h a l f b a k e r yResults not typical.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Please log in.
Before you can vote, you need to register.
Please log in or create an account.
|
We create an autonomous robot with a revolving carousel of magazines filled with mosaic tiles. Software on the robot indexes the colors of an image fed into it, according to a palette created from the colors of the tiles, creates a raster, and then lays out each line of tile, adding a small bit of adhesive
to the tile or floor before pasting it down. A 0,0 origin would have to be set and boundries defined for the autonomous robot to be able to fix it's position and find the co-ords to place the tiles. I'm trying to bake this with Lego Mindstorms right now, but I thought I'd release the idea into the wild to see what happens to it. Lead screws and stepper motors could be used in a more primitive version, but would severely limit the size of the mosaic floor creatable.
(?) Dorkbot
http://www.dorkbot.org/dorkbotsocal/
People doing strange things with electricity. [Novysan, Oct 04 2004, last modified Oct 21 2004]
[link]
|
| |
This would put me out of a job, although my knees are thanking you for the sentiment. |
|
| |
Ideally, this would be done to spec in a factory, with the tiles glued to a scrim. The finished product could then be rolled up and shipped, to be later grouted into position onsite. |
|
| |
How about creating a team robot approach. Instead of using revolving magazines of tiles. You have a base unit for storage of the raw tile material with a built in cutter and dispensor that the roving robot links up to load the precut tiles. The rover robot departs from base to layout the pieces to the floor.
1. Navigation:
a) The Base unit would act as datum point for the rover. The Base would scan the room using a combo sensor package of primitive visual recon as well as LADAR.
b) The Base would 'direct' the rover to its course of layout.
c) If the layout is nonlinear the Base Unit would drive thourgh the course to map out its location. Then choose the datum point most suited for the situation.
2. Tile Prep and placement
a)The raw tiles are kept in the base station for processing. The base would then cut pieces to fit the style of mosaic you wish to lay. [ie: squares or angular cut] The tiles would be adhered to flexable layout strips.
b) The docked Rover would spool up these layout strips into itself.
c) Once spool magazine is filled , rover would be dispatched to location for them to be layed. Once in location the grout is layed down and tiles set into place.
3. Multiple Rovers:
a)Base/Rover approach would make more efficent layouts. Since multiple rovers could be dispatched from the same base station and work on diffrent parts of the same project in parallel.
b)Specialty Rovers could be made to set tile in hard to reach places. Dependant upon the layout and location of mosaic [ie: tight corners, low over hanging cabinets.
4. OR
a) Go with 'ldischler' and measure out the location of where you wish tile to be set. Enter that data into your computer. Design the mosaic in photo editer and email the design to the factory to assemble on scrim sheets. |
|
| |
I really like the queen/swarm approach idea,[MySoulWanders]! And the scrim, [ldischler]! I think using a base unit (the queen) as a datum would work well, lidar and rotation counts for the rovers (drones) and things like ultrasound and IR for object avoidance and safety. Hmm, I just inherited a second RCX..I love extreme halfbaking... |
|
| |
Glad you liked the prior annotation. In referance to your software side of image creation, try this idea to bake your concept.
1)Create a scanning/imprinting fixture.
2)Gather up alot of scrap tile chips that are realatively uniform in size, yet can be any angular shape.
3)Place pieces into fixture one at a time, to low res scan/image the piece. While at the same time it imprints the bottom of the piece with a serialized number and orientation marks.
4)After all tiles are scanned and marked. Load images into one of those "photomosaic" programs to use as "pixels" to create whatever Final mosiac you like.
5)a)Only the "Queen" would have to be physicaly modified to use a rotary indexed place head to read alignment/serialized imprint, then place and align tiles on the strips (scrims) rolls.
b) The "swarm" would not need any modification.
6)Email me pics/video of your Baked Robomosaic masterpiece(s) |
|
| |
Put your flat surface on a conveyor belt. Then move it under bridges that hold stacks of tiles. There is a series of bridges, one for each color and each bridge has to hold multiple stacks side by side over the full width of the surface. |
|
| |
Move the surface - drop the appropriate tiles - move by the surface by one tile length - drop the appropriate tiles - ... That is way easier to control. |
|
| |
If the surface is made from a fabric you can apply the adhesive with rollers through the back. When the adhesive hardens it can also stiffen the fabric. |
|
| |
I'm going to try and get this together for an upcoming dorkbot [see link]! I'll post a link on my website when I have something to show. |
|
| |
Back when I was designing halftone patterns for an inkjet printer company I used to look at them under a stereo microscope and try to imagine what a whole floor of tiles would look like, in a subway station for example. My imaginary tiler had a broad cylindrical rubber wheel, like a steam roller's, with indentations for the tiles. As the machine rolled along slowly, the indentations on the top of the wheel would be filled with tiles of the right colors, surrounded by grout, then set and pressed down as they reached the bottom. The problems with the scheme were numerous, but the mental image of designing a picture, halftoning it, then driving along with it emerging behind me was very appealing. |
|
| |
I had just helped to make a printer mounted on a vehicle that printed marks on bolts of fabric which unwound as it drove, so this didn't seem too farfetched. |
|
| |
I was thinking of this idea as a commission service: color matching robot akin to paint matcher in hardware store squirts custom color tiles according to pixels in image. # of pixels (and so colors used) would depend on ultimate size of image. I could send in a picture of something I wanted as wall art (for example, 40 foot tall image of myself in full regalia dancing a merry jig) then receive a shipment of coded scrims to assemble on the wall or floor. |
|
| |