You are here

37 posts / 0 new
Last post
Posted by:
Cuauv Leader
Cuauv Leader's picture
Tue, 04/11/2017 - 3:41am

6.6 Collect and Classify Samples (Recovery) Discussion

Some ideas we had for the recovery objects:

1) classification of color and shape - four different objects of 2 colours and 2 shapes (ie green and red and triangular prism and rectangular prism). The scoring table has the corresponding shapes and colors. Points for placing any object on any of the tabling targets. More points for placing objects on the correct colour. Full points for placing objects on corrector colour and shape.

2) pipes - pvc pipes about 8” long with “x”s on the ends to stop them from rolling. Four pipes of 2 different colours. on the table, there are corresponding colored rectangles in different direction, (possibly could adjust orientation between rounda). Points for getting the pipes on the table anywhere and more points for placing them in the indicated orientation.

It would be awesome if we could lock down the objects for this in the next couple weeks. Designing, manufacturing, and testing a manipulator is hard to do in a couple weeks over the summer. Hopefully if at least the object shape is finalized earlier more teams will attempt the object manipulation task!

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Sat, 04/15/2017 - 1:24pm

Samples

These are great ideas thanks for sharing, and appologies for being slow it getting things out.  Lately, I've been more away from my home that in it.

There has been some internal discussion about this and other tasks.  I'm going to push to get the ideas out to the competitiors and finialized quickly.

 

I like #2, but maybe using four different colors instead of two.

Posted by:
Bumble Bee
Bumble Bee's picture
Mon, 05/08/2017 - 10:58am

Both sound good to us. We'll

Both sound good to us. We'll prolly comment more when your design comes out haha. #1 sounds like the k-nex we had three years back though.

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Tue, 05/09/2017 - 12:22am

Classify

I have the design finished, and should post it tomorrow.

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Wed, 05/10/2017 - 12:10am

Collect and Classify

Here is the design for the tower and the samples (collect)

https://drive.google.com/open?id=0B6R3K-IZtXFXUlhCWVJsSHJUWkU

 

And for the locations to lay the samples on (classify)

https://drive.google.com/open?id=0B6R3K-IZtXFXQTEySVB0dnozdHc

The ellipses are 4" (10cm) larger along the major axis.  The minor axis is ½ the major.  So 12" x 6" and 16" x 8" (30.5 x 15.2cm and 40.6 x 20.3cm).

Colors can still be changed.

 

 

Posted by:
Cuauv Leader
Cuauv Leader's picture
Wed, 05/10/2017 - 5:56pm

Some Questions

1) In the past each of the recovery objects has been worth equal points, but since the small objects will be presumably harder to grab and are more difficult colors to detect, I am thinking they will probably be worth more points?

2) Would the orientation of the different objects on tower be fixed or would it be randomized? For example, can we assume the blue object is always across from the green one?

3) For the placement of the pipe within the oval, does the pipe need to be oriented along the major axis of the ellipse? Or does the pipe just need to be within the ellipse (fully, more than half)? The pipes look like they would not be immune to rolling around, hopefully the threaded rods at the end will prevent most of the rolling, but will points be based on their point of contact or final destination.

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Sat, 05/13/2017 - 5:14pm

Some Answers

  1. You are correct, the smaller object(s) will be worth more points.
  2. The order (and orientation) of the different colors will be fixed.  My intention is to add a little bit of color to the holder in an obscure place, so that the divers know how to put the classify objects back on the stand.
  3. Full points for getting the correct colored pipe totally within the correct ellipse.  Partial points for getting the correct pipe on the correct ellipse, and some points for getting the pipe on the table.

The threaded rod will keep the pipe from rolling (what are your concerns?), however it won't be immune from being potentially moved from vehicle down thrust (I'm going to weight these a bit).  At the moment, points will be awarded based on the final destination.

Posted by:
Bumble Bee
Bumble Bee's picture
Sun, 05/14/2017 - 1:55am

Couple of comments/questions:

Couple of comments/questions:

1. Can we lengthen the threaded rods so that even if we grab with a single grabber it doesn't slip or slide out due to the currents? 

