4 wheel drive vs. 2 wheel drive

we have been debating whether all wheel drive or front wheel drive or rear wheel drive was better.

General rule of thumb is that you want any wheel touching the ground to be powered. If not, you’re cutting into your team’s available pushing force.

Indeed. If you are imparting robot weight on a wheel, make sure it is powered.

all wheel drive will be the best

That being said, you really want your weight as far back as possible so that your pivot point is as close to the hitch pivot. That reduces the moment that the trailer imparts on you robot, which it needs to counteract to turn, and will allow you to turn better.

Between rear and front wheel drive, it’s really not viable to use a rear wheel drive unless you don’t plan on ever exchanging an empty cell… or being pushed into a corner. That portion is carpeted and if you use rear wheel drive to get there, you’re going to end up with the front wheels on the carpet with high traction and the powered wheels on the regalith with low traction, making it incredibly difficult to move. And backing into the corner wouldn’t work either, there’s a trailer in the way.

Where your weight goes is up to you but please know that any and every wheel touching the ground needs to be powered! Also know that your bot should weigh 119.99999 lbs (without battery and bumpers).

My recommendation would be to run some tests with the kit chassis. We subbed some wheels out on a 4x4 chassis last night and ran it on the FRP and carpet. We did a long wheel-base and a short wheel base. We also shifted the CG from 50/50 to almost all on the rear or the nose. All of these are without the trailer so far, so we will try those tonight.

All I can say is I would recommend doing a few tests for yourselves. The dynamics are very different on the carpet vs. on the FRP. Set-ups that were nice and controllable on the FRP wouldn’t turn if two side wheels got on the carpet, and set-ups that were easy to turn off the carpet were very squirelly on the FRP. Also, the bot lateral grip seemed a lot lower than the inline grip.

That brings up a very important point, there is a hierarchy when you design:

Ideas are great, but can often be flawed.
Fully thought out theories are better, but often involve a lot of assumptions.
Models (calculations and 3D models) are often made of some tests and measurements and some assuptions.
Test based data is even better yet especially if it validates your model assumptions.
Prototypes can be even more accurate.

My point on this is I have read through a lot of threads on how these bots should be built and how they will behave. I have a lot of vehicle dynamics experience, and I was still surprised by a couple of our tests last night. :yikes:

I am going to have to respectfully disagree. Having lived in snow/ice conditions for most my life, the best option is FWD with Rear wheel steering. Weight must be over the front wheels and can be done due to the trailers weight so you font tip over. It is much better to pull weight than push when it comes to this subject. Transverse friction is used by the rear wheels to guide the bot to turn. Also as for the trailer, since sway bars are not used (i tow trailers all the time) all of the weight will be centered and pulling at the tongue of the trailer. So putting the weight at the back would epically screw you over since it would want to lift the front wheels off the ground thus taking powered wheels off the ground (in any set up)

Edit: All wheel drive might help but might be wasting power, motors, and battery life…
(i have done some tests and this is what i get, but its up to you)

All wheel drive absolutely helps.

The physics reason for all wheel drive is this:

  1. Let’s say your robot has 4 wheels
  2. Each of these wheels supports some fraction f[sub]i[/sub] of your robot’s weight. The sum of all F[sub]i[/sub] = 1.
  3. Each wheel will thus have a normal force of m[sub]Robot[/sub]*f[sub]i[/sub]g = 9.81m[sub]Robot[/sub]*f[sub]i[/sub]
  4. Each wheel will thus be able to generate (if powered) a frictional force of
    F[sub]f[/sub] = uF[sub]n[/sub]
    F[sub]f[/sub] = 0.06 * (9.81
    F[sub]f[/sub] = 0.5886
  5. Your robot’s total potential frictional force is the sum of the frictional forces of all of its driven wheels.
  6. For an all wheel drive robot, F[sub]fRobot[/sub] = 0.5886*m[sub]Robot[/sub] (because F[sub]i[/sub] sums to 1)
  7. For a 2WD robot with its weight evenly distributed, we know all the F[sub]i[/sub]s are 0.25 (due to equal weight), but only two of them are driven. So F[sub]fRobot[/sub] = 0.2943*m[sub]Robot[/sub], or half of the AWD’s robot full force.

So based on 5), 6), and 7), we know that by choosing not to drive some of your wheels, you are throwing away potential friction. Wheels that are touching the floor while the robot is driving and are not powered are completely wasting potential friction. A robot with its weight evenly distributed over 4 wheels that only drives two wheels will be two times slower than an identical all wheel drive robot. The two wheel robot could potentially employ a very complicated suspension so that its weight shifts onto its driven wheels when acceleration or decelerating, but it would be a waste of time when you could just drive all four wheels.

if you are going to attempt to prove me wrong, at least understand my argument. I never said an even weight distribution, in fact i did say WEIGHT OVER THE FRONT DRIVE WHEELS, plus my system cuts down on weight thus allowing you to allocate that precious weight to other areas of the robot like the heavy control system, or ballast weight that you can add where needed…

Yes, but even with an uneven weight distribution you’re losing pushing power. Even if you stack absolutely everything over your driven wheels, you’re probably still losing 10-20% of your pushing power from the weight that will inevitably be over your undriven wheels. Plus you’re adding a large design constraint in that all your heavy mechanisms must be in a column over your driven wheels to maximize force.

