Tuesday, September 11, 2012

Ubicomp 2012: some cool things

Ubicomp 2012 just ended, right here in Pittsburgh. I was a student volunteer and had a great time. As I guess is the norm at conferences, I'm a bit overwhelmed by all the information coming at me at once, so here's an attempt to sift through it a little bit by summarizing some talks/papers/posters that I liked. I'm in Week 3 of my PhD program, so this is not going to be super focused.

An Ultra-Low-Power Human Body Motion Sensor Using Static Electric Field Sensing by Gabe Cohn et al. Currently to track motion, we use accelerometers. Their device lets you wear one sensor on your wrist, it can detect when any part of your body moves, and the power usage is about 1-10% of an accelerometer's.

A Spark Of Activity: Exploring Informative Art As Visualization For Physical Activity by Chloe Fan, Jodi Forlizzi, and Anind Dey. So your Fitbit (pedometer) counts steps, right, but it only gives the data back to you in graph form. That's cool, but somehow it's more fun (and indeed, motivating) if it's in a little bit poppier form. She's developed some visualizations and found that people do prefer more abstract visualizations for display/fun purposes. (Graphs are still better if you're looking for concrete numbers.)

Lullaby: A Capture and Access System for Understanding the Sleep Environment by Matt Kay et al. Put this box in your room while you sleep, it'll tell you if there are any disturbances or anything that might be hurting your sleep. You can't tell what's wrong when you're sleeping. Difficult task, and well executed. Also, there are privacy concerns when there's a camera in your bedroom! (They address this.)

RubberBand: Augmenting Teachers' Awareness of Spatially Isolated Children on Kindergarten Field Trips by Hyukjae Jang et al. A system that'll alert a teacher if kids go wandering off. Solves a real problem, and does it in a clever way: clusters the kids based on proximity, then detects if any cluster is getting too far from the other kids, not from the teachers.

Providing eco-driving feedback to corporate car drivers: what impact does a smartphone application have on their fuel efficiency? by Johannes Tulusan, Thorsten Staake, and Elgar Fleisch. They gave drivers an iPhone app to mount on their dashboard that would give real-time feedback so they can learn to drive more efficiently. Cool not so much for the gas-saving effect (3%) as for the idea that they maintain their skills even after they take the phone app away. I'd love to know how they're driving a month or a year later.

SpiroSmart: Using a Microphone to Measure Lung Function on a Mobile Phone by Eric Larson, Mayank Goel, et al. Got lung problems? Need to measure lung function? Toss your $2000 home spirometer, use a smart phone.

MoodMeter: Counting Smiles in the Wild by Mohammed Hoque, Javier Hernandez, et al. They set up cameras and big screens around MIT that would detect who's smiling and who's not. Cool way to measure happiness of different places (in a sense), interesting interactions (everyone would try to make it see them as a smile or not), nice face recognition. What's this good for? As is, just seems neat, but it makes me think of a few other ideas. What if you had reminders to smile in your house? Making a smile causes you to feel happier... how far does this effect go, and is it worth trying to do it repeatedly? Or do we get into annoying cheesy "smile!" dystopias? Also, does the number or percent of smiles actually tell you anything useful about an area?

Enhancing the "Second Hand" Retail Experience with Digital Object Memories, by Martin de Jode et al. They put RFID tags and QR codes on stuff in Oxfam (second hand) shops in the UK, so you can hear the original owner's story behind something you buy. Cool. They had 50% more sales, but they couldn't attribute that to the tags. It makes the world more magical: you could find a secret story in any nook and cranny. Imagine you buy a second-hand fridge, find a QR code in the drawer, and the original owner left a message about a big party they had where they stored beer there. Or a recipe of their favorite thing to make, or a log of when people repaired this dang thing. Also, I'd love to know long-term how this affects sales; I'd imagine it'd make people more likely to buy and sell things. Increased use of second-hand shops is good for the environment, your wallet, etc.

Making Technology Homey: Finding Sources of Satisfaction and Meaning in Home Automation by Leila Takayama et al. They interviewed people who did home automation projects, from the basic to the extreme. People who made their own stuff had more connection to it than people who bought premade solutions. Sometimes other people think they're wasting time, but it brings them some meaning. One guy had a "Canyon Cam" on his vacation home so he could see this view he loved, one guy rigged up a system to take pictures of his cat when it got scared off the counter, one guy would turning off the lights in the whole house as a subtle signal to his daughter to go to sleep. These are surprisingly cool and surprisingly meaningful, and I guess the crucial insight is that people enjoyed them the most when they were connecting with their home, not controlling it.

Augmenting Gesture Recognition with Erlang-Cox Models To Identify Neurological Disorders in Premature Babies by Mingming Fan et al. Put accelerometers on babies, tell if they're doing Cramped Synchronized General Movements, which correlate with Cerebral Palsy. Instead of watching an hour of a baby moving, doctors can just watch 10% as much video to detect CSGMs for sure. Useful in the medical field (clinical trial going on now) and uses a cool variant of hidden Markov models.

Identifying Emotions Expressed by Mobile Users through 2D Surface and 3D Motion Gestures by Celine Coutrix and Nadine Mandran. This is not about general emotion detection, but rather about intentional actions people might use while expressing certain emotions. Neat study: when triggered once a day or on demand, users would do whatever gesture expresses what they're feeling, then rate how they felt on a PAD model (pleasure, arousal, dominance). This is interesting when creating apps that take emotional state into account (e.g. "shake the phone angrily to restart if it freezes"; I don't know whether this would be good or bad, but you get the idea.)

What Next, Ubicomp? Celebrating an Intellectual Disappearing Act by Gregory Abowd. Okay, everyone was talking about this the whole conference, so I've got to mention it. His point, as I understand it, is: "Ubicomp" used to just mean anything with small computers. Now that field is so big, the Ubicomp conference can't continue to be just anything with small computers. Imagine having a "Personal Computing" conference nowadays; it's too broad. Maybe new conferences need to form for subfields or something. Discuss.

Demos, Posters, etc.

Touche by Ivan Poupyrev, Chris Harrison, Munehiko Sato. I feel like I've heard about this, but it's still cool. Make anything touch- and gesture-sensitive.

SenSprout: Inkjet-Printed Soil Moisture and Leaf Wetness Sensor, by Yoshihiro Kawahara et al. Print out some conductive ink and it, well, senses soil moisture and leaf wetness.

Design of a Context-Aware Signal Glove for Bicycle and Motorcycle Riders, by Anthony Carton. Of course I want this.

uSmell: A Gas Sensor System to Classify Odors in Natural, Uncontrolled Environments by Sen Hirano, Khai Truong, and Gillian Hayes. I've never seen a smell sensor before. This allows lots of possibilities.

Big talks:

The keynote by Steve Cousins, CEO of Willow Garage, was cool. A "state of personal robotics". Saw how finding a beer is easy but opening the fridge is hard, folding a towel is easy but finding the corners is hard, letting go of things at the right times is hard, etc.

The talk before the conference by Jun Rekimoto and colleagues was cool too. FlyingBuddy2, a drone you can control with your mind; glass that you can turn transparent or opaque; a drawer system that can tell what's in each drawer; armband that programmatically activates your fingers, a smile sensor before you can open your fridge; a really cute potted plant on wheels that drives around; a fork that makes different noises based on conductivity when you touch food to your mouth. Some of these things I'm not going to argue are super useful, but they're all cool.

No comments:

Post a Comment