Team 2175 Logo Motion Robot Design

[Aren Siekmeier]

This is the product of our first week of brainstorming, prototyping, and design.

http://img834.imageshack.us/img834/6152/orthoextended.th.png (http://img834.imageshack.us/i/orthoextended.png/)http://img824.imageshack.us/img824/7809/orthoelec.th.png (http://img824.imageshack.us/i/orthoelec.png/)http://img96.imageshack.us/img96/4813/frontextended.th.png (http://img96.imageshack.us/i/frontextended.png/)http://img535.imageshack.us/img535/6056/orthoframe.th.png (http://img535.imageshack.us/i/orthoframe.png/)

Our strategy:
-Score pieces quickly. Quick ground to peg time.
-Be maneuverable. This speeds up piece retrieval and helps us place them accurately (particular by strafing).
-Make it simple and reliable. We hope to attend 2 or 3 events this year as opposed to previous years' 1, so we want it to last.
-We also hope to have more of it CADed before construction so it can come together more cleanly and much, much more quickly, leaving lots of time for code, testing, and practice.

Our solution:
We determined that strafing was very near a requirement for placing the pieces because tank turning would be very limiting and almost always slower even with practice. We therefore arrived at the holonomic drivetrain that we prototyped (http://www.chiefdelphi.com/forums/showthread.php?threadid=89478); the competition robot will use four 4" dualies. It favors the forward direction (2 CIMs per side) for faster field traversal, but still strafes (with 1 or 2 RS 775s per wheel) and turns very well over short distances. The frame is an octagon to increase stability (?) and keep turning effortless. It will be welded aluminum box tube, 1" by 1/8" wall.

For a lift mechanism we have decided on an elevator very similar to that of Team 25 from 2007. We got the idea from the 2007 Behind the Design book, and watching match videos confirmed that it worked well. We determined that we needed to pick up off of the ground, because we think most of the pieces will be there at some point, and we don't want to be limited to the slot. The elevator can reach the top peg with just 2 stages and a gripper that rotates the game piece from pickup position (horizontal) to scoring position (vertical). It will be powered by an FP and should be able to lift at 12 ft/s (!), not counting inefficiencies, friction, etc. The gripper is not yet finalized, but will likely be a simple pneumatic clamp.

All the electronics, battery, and motors will fit within the 12" tall octagon (not shown). The minibot will conveniently sit on top behind the elevator, at the appropriate deployment height. Design work to be done includes the gripper and the minibot and minor details for the elevator and the drivetrain.

We haven't decided on a name yet. Suggestions welcome!

[Flyer522]

Looks very cool! Great work! Does the CAD have 6 Jaguars on it?

[DiscoKittyPrime]

This looks like a good design. I do have some questions for you that are born out of curiosity and not criticism.
1.) How do you plan on picking up the tubes from the floor easily with the telescoping lift centered in the bot?
2.) The placement of your batter seems a little difficult to reach from the views that you have provided. Have you guys planned on an easy way to change out your batteries?

Just curious about your design choices but it looks like a solid design non the less. Best of luck.

[pfreivald]

What gear reduction are you thinking for the FP?

[1493kd]

Wow I showed a fellow mentor your photos and he asked me when I drew up our design. We have the same exact design. Hopefully it works.. we may have to compare notes.. Good luck

[dodar]

If anyone on your team likes watching the UFC, they could name your minibot Brock and just call your robot The Octagon.

[electron]

I like how you included the electrical components in your model (wish we had done that last year... :rolleyes: )

I do, however, want to know how you plan on changing your battery. it looks pretty buried/stuck in the bottom.

looks nice!

[AllenGregoryIV]

That looks really similar to our design.
We are only using 4 CIMs on the drive and are wheels are at 45 degree angles to the direction of normal travel(Picture (http://www.chiefdelphi.com/media/photos/36085)). Omni drives like this actually travel faster in that orientation in the orthogonal directions (They travel faster then each individual wheel velocity) Source (http://users.cecs.anu.edu.au/~nmb/papers/omnidrive_dynamics4.pdf)

Our lift is using spectra cord instead of timing belt but overall an extremely similar robot design. I hope we both do well in competition.

[Chris is me]

We are only using 4 CIMs on the drive and are wheels are at 45 degree angles to the direction of normal travel(Picture (http://www.chiefdelphi.com/media/photos/36085)). Omni drives like this actually travel faster in that orientation in the orthogonal directions (They travel faster then each individual wheel velocity) Source (http://users.cecs.anu.edu.au/~nmb/papers/omnidrive_dynamics4.pdf)

They travel faster in that they can be geared higher than two motors would normally push in that direction; however drives like these allow the user to configure a particular direction to drive with more than the equivalent of two motors of "power", allowing a higher free speed / lower gear ratio in the primary direction of travel.

In short, "yes" but "no". :P

[AllenGregoryIV]

They travel faster in that they can be geared higher than two motors would normally push in that direction; however drives like these allow the user to configure a particular direction to drive with more than the equivalent of two motors of "power", allowing a higher free speed / lower gear ratio in the primary direction of travel.

