pic: What's going on here?

[gblake]

Are you guys who are saying that a crab will stay in, or mostly in, the wheel's static friction domain, trying to tell me that crab drives are going to be able to perfectly match their wheels rotation and orientation with the bot's motion vector (speed and direction) AND then, when they want to turn, will be able to magically have the wheel trace out an arc without slipping sideways in the least????

The last time I looked, wheels travel in straight lines and turning a bot required the wheels to follow an arc. Twisting the wheel (even slightly) to cause it to travel in an arc would seem to involve a bit of slip (did I mention that the wheels' rotation and direction have to stay perfectly matched with the bot's velocity vector? - That's the velocity vector that you are trying to change as you turn...).

Color me dubious that a Lunacy crab drive will be able to do this at all, much less have an easy time supplying a non-trivial, static-friction advantage to a Lunacy bot.

Blake

Originally Posted by gblake

Are you guys who are saying that a crab will stay in, or mostly in, the wheel's static friction domain, trying to tell me that crab drives are going to be able to perfectly match their wheels rotation and orientation with the bot's motion vector (speed and direction) AND then, when they want to turn, will be able to magically have the wheel trace out an arc without slipping sideways in the least????

The last time I looked, wheels travel in straight lines and turning a bot required the wheels to follow an arc. Twisting the wheel (even slightly) to cause it to travel in an arc would seem to involve a bit of slip (did I mention that the wheels' rotation and direction have to stay perfectly matched with the bot's velocity vector? - That's the velocity vector that you are trying to change as you turn...).

Color me dubious that a Lunacy crab drive will be able to do this at all, much less have an easy time supplying a non-trivial, static-friction advantage to a Lunacy bot.

Blake

While your concerns are valid, you're skipping a few factors. One being traction control, which will be clarified once we get it working. Another being individual wheel power. Individually powered wheels will be able to "power through" turns, much akin to an all-wheel drive car, or a vehicle with sufficiently advanced traction control.

A decently intelligent thing to do with a crab in Lunacy would be to keep the bot in 4 wheel steering mode, and have it act like a car. This allows for the options of doing "crab" (or side to side) motions whenever the situation requires it. A simple combination of a gyro and individual wheel motors will allow a crab to go in any direction regardless of the trailer.

[gblake]

Originally Posted by CraigHickman

While your concerns are valid, you're skipping a few factors. One being traction control, which will be clarified once we get it working. Another being individual wheel power. Individually powered wheels will be able to "power through" turns, much akin to an all-wheel drive car, or a vehicle with sufficiently advanced traction control.

A decently intelligent thing to do with a crab in Lunacy would be to keep the bot in 4 wheel steering mode, and have it act like a car. This allows for the options of doing "crab" (or side to side) motions whenever the situation requires it. A simple combination of a gyro and individual wheel motors will allow a crab to go in any direction regardless of the trailer.

I am focusing on the situation where a bot in motion wants to change its trajectory.

Driving in a in a straight line, traction control (keeping the torque on the wheels less than the torque that will cause them to begin to slip) does sound like a useful thing.

But any attempt to orient those wheels non-trivially away from the bot's motion vector means they have to (and most certainly will) start slipping sideways. Maybe I am wrong, but I'm thinking that if you are slipping sideways, traction control will be dealing with the slipping coefficient of friction, not the static one; and that will put a big crimp in its usefulness during turns.

Eventually the bot will stop sliding sideways and the traction control can be used to move efficiently in the direction the wheels are pointing now.

I can see that a crab drive might turn you a bit more tightly or quickly, but you can still color me dubious that a crab system will be worth the trouble and weight.

Separately, if a bot is standing still and uses a crab drive to head sideways from a standing start, I can see some advantage accruing to a team that needs that motion to carry out their strategy and builds a crab drive. I can also see the trailer wrecking havoc on any attempt to drive a crab sideways without having your wheels slip sideways (away from the direction you point them).

Blake

[gblake]

Originally Posted by gblake

....

Folks,

I hear through the grapevine that instead of tossing down the gauntlet in a good debate, I came across rudely here (as we all know, email is tricky).

For anyone who perceived that way, sorry about that, it wasn't intentional - Instead I expected to see some equally spirited replies come back.

I definitely am eager to see what the real world has to say about both sides of this debate (and the similar ones about other design topics). In a few weeks if I need to enjoy eating some yummy crow, I will gladly bow to the masters.

It should be an interesting season.

Blake