2.  Can we get more dimensions for the structure: 

  • How long is a side of the square frame(top down view)
  • How tall is the entire structure from the base to the objects to pick up
  • How tall is the pinger on the stand at the center
Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Sun, 05/14/2017 - 2:40pm

A couple of replies/answers

  1. As currently drawn, they are 2" (51mm) long threaded rod.  I was also going to look into 3" (76mm) and 4" (102mm) threaded rod
  2. Sure!  The tower is made from 1" PVC.
    1. As currently draw, the PVC used for the "square frame" top down view is 12" (305mm).  If I make any change, I'll only make it larger.  But I think I'll stick with how it is currently drawn, unless someone see's a problem.
    2. Based on the preliminary rules, the top of the tower (include the classify pieces) is 4-5ft (1.2-1.5m) off the bottom.  I'll stick to that.
    3. The pinger is 2ft (0.6m) off the bottom.

What I just noticed is that I didn't include a side view, so you could see that the larger objects are higher up than the smaller one.  I'll fix that image and repost it.

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Sun, 05/14/2017 - 3:07pm

Image

The image for the tower and samples has been edited.  The above link still works.

The offset distance between the larger and smaller samples is 4" (10cm).  Similar to the size of the vertical tube "U"

Posted by:
Cuauv Leader
Cuauv Leader's picture
Thu, 05/18/2017 - 1:26am

Offset Heights

We must say that we are against the height offset between the samples. This is a significant change to how the recovery tower has been in previous years and with the rules being released so late, it would be very difficult to modify designs around this. 

It is standard that we, like most teams, must design our vehicle without knowing the exact mission and try to design to be flexible. When the rules were released in December/January this still left us time to make modifications to our design and fabricate these. However with the rules being released later and later it becomes neccesary to design and make everything except the actual gripper before the rules are released. The differing heights will require either specific placement of gripper so that the vehicle is always on the outside of the tower or an extremely long gripper. Either of these would be reasonable given enough time to make design modifications, but doing so with just over two months to Transdec (wow, really?) will severly cut into our testing time. 

The unique design of the samples, coupled with the different difficulty objects, and two new colors for objects provides plenty of novelty and challenge for the task, particularly when it is already so close to competition.

If you are against making all objects the same height, we would ask that the tower is made wider to lessen the chance of vehicles contacting other portions of the tower.

What are your thoughts Dave, and any other teams?

Posted by:
Boulder Robosub
Boulder Robosub's picture
Thu, 05/18/2017 - 9:06pm

Seconded

We agree with cornell on this matter. It is already dificult to fabricate a retrevial method for these objects and So few teams even attempt this task in the first place, what is the point in adding to the dificulty over them being on a level plane? 

Posted by:
Bumble Bee
Bumble Bee's picture
Fri, 05/19/2017 - 12:26pm

Thirded!

Now that Cornell mentioned it, better if they were of the same height. Also what diameter are the threaded rods gonna be? It challenging for the something that small to have varying depths as well. Its hard for the robot to tell what depth the object is at from each other.

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Fri, 05/19/2017 - 10:11pm

OK!! OK!!!

   :)

Well, the idea was to add another level to the challange (both literally and figuratively) .  However I agree, the added complexity is unacceptable at this late date (and I appologize for that as well).  I want you to do the best you can with the time you have.  I'll make them all the same height.  I'll update the image this weekend and upload the changes.

As drawn, the threaded rods are 1/4" x 2"

Posted by:
Cuauv Leader
Cuauv Leader's picture
Sat, 05/20/2017 - 2:59am

One More Request

Thank you, we do like the idea and hopefully it can be in next years competition, but with our last day to CNC the grabber being tomorrow...

One other thing came up while we were playing around with our tower yesterday. We discovered that the threaded rods sticking out of all the samples really makes it seem like the sub is descending onto a tower of spikes that are just waiting to crack a DVL or jam a thruster. We would like to propose having the threaded inserts only extend out of the one side of the tube going into the tower. We don't think this should have much of any effect on the actual task, but will make testing it less nervewracking. Thoughts from any other teams? Perhaps there is a better way to hold the pipes on the tower?

