Click to see the full WalkType poster

OK, so maybe this isn’t the greatest invention for society, but purely as a technological trick, it’s pretty neat.

A group from the University of Washington has figured out how to adjust the sensing mechanism behind a mobile keyboard to take into account the effects of walking — making it easier to accurately text a friend or send an email from a phone while on the move.

The project, dubbed WalkType, uses the accelerometer inside a phone to sense when someone is walking and make the adjustment on the fly. The visible keyboard doesn’t move on the screen, but the behind-the-scenes sensing mechanism adjusts for the effect of stepping.

The team (Mayank Goel, Computer Science and Engineering Ph.D. student; Dr. Leah Findlater, UW Information School postdoctoral fellow; and Dr. Jacob Wobbrock, UW iSchool professor) has implemented a prototype inside a custom iPhone app, and they’re now planning an Android implementation, which will allow for better access to the underlying operating system.

Maybe this is a commentary on my own bad habits, but WalkType was one my personal favorites among some 70 projects on display last night during an open house and poster session at the Computer Science and Engineering Department’s annual Industrial Affiliates meeting.

Gabe Cohn

The annual Madrona Prize, handed out by Madrona Venture Group, went to UW student Gabe Cohn and the “Humantenna” project, which uses a receiver on the human body to determine a person’s position in relation to electrical noise emanating from a home’s wiring system.

With a bit of machine learning, a computer program can analyze the variations in that signal to figure out the particular gesture a person is making — raising the possibility of interaction similar to the Xbox 360 Kinect sensor, without actually requiring the Kinect sensor. Other applications include the ability to control home systems such as lighting by touching a wall, after programming the computer to recognize and act on that gesture.

Cohn is the lead researcher on the project, working with Microsoft Research’s Desney Tan and Dan Morris, as well as UW professor Shwetak Patal, whose work involving home electrical signals recently won him a MacArthur Genuis Award.

The People’s Choice Award went to “Paper to Digital,” a project by students Nicola Dell and Nathan Breit that uses a mobile device to accurately scan and digitize paper reporting sheets from health workers in the field.

(Note: The references to Leah Findlater and Jacob Wobbrock have been corrected since the original post.)

Comments

  • Jamesian

    Humantenna was mentioned in The Economist Babbage blog yesterday

    http://www.economist.com/blogs/babbage/2011/10/innovation

  • Jack

    Interesting project, but doesn’t this encourage lack of awareness for those who text? It’s already irritating as is to avoid walkers that text, but to make it easier for them? It’s like saying, “Here, let’s make this GPS system more convenient/accessible for drivers who program it while driving.”

    • Joe

      As an owner and I’m assuming chronic user of a phone in this manner, one should have the ability to multitask enough to not plow into people when walking. Consider the alternative of standing (often in the way) of people while texting; the same hassle is there but for the ones around you. 

    • Guest

      I use an SMS program with camera integration. That way, as I’m walking and texting, what I would normally see with my eyes is simply projected into the background. Eventually I expect to replace my eyeglasses with a similar solution.

Job Listings on GeekWork