Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
Cellphones Crime Security

Smart Phones Could Know Their Users By How They Walk 96

mirgens writes "Technology Review has a short article on new work on gait analysis with the accelerometers built into many smart phones. The work was done at the Norwegian Information Security Laboratory ('Nislab'). The need for more security on mobile devices is increasing with new functionalities and features made available. To improve device security, Nislab proposed gait recognition as a protection mechanism — in other words, if somebody else walks away with your phone, it locks up. While previous work on gait recognition used video sources, for instance to identify people in airports or secure buildings, the Nislab researchers collected the gait data using a Google G1 phone containing the AK8976A embedded accelerometer." What if you're running from a mugger and want to dial 911?
This discussion has been archived. No new comments can be posted.

Smart Phones Could Know Their Users By How They Walk

Comments Filter:
  • by TheLink ( 130905 ) on Saturday October 09, 2010 @07:41AM (#33844390) Journal
    How's it going to work for ladies who carry phones in handbags etc (many don't use the same bag all the time).

    Those ladies actually have a high chance of getting their phones stolen - the thieves steal the bag with the phone inside etc. Many ladies typically don't wear any garments that have pockets. Or worse there are pockets but they are sewn shut so that you don't use them by mistake and make an ugly looking bulge ;).

    As for guys, it might work, but I doubt a significant number of us would intentionally buy a phone with this. If us guys wanted an antitheft phone we'd just buy a really cheap phone. Or "customize" it to the point where its fence value drops immensely.

    A way to reduce phone theft is by phone makers making their phone IMEIs very hard to change (and ensuring that they are unique), and the cellular providers blocking stolen phones (even globally).
  • by vtcodger ( 957785 ) on Saturday October 09, 2010 @07:45AM (#33844406)

    Great!!! I've always wanted a phone that won't work if I am jogging, riding a horse, skiing, walking on ice, sprain my ankle, having a gout attack, riding a bicycle, fleeing for my life, .... When can I expect to be able to buy this wonder?

  • by Anonymous Coward on Saturday October 09, 2010 @08:17AM (#33844488)

    Interesting. It might be difficult to get the software flexible enough to deal with all possible paces for the same person, but consider this: each person's legs are a specific length. The way a person walks is intimately tied to the length of their legs, as well as the shape of their hips and their overall weight. It is not unreasonable to suspect that sufficiently-accurate accelerometers could distinguish between peoples' movement no matter how fast they were moving (walking, running, whatever).

    The main way to fool the device, I think, would be to purposefully chop your steps short. But that is very difficult to do for very long, and of course you couldn't very well imitate somebody else by doing so.

    I do not buy that placing the phone in a different bag would have any effect on the estimated gait. Also, things like riding in a car or on a bicycle would look extremely different from walking, so the device obviously shouldn't lock up in those situations. Furthermore, it would be impossible to distinguish between being a passenger and a driver, making the very possibility of locking a phone when it detects you're in a car problematic. Anyway, obviously as an anti-theft device, this will, if it is successful, be marketed as a feature to consumers, meaning that the owner of the phone can choose whether or not to use it. If the phone is routinely locking itself due to this feature, users will simply turn it off.

  • by realityimpaired ( 1668397 ) on Saturday October 09, 2010 @08:24AM (#33844502)

    I'm more worried about the fact that my gait changes based on whether I'm wearing shoes with a lift or not, and how high that lift is. Admittedly not as much of a problem for the male folk in the room, but in my closet there's flats, as well as shoes with heels varying from 1-4", and I very rarely wear the same height of shoe two days in a row.

    I don't like the idea of my phone locking me out if it thinks I'm driving. In theory that would be temporary, until you stop moving. But I *really* don't like the idea of my phone locking me out because I'm wearing a different pair of shoes than I did yesterday.

  • by Anonymous Coward on Saturday October 09, 2010 @08:31AM (#33844518)
    You had me until, "Somehow I can talk on a cell phone without disregarding all awareness of my environment, it comes from not being a bovine idiot with a deer-in-headlights outlook on life."

    Every single one of them thinks the same thing. Newsflash: you're one of them.
  • Gait vs password? (Score:3, Insightful)

    by Freddybear ( 1805256 ) on Saturday October 09, 2010 @09:41AM (#33844772)

    Don't completely disable the phone. If the gait analysis comes up "wrong" then require the user to enter his password again.

  • by tepples ( 727027 ) <tepples.gmail@com> on Saturday October 09, 2010 @09:45AM (#33844794) Homepage Journal

    If you are a passenger, you are just outta luck.

    In that case, such legislation would never get out of committee because it would impose an undue burden on people who carpool or ride public transport. All seats in a car or especially a bus (except one) are reasonably safe for making phone calls.

  • by tverbeek ( 457094 ) on Saturday October 09, 2010 @10:01AM (#33844880) Homepage

    There's no way this feature is going to be 100% accurate, and certainly not in 1.0. Every "recognition" technology ever has an error rate, and this will be no different. If it's intended as a security feature, the developers will have to calibrate it to err on the side of denying access, otherwise they'll open themselves up to criticism (and probably legal suits) over its failure to provide the advertised security. This means that there will be false positives, in which the phone denies its legitimate owner access (wearing new shoes, walking on unusual surface, injured, tired, listening to "Beat It"), and that will get the phone chucked across the room in pretty short order.

"When the going gets tough, the tough get empirical." -- Jon Carroll

Working...