomes easier to turn because your wheels can provide the maximum torque about the point of rotation.

Re: tank drive in this years game
 

i dont think any kind of special drive trane is going to be that much of a help because the wheels slip around no matter which way they are oriented. persoanally i love tank drive but i am thinking about arcade (which i hate). the reason being is that both sides of the robot will turn exactly the same rate. on a normal playing surface the human error between the two sides was negligable but on this surface any amount of wasted traction is a lot. but i dont know yet, we need to test out on the regolith on wide set base. its been agreed that a wide set base is definitely the way to go. this will allow more space in out colletion system, and the large space between the wheels will lessen the differential between the wheels. but of course, we have to try it first.

Re: tank drive in this years game
 

Theoretically, crab drive will give you an advantage.

Skid Steer is just that - skidding your wheels. Which, as we all know, takes a substantial hit in the friction depart. Theoretically.

The crab drive's advantage is threefold:

One, you do not skid the wheels when turning.
Two, you do not need to impart angular momentum to your robot to turn. In fact, if you drive in the direction you want to go, the trailer will help turn your bot for you.
Three, you can turn your wheels sideways. According to the theoreticals, the sideways friction is much higher than the linear. So, again theoretically, a tank drive robot trying to push a crab robot will not be able to if the crab turns his wheels sideways.

In the past, we were always limited by torque. Now we're limited by friction. However, the weight of crab may be an issue for many teams less familiar with it (like ours). You're going to need a good deal of manipulation on your robot - pickup, accumulation, release... those probably won't be lightweight.

Re: tank drive in this years game
 

I am becoming increasingly convinced that people are underestimating the tried 'n' true skid steer bot this year. All indications I've seen from those who have tried it with the rover wheels are that turning and moving aren't really problems.

Will a crab drive give better performance? Sure - it always has and always will. But the drive system style is dwarfed by strategy, driving ability, and scoring function when it comes to who gets to stand in the winner's circle. Some teams have the resources to make a crab drive to go along with an excellent robot in other areas. Others don't. I am not convinced that you will see any more crab drives hoisting blue banners than in other years.

Re: tank drive in this years game
 

I think that 4W drive may be the best solution. We looked at crab drive systems. However it seems that once you take into account the trailer, you will have less control given that a sudden change in direction would cause the trailer to jackknife and spin the robot around.

Have any teams done testing with a crab drive and a trailer?

Re: tank drive in this years game
 

Would anyone be able to show me a link to a CRAB drive? I have an idea what it is but would like to see me about it.

Re: tank drive in this years game
 

There are several ways to do a crab drive, but this will give you an idea...

http://www.chiefdelphi.com/media/photos/28618

Re: tank drive in this years game
 

Here's a fantastic white paper describing many types of drives.

Re: tank drive in this years game
 

Here is why I do not like the tank drive. The field still has carpet along the edge. If you get two wheels (based on a 4 wheel system) of one side of the robot on the carpet, the other side of the robot will not have enough traction to turn. So your robot will be stuck on that edge of the field for the entire map.

This is why I am pushing for a hybrid between a swerve and "monster truck" style steering. Wheels are powered individually and move to steer, so if you get caught on that carpet, you can just turn off it.

Re: tank drive in this years game
 

Thanks Taylor and Taz!!

Re: tank drive in this years game
 

Is their a limit to the number of wheels you can use?

(I am thinking inline skates) Not sure on the physics since friction is a product of weight and area, putting more wheels on increases area but lowers the force on each wheel, so I am not sure if overall traction, ability to accelerate improves. Tank style steering is also called 'skid' steering for a reason. Putting more wheels on will make turning harder, so aiming the robot to shoot is harder (Unless your shooter is independent of robot position). Most of the successful 6 wheel drive systems really cheat a bit to let them turn, typically the center wheel is a bit lower then front and back, so the robot rocks a bit when turning - and typically only 4 wheels are providing traction at any one time.

This year's game is going to be a gas - I think rookies are going to have a good chance to do well. I don't think we'll see a lot of use of G14. Just gonna be to hard to score. I hope you teams are gonna recruit good human players. I'd start bringing donuts to the basketball team now. A good three point shooter could spell the difference. If your school does hockey, I'd recruit my driver from there too!

Re: tank drive in this years game
 

I've seen the difference between the coefficient of friction in line with the wheels or sideways to the wheels mentioned in a couple of places. I can't think of a reason for this. Was physics class so long ago that I have forgotten some fundamental concept?

Ivan

Re: tank drive in this years game
 

Experiment trumps theory on this one. We took last year's frame, loaded it up with ~140 lbs, and put it half on FRP and half on carpet... While towing a hand truck (casters on the front, high-mu in-line rubber wheels on the back), and turned with no problem. The wheels on the carpet spun more than the wheels on the FRP, and the bot just pivoted right off, easy as pie. Pie on ice. Or the moon. But pie.