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Sat, 05/20/2017 - 12:43pm

LOL

I agree.  Yikes!  Why is this you last day for the CNC?

I didn't think about that aspect of the tower.  I could have them only extend out the "bottom".  They could still have the potential of rotating 180° (which wouldn't move them a huge distance).  I can also put some acorn nuts on the top.  So it would just be the acorn nuts on the "top", and the threaded rod extending fully out the "bottom".  The acorn nuts would have a slight chance of stopping the rotation.

I'm open to other suggestions.

FYI, I've updated the picture of the tower.  The link above should still work.

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Sat, 05/20/2017 - 3:07pm

Kinder Gentiler Samples

Does this look safer and more inviting?

Samples w/ Acorns

 

Posted by:
Boulder Robosub
Boulder Robosub's picture
Sun, 05/21/2017 - 12:19am

We Like the Acorns

This makes us feel much safer about our DVL

Posted by:
Bumble Bee
Bumble Bee's picture
Tue, 05/23/2017 - 11:07am

Agree with the suggestion!

Agree with the suggestion! How could we almost have missed this out! 

So we've been testing on our tower today and realised that the pipes are really small and difficult to detect. In previous years where we had the knex which was similar, there would be a big border around or the structures would be large and significant enough to be detectable. Can we enlarge the diameter of the samples by like double so that it would be easier to see? The alternative is to make an obvious border around the samples for us to track properly.

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Wed, 05/24/2017 - 6:51pm

Bigger Samples

I originally looked at using 1" PVC pipe for the samples, but that seem to make them very large.

I could revisit the 1" samples, but I can also look at putting a border around the samle (the same color) to help with ID and tracking.

Posted by:
Alan Albritton
Alan Albritton's picture
Thu, 05/25/2017 - 1:02pm

End caps on samples

Can endcaps be placed on the end of the samples? We found anytime there is a slight angle to the pipe when dropped the pipe follows the angle its at, as opposed to falling straight down. If endcaps are added it prevents water from flowing through the pipe thus it falls relatively straight down no matter the angle its pitched at when dropped. This would make it much easier to drop on the sample board in a predictable manner. 

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Fri, 05/26/2017 - 11:35pm

Capping the ends

Interesting

Will the samples fill with water with the currently drilled holes used for the 1/4" x 20 thread?  (Probably)

I can always make sure they are filled before securing the threaded rod in place, or make the "bottom" holes (the ones opposite the acorn nuts) slightly larger to make sure water can enter.

I can also drill out the endcaps (1/4"... 1/8"...?)  to make sure they are filled with water...

 

Thoughts?

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Sat, 05/27/2017 - 2:54pm

Bigger samples, take 2

Here's an image with the current origial size of 1/2" PVC, and one with 1" PVC.

Tower top

I probably could go to 1.5" PVC, but with end caps, it will start to get heavy.

Go with bigger PVC, or with a border (of the same color as the sample).  I can also color the vertical portion of the holder to help with visual identification (as shown in the drawing).

Thoughts?

Posted by:
kushal.agar@gmail.com
Sat, 05/27/2017 - 6:43pm

Thoughts

Interesting discussion. I'd vote for bigger PVC, which will definietly help make the samples easier to see. The mass of the samples is not that big of a concer, especially since most of it is offset by buoyancy.

 

I wouldn't want the stationary part of the tower to be colored, because if you do, vehicles might confuse those sections with the sample itself and end up grabbing the wrong piece of the structure.

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Sat, 05/27/2017 - 7:50pm

Mass of samples

PVC and Nylon are heavier than water, and the "sample" will be filled with water, so at the moment, there will be no buoyant force to offset the weight of the "sample". 

I could add some floatation inbetween the threaded rod to help offset some of the weight.  However, some teams like a heavier object to pick up.

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Sun, 05/28/2017 - 5:23pm

Endcapped samples

Here's what the 1" PVC longer sample would look like with end caps.  I am leaning towards making the bottom holes bigger to allow water to fill the tube.