I understand your idea (though I did gloss that over in my original post, sorry) and it is definitely the most optimal way to implement a 2WD system, I just think that any 2WD system is far from optimal this year.

well from concept everything is in the front anyway, but in this situation pushing power doesn’t matter. I you want control you want pulling power. The same reason drift cars are rear wheel drive is the same principal rear wheel drive results in loss of control, and 4wd is going to work too hard to steer sicne the wheels would rather slip inline than transverse (which makes skid steer out of the question) i will admit that 4wd is a valid point, but all drive systems of this year have MAJOR tradeoffs, thus making not one type (except for rear wheel drive) the best, it will be interesting to see at competition…

Agreed. My comments were was based upon the assumption of a standard kit style frame with no steering and the choice being between 2 or 4 wheel drive. Most steered mechanisms will be better than that, but if you are limited to a standard kit drive system, it is my experience with the kit base and attached trailer that the more rearward your pivot is, the better able you are to turn.

i see, the introduction of a powered steering system changes things quite a bit…

right but since there are no sway bars present, the tongue weight and force it exerts will be concentrated at the hitch, if the hitch is in the center of the two steering wheels, the force will be distributed evenly between them giving the needed down force to turn the bot…

To everyone considering not using all wheel drive: DON’T DO IT!!!

I don’t care if you are steering, not steering, using 2 wheels or 18. With the limited amount of usable force available you want every wheel touching the ground to contribute to converting normal force into tractive force.

Every non powered wheel you have touching the ground will reduce your ability to accelerate and push.

Pushing other robots around isn’t going to be viable this year, but you still need to be able to use all your traction effectively to navigate the field quickly. The problem with FWD and front wheel steering is that you’re using up all the traction in the front and wasting the traction in the rear, and you’re just going to understeer all over the place and not be able to do anything about it. Rear wheel steering wouldn’t be the way to go this year because you’d pivot around the front wheels, which would make steering with the trailer much harder. Since you want to have your COG and pivot point towards the rear to be able to turn with the trailer easier, the optimal 2WD setup would be RWD with front wheel steering.

The same reason drift cars are rear wheel drive is the same principal rear wheel drive results in loss of control,
Drifting isn’t losing control. In fact, it’s being able to control the car with the throttle even when you’re breaking traction. Watch the sport at at competitive level, and tell me those drivers don’t have control of their vehicles.

and 4wd is going to work too hard to steer sicne the wheels would rather slip inline than transverse (which makes skid steer out of the question)
Actually, testing has shown that the wheels are so rigid that they don’t much care if they’re sliding inline or transverse. Skid steer can work perfectly fine this year. And if you’re comparing AWD with car-style steering to 2WD with car-style steering, the AWD system is going to destroy the 2WD one. Watch some World Rally Championship, and you’ll see the advantages in car control that AWD gives you on low-traction surfaces. Being able to four-wheel drift through corners would be a huge advantage this year, as it would let teams maneuver while maintaining speed to avoid being scored on.

i will admit that 4wd is a valid point, but all drive systems of this year have MAJOR tradeoffs, thus making not one type (except for rear wheel drive) the best, it will be interesting to see at competition…
Like I said, as far as 2WD car-style drivelines go, RWD with front wheel steering is going to be optimal. The ability to control the rear of the robot with the throttle in a RWD setup would give you better control than in a FWD setup which can understeer much more easily than it can oversteer.

And by the way, the whole “pushing is better than pulling” thing is more or less a myth. RWD makes more efficient use of it’s traction than FWD, no matter what surface they’re on. The reason why FWD is better in the winter is that when they’re varying levels of traction on different parts of it road, it’s easier to control, and understeer is safer than oversteer. Manufacturers don’t like selling cars that oversteer, is because the majority of people don’t understand vehicle dynamics, and don’t have the experience to correct oversteer instinctively. If a RWD car starts to slide out, most people are going to slam on the brakes. This transfers weight to the front of the car, increasing the grip on the front wheels, and decreasing the grip in the rear, causing the car to pivot further. To correct a slide in a RWD car, you need to use steering input and throttle modulation, not panic braking. In the context of Lunacy, where the traction on the regolith is entirely uniform, and it doesn’t matter if you slide into other robots, RWD would be far superior.

Just because traction is reduced this year doesn’t mean that the laws of physics went out the window. FWD still sucks. In fact, it sucks even more this year, because the robots are going to be driving at the limit of their grip all the time.

8-wheel drive
Use pairs of wheels

Kind of late, but thought it was worth posting.

Our final configuration uses 48 wheels.