...and Ivan, the tests we've done show no difference in inline vs. lateral coefficient of friction. I don't know where the official numbers came from, but they don't jive with the tests done by several other teams, either.

Patrick

Re: tank drive in this years game
 

Thanks Patrick, that makes more sense. Thanks also for saving me the time of doing the test myself.

Ivan

Re: tank drive in this years game
 

Hysteresis is usually a ferromagnetic term, but in this case it's basically the tendency to over-steer. (official definition: the lag in response exhibited by a body in reacting to changes in the forces...affecting it.)
So, in it's simplest terms--which is about all I can actually explain, though I'll try more if you'd like--the "easier" it is to steer, particularly the less you have to skid, the less hysteresis you'll have. I feel like I should explain its relation to static and kinetic frictional coefficients, but I'm not sure that I can do it understandably.

There is no limit to the number of wheels (except for your maximum expenditure and the number that can fit in the 28"x38"60" box).
In a perfect world (i.e. Coulomb's), surface area has zero effect on friction. Ff=mu*Fn, and we can all go home now. In reality, surfaces aren't perfect, and their mu's can vary across the surface. At first I didn't think this was a big deal in this game, but from what people say about floor wear-and-tear, it might be a minor advantage. I still seriously doubt it could play a major part.
Can you explain your more wheels=harder steering conclusion? My train of thought (and experimentation) was to consider the difference angular distance for wheels closer and farther from the center of rotation. Without sacrificing wheel base (stability), more wheels means wheels closer to the com, which means less skidding. Our prototypes have supported this; 6wd (even without dropped center) does turn better, chiefly because the center wheels do not have to skid as much. Someone on our team put together a beautiful analysis of this, I'll ask him to post it.

Are your wheels worn in? We did a loaded incline test with the fresh wheels and found inline as .09 and transverse as .13 (static). Not quite FIRST's numbers, but I can't blame them for giving us the bottom of the range. Does this change as the wheels wear?
To answer Ivan's question, in theory the difference between transverse and inline coefficients is based on the structure/pattern of the material. Almost all materials have these differences. However, if the difference disappears with wear, I might be inclined to blame it on that nice seam running down the wheel. If not, it's surprising given that there are no observable significant grains on the wheels, but certainly not physics-defying.

Re: tank drive in this years game
 

I was thinking along the lines of how Ice skates work. Also how the keel in a sailboat tends to keep the boat going straight. In skating you turn to the side in order to slow more quickly, but I am not sure this would be a good analogy. I think it works in skating because the skate actually starts breaking the ice and making snow cones, as opposed to straight skating where the pressure of the skate melts the ice and forms a frictionless layer of water.

This is all academic - teams will experiment and find what works best for them. From what I see 4wd can turn and accelerate and clever speed control programming along with driver practice will pay off. Be aware the trailer is really part of the robot this year - it will sort of act like a big sea anchor. It probably will make it easier to go straight but harder to turn sharply.

Has anyone found a cheap substitute for regolith? Would something like freshly acrylic waxed linoleum be close? Looks like I may be mentoring some rookies this year so any suggestions to cut costs would be appreciated.

Re: tank drive in this years game
 

Not tremendously, no -- but they're not perfectly fresh, either.

Patrick

Re: tank drive in this years game
 

Ours have become a bit dirty, but not worn either.

Re: tank drive in this years game
 

Not sure what makes you think more wheels means more traction. In general it doesn't.

Siri: Thanks for the explanation, to mirror it back and ensure my understanding, the lag between steering input and robot reaction could be called Hysteresis.

Don

Re: tank drive in this years game
 

We built a simple "wide" test bot today with two caster wheels made from 2009 "slick wheels" on the front and 2 powered "slick wheels" on the back. We weighed the test bot to 140 lbs and attached a 35 lb trailer on the back. I was shocked at how well the thing actually did. I was expecting poor performance but it wasn't too bad. We have something more complicated in our build plan. If that doesn't pan out, We may go back to the caster bot.

Re: tank drive in this years game
 

I found a similar but different hysteresis definition here that I like a little better.Blake

Re: tank drive in this years game
 

Ed,

First, I'd like to say that 'Sparks' is a great name for a FIRST mentor. If only you mentored SparX (1126), nothing could stop you!

Second, I'd like to suggest that you not consider how the robot performed in isolation... It may turn well, and even accelerate/brake acceptably (although I would be concerned about losing 50% of my traction by only having half of my wheels driven)... But when it comes to pushing or pinning, you're *doomed* against a robot with four driven wheels. ...not to mention if you get pinned with the casters on the carpet and your driven wheels only on the regolith. This could be very bad for you.

Then again, much of the doom-and-gloom bounced around on this forum turns out to be no big deal, so...

Patrick

Re: tank drive in this years game
 

Yep, Good point ........