End caps

 

Posted by:
Alan Albritton
Alan Albritton's picture
Sun, 05/28/2017 - 11:35pm

Endcaps and larger tubes

I'm definitely a fan of larger tubes and the end caps. Making the bottom holes larger would definitely speed up the process of allowing air to escape.

Posted by:
Alan Albritton
Alan Albritton's picture
Wed, 05/31/2017 - 7:40pm

Length

With the end caps being added will the 8 inches and 12 inches include the end caps? Or will the actual pipe be 8 or 12 inches plus the bit of length the caps add?

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Wed, 05/31/2017 - 10:26pm

Sample length

Since I had all ready cut some items to length, I kept the length of the pipe the same (8" and 12").  So the entire length of the sample will be longer with the addition of the end caps.

Posted by:
Cuauv Leader
Cuauv Leader's picture
Thu, 06/01/2017 - 11:21am

Classify Table Size and Colors

In the preliminary rules it states the table for samples will be 4'x4', but the figure with the ellipses appears to only be 2'x2'.

Is the table actually 2'x2', or is there a 1' white border around the ellipses or should the ellipses be scaled up in size?

Also, what is the thickness of the black line outlining ellipses?

Do you have specific paints/tape picked out for the different colors of the objects?

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Fri, 06/02/2017 - 5:02pm

Ellipses

The image is in full scale.  To make my life easier, I send the image off to get printed onto a banner, I then cut out what I need and glue/tape it on to the surface.  I squeezed all of the four ellipses into a 2' x 2' square, so that I can get more printed out per banner (which is usally a 2' x 4' banner).  Each of these ellipses will be placed on their own 2' x 2' square to make up the 4' x 4' table.  If you take the perliminary rules and swap out the circles for these ellipses, you'll have the final task.

The black outline is 0.25" (6.4mm) thick.  That was a good question, since I hadn't finalized on a size.  I've update the image to reflect the 0.25" black border, but here is the link as well

Ellipses

The colors used:

  • Red: 255, 0, 0
  • Green: 0, 255, 0
  • Blue: 0, 0, 255
  • Orange: 255, 102, 0
Posted by:
Cuauv Leader
Cuauv Leader's picture
Fri, 06/02/2017 - 7:12pm

Locations and Orientations of Ellipses

Okay, so if I am understanding correctly, the layout doesn't actually look like the image you posted, instead there will be one ellipse in each 2' x 2' quarter of the table. This brings up a couple questions:

1) Are the ellipses centered within their 2' x 2' box?

2) What is the orientation of the ellipses?

3) Which color correspoonds to which quadrant of the board?

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Sat, 06/03/2017 - 3:27pm

Location and Orientation

Sounds like you are understanding things correctly.  I'm writing the final rules this weekend, so hopefully this will be better cleared up after that.

  1. They will be ~centered within their 2' x 2' box.
  2. They could be in any orientation
  3. If you are asking for an order for the ellipses (i.e. relative positions to each other within the 2x2 array), they could be in any order.
Posted by:
Alan Albritton
Alan Albritton's picture
Fri, 06/09/2017 - 3:22pm

Colors

Is it possible to have the orange changed to yellow? The orange and red are fairly close in color and may cause issues.

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Sat, 06/10/2017 - 2:55pm

Red and Orange

I've used both of those colors (Red and Orange) close together before without any problems (as far as I know).  My concern with the yellow, especially on the table, is that it would get washed out with the dapled sunlight.

Posted by:
Bumble Bee
Bumble Bee's picture
Fri, 06/16/2017 - 8:45am

How far would the tower be

How far would the tower be from the samples? I would imagine it to be side by side and not some wild distance away else we would have to search like mad for it.

Posted by:
dknovick@yahoo.com
dknovick@yahoo.com's picture
Fri, 06/16/2017 - 6:11pm

No Wild Distance

While some of the images in the preliminary mission are "not to scale", the relative distances shown aren't a bad representation.  So the classify table will be placed right around the boundary of the octagon floating on the surface.

Log in to post comments