after observing the competition at saint Louis, and how tank drives were far superior in most respects this game, i remembered something Ive read about called melty brain.
the idea is to have a tank drive robot, that spins very quickly. by altering the amount of power to each side at different angles within a spin. omni-directional movement is achieved. a standard 2 speed tank drive should be able to spin fast enough to make this work ( to some degree) in a fast gear.
this type of movement has one big advantage that i can see, a very intimidating defensive robot. because of all of the added energy stored in the robot, even a robot with tank treads would have a difficult time passing a defender with this type of drive-train. especially if it could move in any direction.
all of the open source software can be viewed above.
my question is, how do you think this would work in a match, would it be legal, and how tough would the spinning robot need to be. would it pose a danger to other robots on the field. ( the goal is to intimidate, not actually destroy) if the answer to the last question is yes, what is a reasonable speed to spin. ( our robot this year with 6 wheel tank, powered omni wheels on the ends, could spin approx 4 times per second.)
also what other potential does this type of control scheme have in FIRST. if it werent for the trailers i think it would have been THE best drivetrain for 2009. and it would have been an intimidating contender for a simple omni directional drive that rookie teams could pull off. ( direct driving 8 inch wheels from cimpleboxes) without buying ANY additional hardware.
no. the idea is to achieve omni directional movement. defense would be one possible use, however it cannot hold more energy than the same robot traveling at its top speed. the goal of this would be to make a robot more difficult to push. in past years teams without any predictable means of scoring have had the ability to simply defend. however when more teams than ever have multi speed gearboxes, adjustable stances, and other methods of altering their drive to serve to score swiftly as well as push through defense, this is no longer really a viable option. having a spinning robot removes traction and torque from the factors that influence how difficult it is to push another robot. this is a way to increase defensive ability, without adding any significant complexity such as gyros or legs. the intent would not be to damage any robot, but to make an omni directional robot very simple mechanically, and to gain an edge defensively. because the robot is always spinning it always holds the same amount of kinetic energy as it would going in a straight line. in battle bots the spinner bots typicaly spin more than 1000 rpm. i was thinking more like 120 - 360 rpm.
Isn’t the rate of “linear” motion related to the rate of rotation in this design? So wouldn’t slowing down that rotation, in turn, slow down how quickly you “move” around the field?
I also question the safety of this design on the FRC scale. Even if you keep the design relatively compact (say 18"x18" without bumpers) and rotate at the bottom end of your suggested range (120rpm), the outside corners of your bumpers are already moving faster than 1000 ft/second. Much worse are the hard corners of your frame (which are traveling 800 ft/second). There’s a reason that videos like this one give warnings about “melty-bots” being able to kill you if they’re out of control. There’s a reason why these types of robots are mostly seen in robot fighting, because their very functionality relies on large amount of kinetic energy.
The common difficult issue for “melty brain” designs is a reliable heading indicator. If the robot doesn’t accurately know where it is at least once per cycle, it can’t accurately translate.
Even fully functional melty brain designs achieve rotating translation speeds that are very slow compared to the vehicle purely translating. It would be an incredibly easy robot to just juke around.
Mauler 5150 v.s. Jabberwocky
Battlebots Televised version.
Watch what happens.
P.S.(I am totally not posting that because my mentor drove the winning robot) noo…why would I do that.)
It’s an interesting concept for driving, but I’m pretty sure the mechanisms team would stage a coup after their end effector fell of for the umpteenth time due to excessive stresses on the joints.
Though your numbers are off by an order of magnitude or two, you are right in theory. It would be much faster to just turn a tank robot and drive in the desired direction than it would be to wait for the robot to “drift” in that direction.
There’s actually a robot with these in one of the labs here at the University of South Florida. I believe a master’s student designed it for his thesis a few years back. I’ll see if I can get a picture next time I’m there.
the main goal of this is not to make the robot faster, it is to remove traction and torque from the factors that influence how difficult a robot is to push, thus allowing a team with single speed gearboxes to effectively block a team with 2 speed gearboxes.
Curiosity:
How does this eliminate traction and torque as factors?
On traction:
A single speed robot CAN block a 2 speed robot.
Assume:
The single speed robot is oriented so that the motors do not backdrive when being pushed.
The tractive force between the robot and the surface is larger for the single speed robot than the two speed robot.
The two speed robot has enough torque to provide more force than the frictional force provided by the wheels and “spin” the wheels/tires.
The frictional force between the robots contact surface (ussualy wheels/treads) is what keeps it from being pushed. For the most part, in FRC, **any robot with AM tread can be an effective “blocker”. **
in my experience teams that attempt to block sideways are easily “spun” also AM tread is not even close to the grippiest material used, the blue mcmaster tread is much gripper.
If you watch videos of meltys (google: spinning tortoise), you’ll note as soon as they contact something they bounce away in the opposite direction. All the offensive robot would have to do is bump into the melty, and the reaction force would move it back and to the side. The offensive robot would move back a little but much less because it has much better traction.
I admit it would be cool to see, but i don’t see it being a viable design for FIRST.
