12 Sep 2005
ubicomp day 1
First day of Ubicomp
Anil and I have been writing some notes about some talks and demos that we've seen using SubEthaEdit. So I'll stick to personal views and opinions about the presentations and demos.
The keynote was given by the Co-founder of Sony CSL (Computer Science Laboratories.)
The main points to take home here was that:
(a) technology takes 30 years to mature (eg. internet, personal computers)
(b) CSL only has 30 researchers
(c) CSL only employs people who create new research fields and not any second tier researchers.
(d) Rekimoto did augmented reality
in 1994 way before the bat system, badge and also work done by some ex-LCE people.
(e) They did some cool videos, although recently their funding has been reduced somewhat
(f) New fields of research they've been looking into are cross between computing and biology. But instead of using biological processes on computing, use known systems theory in computing to analyse biological processes!
The talks were mainly average. The most interesting thing is Carpet LAN
from NTT Docomo Research Labs. The joke paper of the day would be the analysis of chewing sounds
. They tried to justify the paper buy saying we might want to monitor diets. It seemed improbable that you can determine the fat content just by chewing sounds. And also, what about drinks?
Another odd presentation is the u-texture presentation
. Maybe a bit was lost in the translation, but it seemed like the implementation didn't do the concept justice. The concept of rearranging blocks physically in order to program the "textures" to do a specific function. An example of it was rearranging the textures into a shelf so that if you place a CD with an RFID tag on it, it will play the right one with your mp3 player. It seems like if now that you don't want to play music, then you have to rearrange your textures to do something else. I believe this would be much better if the textures were not fully fledged tablet PCs but instead small tangible objects like speakers, displays etc, rather than homogenous bits.
The best and simpliest idea was the Preventing Camera Recording by Designing a Capture Resistant Environment
. Basically, the idea is to detect CCD cameras using the fact that the film between the lens and CCD is an IR reflecting film and hence can be detected with one of those night light camcorder. Then you localise where the camera is and use a projector to shine a light on the lens of the camera.
There are some very cool demos, especially the sponsor demos (eg, they've paid to place a demo here.) Even though the mechanism to choose the demos is dodgy because they have essentially been paid for.
Cool demos include the "QR Code, RFID and Coin operated Coke vending machine", Suica direction finding pole and Palm Vein Biometrics.
So there's this advanced vending machine
where you can buy coke using a variety of methods. One way is to use a QR Code and QR Code enabled phone. So you rock up to a vending machine, and then you are thirsty and buy a coke. No change in your pocket, well, have a QR code enabled phone? So what you do is go to the coke website using your phone. This could probably be done by clicking on a QR Code already on the vending machine, or maybe manually typing it in. Then you pay for a one time QR code online which will be sent to your phone. Show the QR Code to the vending machine's camera and it will recognise it and give you a coke!
The other way you can do it, is just have an RFID payment system attached to your phone. They sell such phones here already (that I'm lead to believe.) So you just use that like you would use Octopus (in HK) or Oyster (in UK) or Suica (in Japan.) Obviously this is the easiest way, but not many phones have this yet.
The funny thing though was I asked the guy where I could find one of these machines, he said that there are 4000 of these in Japan. Asked him where they are in Tokyo so I could try it out. Turns out there's only 1 in Tokyo (so where are the others?) and it is in Akihabara. One of the other guys in the stand was very enthusiastic about finding such a machine there!
Anil took a photo of the machine here:
Another cool sponsored demo was the NEC direction finding pole that can be distributed around say a train station or museum, or basically anywhere you need to find directions. So imagine you are in a huge shopping mall and you want to find a particular store. You swipe your Suica card (RFID payment card) over the store you want to go to which is listed in the shopping directory. Then you walk up to these direction finding poles around the environment and place your card on the reader. The reader will recall where you want to go, and then the top of the pole will rotate and flex to point to the direction of the place you should be heading. It is very cool and simple. They have an LCD screen on it, but it doesn't add that much to the pole. The pole itself is very well designed and useful. I think this would definitely work and it is easy enough for people to use.
Here's a short movie of it in action:
Other cool thing is the contactless palm vein biometric authentication system
. Basically it has an infrared camera which is able to highlight veins on your hand and apparently that is a good biometric?
The posters aren't particularly interesting. There are some pretty funny ones which I'll cover in the next post.
You can reply to me about this on Twitter: