This is a page with research for an interactive installation that’s gonna be presented at forsningsdagene here in Oslo. More info is coming soon..
At the end exhibition at AHO the 13 of june, this project was awarded with a price for best Universal Design. The price was handed out of ITFunk, and included the trip to Kristiansand, hotel and participation on an international conference on Universal Design.
The conferance t4p (technology for participation) was held in Kristiansand the 25-27 of june. even if the reward itself came with quite short notice, in the middle of the vacation, I am glad me and Sniff packed a bag and jumped on the train to Kristiansand. I brought the presentation material used at the exhibition and got a lot of good feedback on the design. I also met some people who might be important in the future for the project.
For the occation I had updated the design on the Emotion Stickers according to the improvements that were mentioned during my visit at Tambartun. I had also prepared a new game, a Melody Game, where the small orange dots (not designed, the Rfid’s are produced like this) activates parts of a song, in this case a common child song. The nodes could be mixed and the task could be to place them in correct order, or you could just explore them with the dogs sniffing and try to get the small melodies to be played on rhythm. I also got people to try the Memory Game and the Stickers placed in a book or at the poster.
One very important thing that I learned was the neccessity of a volume control. The contrast in background sounds was huge when somebody came strolling alone, perhaps on their way to the toilet, and in the coffée breaks, where a lot of people was mingling around. In the last situation the sound from Sniff was almost not hearable. For the second day I had to drill small holes in the front cover to let more sound out. I would definitely equip Sniff with a pretty wide controllable range in volume.
Generally, I leraned a lot during the conferance. Even though some themes were out of the for me well known discourse in designer environments, it was interesting to meet the somewhat more political parts. And from the feedback I got the impression that a lot of them found a little design in the coffée break amusing. Others were even more interested, I now have a handful of contacts in and outside of Norway who actually seem interested in developing Sniff further. It was also fun to tell people about the process and to explain that I actually built the whole dog from scratch, not just tampered an off-the-shelf technological toy, like somebody thought (even though I dont really understand that, he dosen’t exactly look machine made). The most inspiring was when professionals within the field told me they thought the concepts were “spot on” in stimulation therapy for children with visual impairment. This is definitely going to be continued..
I have updated the report to include the concept phase of the project as well, which makes it even longer but gives a more true picture of the whole process. The previous report in Word, found in the post below, is now also updated. Here is a pdf version of Sniff Project Report.
The end of the project has of course been hectic, so the blog has not been updated for a while. After the final presentation on friday 1st of june, we have set up an exhibition 11-15 june. On the grand opening of the exhibition on wednesday the 13th, my project was nominated to best student project in two cathegories; Innovation and Universal Design, and it was awarded in the Universal Design cathegory.
I’m planning to update the process here on the blog chronologically back in time, meanwhile I post a link to the Sniff Project Report.
To see even more pictures from the prototyping process and user tests check out the slideshow.
The final mounting of the motor circuit turns out to be more complicated than I first expected. This takes almost one whole day.
First, where it’s needed, I pack the individual units into something to shield them. The speaker gets a thin layer of foam to soften the sound. The two vibrating motors are placed in the piece of mounting shell that they had in their original products. The smallest of them get a whole piece of the mobile phone shell. It also gets a piece of tape to cover up the moving parts, since it’s to be placed so high up inside the body, where there’s a ot of wires.
Then I have to place the motors at the correct spots, try to make enough room for both them, the battery and a small piece of bread board that I use to connect the wires from the head with the motors. Then I have to assure that they are securely fastened to the shell, solid enough to stand the vibrations. Especially the biggest motor, a quite powerful vibrator motor from a playstation hand control, needs securing plastic walls of the same type as the one I used in the dog’s head to support the bread board. I glue it all into place with plenty of epoxy.
I suddenly come to think of that I need one extra switch; I need to be able to disconnect the battery when connecting the usb cable.
I’m not sure, but I suspect the arduino is becoming electrocuted if I don’t. I have to make a big bright red control for this function so I never miss it. This switch might be handy for easy access to the saving of battery power as well. When the arduino board is ready and loaded the switch can be tuck into the bum inside the plastic shell again. Good that the pear shape give me some extra space in there.
Finally, at the end of the day, everything works. It’s really something special to be able to test the dog out when it’s running on battery power without any external wires. When everything finally was tucked in the head it was a magical experience to test it after testing a totally open system, and now I experience a kind of magic again. When carrying the dog to the workshop for a small cut in the body to make an opening for the newest switch I catch myself to think of him as almost alive. All motors work as well, but the sound box don’t make the same rattling sound anymore when it’s tightly glued to the wall, so I have to operate this one out and make a new box for it (need to buy more eggs). Another special thing I notice just before I call it a night, is that the dog now gets into the multi-reading sphere a lot easier than it did before. Strange - I have not added any extra distance on the nose, quite the contrary actually, since I’ve glued the reader in an even more vertical position the closest it gets to the front inside the shell (a few millis). Have to test this out more carefully. I also notice that the dog needs a stronger neck..
Results from initial testing shows that when on battery the reader detects a tag over and over again, even when the tag is held inside the active sphere! Seems like the battery makes the reader connect and disconnect over and over. I don’t think this has to do with the inbetween multi-reading sphere anymore - when testing with usb connection the behaviour goes back to normal. Have to search around for clues around what might cause this.
