Falling Around

MOOL: Construction Planning

This post could just as easily be title, “So, you want to build a multi-touch table”. The concept is pretty straightforward, and based on a little idea called FTIR (Frustrated Total Internal Reflection). You can read more about the general idea by doing a quick Google search for “FTIR multitouch”.

The specifics for the MOOL system were derived after a bit of research by the project team on a variety of websites. After reading and digesting all we could, we settled on the following setup (basically, it’s designed almost in “layers”, so I’ve listed the top-most layer down to the bottom-most):

- Roscoe Grey projection material
- Silicone compound (more on this later)
- Acrylic (1/4″ thick, picked up at Lowe’s)
- IR LED’s
- Unibrain Fire-i B&W Camera (more on this later)
- Projector

The camera sees the light reflected by the fingers of someone touching the screen, and then we use this data to power the interactivity of the programs. (We still haven’t decided if we’re going to be using TouchLib or Reactivision just yet, although are based around using OSC and the TUIO protocol to communicate with the Flash apps that we’ll be developing).

For our silicone solution, we first tried to use a silicone caulk that we purchased at Lowe’s, and dillute it in Toulene in order to make it thinner and easier to spread on the acrylic. However, the first caulk that we purchased didn’t dissolve in the toulene at all! So, we did some more research, and finally settled on a compound called Lexel. We were able to purchase a few tubes at our local hardware store for about $12. It dissolves very well in the toulene, and when it dries it’s incredibly clear. We do have a few mistakes (bubble, uneven places, etc.), but it turned out pretty good for a first attempt. In the future, we’re thinking about dilluting the solution down to a very thin liquid, and then using a paint sprayer to spray it onto the acrylic, which should hopefully eliminate bubbles and give us a better way to spread it evenly across the entire surface.

After some research, we settled on the Unibrain Fire-i B&W Camera. We selected the B&W camera instead of the color one because it actually outputs a high-quality picture, faster than the color one does. We obviously don’t need color for this application, so there was no reason not to get it. It was a bit more expensive, but we’re thinking that the investment should pay off in spades in terms of responsiveness. If you do decide to get this camera, be sure to get a compatible lens (we ordered everything from 1394store.com, and it was all listed in the same place) WITHOUT an IR coating (since we’re trying to detect IR light, after all!). In addition, you’ll want to create a simple bypass filter (that will block out visible light and only let in IR light). We created ours by cutting out two small squares from the film inside an old floppy disk, and placing these (overlapped) on top of the CMOS sensor (under the lens of the camera). While we haven’t finished the MOOL yet, so we don’t know if it will work in the final form of the table, in our tests it blocked out all the ambient light of a very well-lit room, and we were still able to see the IR LEDs just fine.

For more information on why we chose the products we did, and to see some other examples of multitouch tables, be sure to check out the forums at NUI Group. Also check out AudioTouch, another multi-touch table project, where we found a lot of good advice for getting started.

Multi-touch Object Oriented Learning (MOOL)

I’ve started a pretty exciting project at work recently, and I’m willing to bet that a good portion of the upcoming posts on this blog will be devoted to it. It’s a multi-touch interactive table, similar to the Microsoft Surface and countless others that seem to be coming out these days. We call it MOOL (I like to pronounce it “mule”).The goal of this project, however, is to enhance the learning of students at Heartspring, by utilizing this new technology to transcribe existing teaching methods onto the multi-touch platform.

For example, one task that the students perform on a regular basis is called “sorting”. Sorting involves taking a set of objects (either the objects themselves, or cards with pictures of the objects), and then having the student sort these objects into groups. For example, cars vs. fruit. The student will repeat this task over and over, to help them learn basic reasoning skills. Throughout this process, the teacher will help the student, and at the same time record the results (i.e. “Billy got 3 correct and 4 incorrect”).

With MOOL, the idea is that custom software will be developed that takes this existing learning technique (sorting), and transcribes it onto the multi-touch platform. The objects will be digitized, and the student will simply drag them from their starting place (in the middle of the screen) across the table to one side or the other, depending on which group the student thinks the object belongs in. It’s the same as moving the objects on the table as before, only now it’s a touch-screen table instead of a wooden one. In addition, rather than the teacher recording the results, the computer will be able to automatically determine if a correct or incorrect response is given. This frees up the teacher to concentrate on the student throughout the entire activity, rather than worrying about recording correct responses. In addition, the computer can calculate statistics on the fly, giving the teacher an up-to-date overview of the progress of the student. The computer can also employ additional teaching techniques, such as “back-chaining”, to further enhance the exercise. Finally, it can instantly store and process the data being gathered, presenting an up-to-date report to any teachers or staff regarding the student’s progress with the “sorting program.”

As you can see from this example, the MOOL system has great potential for really revolutionizing the work that we do with these students.

Now that we’ve covered the general idea, I’ll cover the specifics in my next post.