[757labs] Augmented Reality Indoors?

Michael Ross 757labs at oneross.com
Sun Apr 22 16:24:35 EDT 2012


New broadcom chip promises to help of this front, approximately 1" position
resolution including vertical:

http://singularityhub.com/2012/04/19/new-smartphone-chips-will-pinpoint-your-exact-location-down-to-the-inch-even-inside-buildings/


On Thu, Sep 15, 2011 at 3:58 PM, Steve Nelson <snelson at webapper.com> wrote:

> damn this mouse (it click send for me. thanks glitchy mouse!)
>
> I'm pretty sure I have a tripod.
>
> Steve
>
> On Thu, Sep 15, 2011 at 3:58 PM, Steve Nelson <snelson at webapper.com>wrote:
>
>> I'm still messing around with this, but here's the latest version with
>> nearly accurate measurements for the walls of the first floor.
>>
>> Does anyone own a camera that will take panoramic pictures that we could
>> borrow for an afternoon? I'm pretty
>>
>> Steve
>>
>>
>> On Tue, Sep 13, 2011 at 1:17 PM, Steve Nelson <snelson at webapper.com>wrote:
>>
>>> That's exactly right. I've been goofing around with that a little this
>>> morning. Getting the exact coordinates is a bit tricky because each of
>>> their different views is off by a tiny bit. But I don't think it'll matter
>>> much. Incorporating two floors will be a little hard too.
>>>
>>> Steve
>>>
>>>
>>> On Tue, Sep 13, 2011 at 1:07 PM, Trevor Lewis <trevorl.salad at gmail.com>wrote:
>>>
>>>>
>>>>  Hmmm, actually, I think that's an easy one to solve. Find the
>>>> coordinates of the outer extents of the building/space, and do a little
>>>> creative division. I think my phone will pull down 7 or 8 digits of
>>>> precision for GPS. It should be straightforward to splat that into an grid,
>>>> and overlay the floorplan of the lab right on top of that. It won't be
>>>> perfect, longitude is variable, but for the purposes of mapping a space it
>>>> should be just fine.
>>>>
>>>> On Tue, Sep 13, 2011 at 12:33 PM, Steve Nelson <snelson at webapper.com>wrote:
>>>>
>>>>> I knew someone would know how to do it! You should come in sometime
>>>>> and hook up with one of the Android developers. I bet after a little
>>>>> tinkering you could get something to work. If you/others can produce a GPS
>>>>> coordinate, I'll give you a way to plug it into a google map that zooms
>>>>> beyond 21.
>>>>>
>>>>> Steve
>>>>>
>>>>>   On Tue, Sep 13, 2011 at 11:58 AM, Trevor Lewis <
>>>>> trevorl.salad at gmail.com> wrote:
>>>>>
>>>>>>
>>>>>> I'll start off by saying "Hi!" I've been following the 757labs
>>>>>> exploits for a while, and just haven't had a chance to make it to one of
>>>>>> the open house events. Anyway...
>>>>>>
>>>>>> This is an idea I've spent a lot of time thinking about. Wifi is an
>>>>>> abysmal way to map a space, too much noise, too many things to degrade the
>>>>>> signal. If I were forced to do it I think the easiest way might be to use
>>>>>> multiple routers blanketting the same area at different frequencies. A
>>>>>> network of small circuits, sensitive to each frequency, will allow you to
>>>>>> determine the relative strength of the transmitters at a particular
>>>>>> location. A person walking through the space is going to occlude
>>>>>> transmitters and receivers, changing the map of the RF in the room. I just
>>>>>> don't know if a person is going to make a large enough occlusion that it
>>>>>> can reliably be detected. Honestly it's just a reallllly cool smart motion
>>>>>> sensor, or an excuse to wear tinfoil hats. :)
>>>>>>
>>>>>> Some issues would be gathering the data from the receivers, and
>>>>>> turning that into a map. You're also not going to be able to reliably tell
>>>>>> the orientation of the person moving through the space. It's a safe bet
>>>>>> most people will be facing the direction they're moving, but y'all could be
>>>>>> throwing a moonwalking competition or something.
>>>>>>
>>>>>> Have you thought about using something like the edge-finding
>>>>>> algorithm that Goggles uses? It's actually a really good at what it does,
>>>>>> and would be even better if you could make Goggles not search Google, and
>>>>>> just work from a limited data set consisting of patterns built from the lab
>>>>>> space.
>>>>>>
>>>>>> The ADK is also an easy way to marry external sensing with a phone. I
>>>>>> had it set up and working when it was first released, but got distracted
>>>>>> with other projects before I got it to do anything more interesting than
>>>>>> run the sample/demo code.
>>>>>>
>>>>>> _______________________________________________
>>>>>> 757labs mailing list
>>>>>> 757labs at 757labs.org
>>>>>> http://757labs.org/cgi-bin/mailman/listinfo/757labs
>>>>>> If we knew what we were doing it wouldn't be research.
>>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> 757labs mailing list
>>>>> 757labs at 757labs.org
>>>>> http://757labs.org/cgi-bin/mailman/listinfo/757labs
>>>>> If we knew what we were doing it wouldn't be research.
>>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> 757labs mailing list
>>>> 757labs at 757labs.org
>>>> http://757labs.org/cgi-bin/mailman/listinfo/757labs
>>>> If we knew what we were doing it wouldn't be research.
>>>>
>>>
>>>
>>
>
> _______________________________________________
> 757labs mailing list
> 757labs at 757labs.org
> http://757labs.org/cgi-bin/mailman/listinfo/757labs
> If we knew what we were doing it wouldn't be research.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://757labs.org/pipermail/757labs/attachments/20120422/e1adb680/attachment.html>


More information about the 757labs mailing list