I also get to test the reading range of the stones. They are detected at approx 5 cm distance, and the sticker tags are detected at 6 cm. Considering the thickness of the material in the stones this means they’re almost the same. The material haven’t reduced the range of the tags remarkably, even though it is very dense and quite heavy. Let’s hope that’s the issue with the nose as well. If not I might just make it thinner by hollowing it out on the backside. I think 5 cm is a good reaction distance, then you have to come close enough to sniff the arphid but without having to make the parts touch each other.
Preparing for the upcoming sewing session, I cut the tempur foam in suitable slices and wrap the whole toy body in it. I vary the thickness after where the grip should be placed, and also create a flatter appearance of the head through form cut pieces to put in the chins. And I make a bendable tail out of braided hard wires packed in foam.
Then I test the circuit again, preparing for the final mounting of the motors. I create a tag for each feedback motor and test them all out with a piece of new code that directs the feedback to each of them. A mixture of theese feedbacks is what I’m gonna use to create the emotions of the dog.
The memory load issue is solved :) A good thing is it was not even the problem. The code I used had a hangup for another reason, a serial overload, and that’s worse since I created it.. To tangle out the problem and fix it was a simple task for an experienced programmer (not me), and I therefor managed to move into some coding of sounds as well.
After a couple of tests and a lot of mathematics I decided to create the sounds in Sound Forge first, and try to program them later when I know what I want. In Sound Forge you can hear the difference between a square, a saw and a sinus shaped sound wave. There’s quite a big difference. I don’t think it should be impossible to program a sinus wave with electric impulses, even if it’s probably pretty hard. Searching web for examples without luck. Don’t know yet if it’s worth the effort either - it might just make the code a lot bigger and cause an actual memory overload. But the tones where the pitch slide in a sinus wave is not too unpleasant even if they’re built out of square sound waves as ground component, so I might try create some of those. The sliding tones I’ve got now increases or declines in a straight line.
Time for catching up with the 3D design. Handy when my own programming brain also is temporary overloaded. I decide to produce my own objects for the memory game, either casting them (plaster or concrete) or shaping them in the actual material (the best would be stone - but guess I’m landing on wood or cibatool). The idea with the objects is to have no visual clues, not even tactile symbols in this trial version of the concept that I’ve decided to go for. The only detection should be the reaction sounds from the dog. The shapes should be small enough to be enclosed in the kid’s hand, a possible secret treasure. Inspired by the natural irregularity of beach stones, I model out some suggestions in wax.
The cubic shapes is supposed to be softly geometrical with different radius and slightly concave/convex surfaces. The more naturalistic stone shapes is even softer. Since the arphids I wanna insert are flat, inexpensive 20 mm chips, the shapes must be a bit flat to not becoming too big. A wish from the blind center is that they also should be flat enough on at least one side to not roll away on the floor like an ordinary marble.
To make them desirable for blind kids (as well as seeing) I wanna focus on the quality when touching them, both in surface materiality and in shape. Their sublime irregularity should give them a tactile exciting quality, as well as being interresting to collect and also to feel under the fabric covering the dog’s stomach. I’m imagining the objects being produced in enough hard material to have a clinking sound when they touch each other. The best would be if they also had some natural weight (I’m now skipping the idea of a built-in weight in led because that would cause interference with the arphid detection), and since they are so small the material should be as massive as possible. I have to test out how much that weakens the detection though.
Just when I was having so much fun.. When testing the handy headshaped reader together with all the four feedback motors in the code I’m up for an unpleasant surprise. I make a couple of new tags, one for each feedback function to test them out. I finally get them all up and running with the code, after a small discovery that pin no 6 is switched high automatically when uploading code to the arduino mini (Katarina had the same problem, but none of us has find any explanation of why it’s like this). But then the reader starts striking. Turns out that it works ok when reading the same tag over and over again in almost eternity, but when feeding it with different tags the board cuts out reading after just six or seven tags. This really sucks, since my prototype depends on the testing of different type of feedback.
I search the web for arduino and memory overload without getting closer to a solution. I realize it’s time for me to overcome my fear for the unknown, and change my habit of just discreetly sniffing around in the arduino-forum that I’ve joined for the prototyping occasion a couple of weeks ago, where I’m one of the YaBB Newbies, and try get some help from the pro’s. People do seem very helpful in there, and questions are posted in all complexity levels. That makes me feel entitled to try it out. It also helps that this time I suspect what might be wrong (seems like there’s some sort of memory of the arrays that has to be deleted, otherwise it’s filled up with traces from earlier readings fast) but I don’t know how to solve the problem in code. Never liked working much with arrays in Processing either, so far I’ve learned they generally means trouble.
I get an answer within one hour :)
What a beautiful sight when all the spreaking wires are neatly stuffed into the milky white head. I’ve built a wall inside the head that gives the reader a separate room without any pressure on the sensitive solderings on it’s back. At the same time it supports the breadboard so that this is kept in place even when I’m connecting the USB-cable or pushing the buttons. I’ve made one opening for the USB-connection at the back of the head and two more for the push-button and the switch-button on each side of the neck. When it’s all mounted the board is snapped in place between the push-button and the internal wall. The wires are directed out from a fourth hole. I plan to reinforce the neck with a braid made out of the stiffer wires, that is thick enough to manage to hold the head upright and also being bendable.