Re: [Geopriv] RE: For those interested in WiFi location determination

From: Henning Schulzrinne ^lt;hgs@cs.columbia.edu>
Date: Tue Nov 27 2007 - 09:23:06 EST

I hope we're agreeing that Columbia University (where indeed these
coordinates are from) would send civic location data to the PSAP or
the pizza parlor, so that there's no need to convert anywhere. As
Hannes pointed out, that's what the indoor location systems generate
naturally (since that's what's useful for other purposes, such as
equipment or personnel tracking.) If it were to send geo data, the
PSAP or pizza parlor would have to convert to building-and-floor data,
which is clearly what we want to avoid. Since the WLAN positioning
system doesn't use geo coordinates internally, they don't enter the
picture unless somebody were to manually convert rooms or points on
the floor plan to such coordinates, which also seems pointless and
error-prone.

We've had this discussion before (and before that...).

This means that worrying about indoor geo positioning, i.e., 3825,
seems unnecessary. In other words, we've narrowed typical 3825
applicability to the outdoor/campus WiFi AP situation, where we are
back to the roughly-RF-range discussion, i.e., some rough indication
of 802.11 range that's pretty much location-independent (since nobody
is going to map it precisely).

Henning

On Nov 27, 2007, at 8:54 AM, Hannes Tschofenig wrote:

> Hi Marc,
>
> Marc Berryman wrote:
>> We have witnessed first hand how a geodetic location can easily be
>> converted into an incorrect civic address due to differences in the
>> spatial data that is being used for the geodetic to civic
>> translation.
>
>> It is far better to let the local PSAP (in this example) do the
>> translations because they are (should be) using the most up-to-date
>> geographic information and are most familiar with their area.
> But the PSAP is less likely have the floor plan for the building of
> your company site.
>
>> Imagine trying to locate 40.809403, -73.959223 on several different
>> sets
>> of geographic databases. One put the location at the NW corner of W
>> 120th St and Amsterdam Ave, one puts it at say 1207 Amsterdam Avenue,
>> one puts it at the back corner of the Whitter building at the
>> Teachers
>> Collage, and still another places it at the Engineering Terrace at
>> Columbia University. Too many variables (and liability) to be letting
>> anyone perform a geodetic to civic transformation other than at the
>> local level.
>>
> But isn't it most likely the case that the local network operator
> knows the location best?
>

_______________________________________________
Geopriv mailing list
Geopriv@ietf.org
https://www1.ietf.org/mailman/listinfo/geopriv
Received on Tue, 27 Nov 2007 09:23:06 -0500

This archive was generated by hypermail 2.1.8 : Tue Nov 27 2007 - 09:23:21 EST