[Cyborg] indoor positioning system using magnetometer
tomm.fire at gmail.com
Tue Jul 10 22:05:44 UTC 2012
This paper has some hard data on using a magnetometer to determine
In the Data Analysis section, they show that although they were accuracy
to 1 meter 75% of the time, they also had some instances of being off by
50m in a 70m hallway, for an average accuracy of 4.7m. Adding in
heuristics bumped them up to 88% accuracy, but still with some errors >60m.
But add this kind of data to step counting, as the calibration procedure
appeared to do (and every smartphone has an accelerometer), and it's
probably a worthwhile system. Cool stuff.
On 7/10/12 2:20 PM, Tomm wrote:
> It looks to me like their "secret sauce" is calibrating step size, then
> counting steps to determine distance traveled. I think the whole
> magnetic field thing is just to determine direction of travel on the
> map, ie, orientating the map relative local north. Which of course will
> be incorrect if you walk by a fridge compressor that wasn't on the last
> time you walked by...
> As far as rough vs. fine location, maybe they have an NFC touchpoint
> when entering the building to establish "you are here"? It's a solution
> only a geek could love.
> I see the future: someone asks their phone, "where are the canned
> tomatoes?", and their phone guides them to aisle where the tomatoes were
> until last week, when the shelves were reorganized and no one bothered
> to update the app. And the only reason the user asked is that they know
> where the tomatoes used to be, and they want to know where they are _now_!
> On 7/10/12 7:33 AM, Eric Boyd wrote:
>> It's hard for me to believe that the magnetic field at each location
>> would be globally unique - at least at that resolution. Seems more
>> likely that you have to give the app a rough location (e.g. this
>> building, or here's a nearby wifi/cell tower location...) and from there
>> it can pin-point you.
>> Also, anyone else notice the curious total fixity of the cell phone
>> angle and location with respect to the camera & floor? They must have
>> had some kind of fancy rig to hold them that rigidly. Maybe some smoke
>> & mirrors? It also makes me wonder how well the indoor positioning app
>> works if the phone tilts, or shakes...
>> Still, really cool technology. I imagine it would be super useful for
>> things like google glass :-)
>> Cyborg mailing list
>> Cyborg at lists.noisebridge.net
> Cyborg mailing list
> Cyborg at lists.noisebridge.net
More information about the Cyborg