In short, "yes" but "no". :P

I am not sure I understand the first sentence here. A robot with four omni wheels placed 90 degrees apart will travel approximately 1.4 (or square root of 2) times faster than wheel velocity in it's orthogonal directions. At directions parallel to any wheel rotation the robot will only be able to travel at wheel velocity like any two wheeled robot. This is not saying it is geared any different. The robot is actually moving faster than any one wheel is spinning.

Your second statement is correct that by adding motors you are able to get a higher power, this is always the case. However this also makes doing the calculations for holonomic control of the robot very challenging since different wheels are able to travel at different velocities and will have different accelerations.

So to sum up yes their drive train will have more power and acceleration when traveling forward and backward than one that uses 1 CIM on each wheel (assuming consistent gearing). The drive with 1 CIM on each wheel will have a greater top speed when strafing and consistent control to all 4 wheels. These are just two different design and both have there ups and downs. And as always you can gear any drive train to have more speed when you sacrifice torque.

[Chris is me]

I am not sure I understand the first sentence here. A robot with four omni wheels placed 90 degrees apart will travel approximately 1.4 (or square root of 2) times faster than wheel velocity in it's orthogonal directions. At directions parallel to any wheel rotation the robot will only be able to travel at wheel velocity like any two wheeled robot. This is not saying it is geared any different. The robot is actually moving faster than any one wheel is spinning.

I meant that the net speed, all things considered, is potentially greater for a robot with four motors pointing in the intended direction of motion than a holonomic system, if the gearing is adjusted to compensate for motor differences.

Sorry about any confusion.

[Dr Theta]

How does this pulley system function? It doesn't appear to be anchored to any of the lift components and therefore would essentially free spin like a belt in a loop. Just curious I actually love the design and have seen some ways we could fix some of the problems we have been having in our own design. Thanks for posting. The video of the omni drive is also pretty awesome I hope I get time to sneak over to Mariucci arena to get a look at this thing first when it's done.

[Dr Theta]

Never mind, I believe I spotted the anchor point, very clean. I'm impressed.

[Aren Siekmeier]

Does the CAD have 6 Jaguars on it?

Right now it only includes the jags for the drivetrain. There will be at least one more for the lift mechanism, and possibly more depending on minibot deployment, gripper design, etc.

1.) How do you plan on picking up the tubes from the floor easily with the telescoping lift centered in the bot?
2.) The placement of your batter seems a little difficult to reach from the views that you have provided. Have you guys planned on an easy way to change out your batteries?

1. The gripper will have two positions powered by a cylinder. Vertical for starting the match within the frame perimeter and for scoring. And horizontal for picking tubes up off the ground. In the latter orientation, it will have to reach forward and over the bumpers to the ground.
2. The battery can just be lifted a few inches, tipped, and dropped out the bottom. That's the idea anyway, we haven't checked that the clearance will actually allow that. If not, there will probably just be thumb screws or something on the brackets holding in. Either way, it will be accessed from the bottom.

What gear reduction are you thinking for the FP?

A 20lb arm climbing 10 ft in 1s is about 270W, appropriate for an FP it seems. With a 1" radius pulleys, we're looking about 8.5:1 off the FP. We're also considering two RS 775s, but we're unsure of the transmission options for that.

And actually, with four equally powered wheels, the robot goes slower diagonally, just with more torque (the wheel speeds don't add up, but the torques do). However, we can go twice as fast in one direction by doubling the CIMs there. Since we don't care as much about strafing, we demote them to RS 775s.

Yes, the belt is fixed to the final (gripper) stage. Powering the belt pulls this up and down at the same time as the first stage (or perhaps in sequence...). We realized that a closed belt would simplify things a lot, but at first were failing to see how to make it work. But conveniently the length lost between the stages will always equal the height change of the final stage, so the belt stays the same length (assuming its vertical). We actually verified this in Solidworks down to a thousandth of an inch. Pretty sweet. Of course, who knows if the real mechanism will work that well. The double clamp on the gripper stage also allows us to quickly tension the belt as needed while keeping it all working smoothly.

@Dr Theta: We might also be attending Lake Superior, depending on funding, so perhaps we'll see each others' handiwork there?

[AllenGregoryIV]

And actually, with four equally powered wheels, the robot goes slower diagonally, just with more torque (the wheel speeds don't add up, but the torques do). However, we can go twice as fast in one direction by doubling the CIMs there. Since we don't care as much about strafing, we demote them to RS 775s.


I'm not sure how to explain this better than other people already have before me.

This gif file on this page (http://www.chiefdelphi.com/media/papers/2390) shows the relationship between standard drives, omni drives, and Mecanum drives. Ether does a great job of explaining the math and if you have questions he is usually happy to answer them.

The source (http://users.cecs.anu.edu.au/~nmb/papers/omnidrive_dynamics4.pdf)I sited above also shows how the other motors are able to induce velocity (not force) perpendicular to the direction of rotation of any one wheel.

In fact wheel torque is .707 (1 over root 2) that of any individual wheel when traveling at a 45 degree angle to any wheel.

[Dr Theta]

We're actually attending 10000 lakes. Anyway it looks very good. Out of curiosity what were you planning on using for the belt material?

[mplanchard]

Very nice and thank you for sharing.

For all robot designers, learn from this example:let SolidWorks or any other cad system take the math values out many decimal places. Don't round and put in values. Keep them out to as many places as the software can handle in the part and assembly files.

The in the drawing you specify tolerance and precision for machining.

Marie