point, then if there is any structure to the plan you are referencing from
(either a 3D model or a 2D floor plan), you should be able to locate the
reference point.
There is a small problem with the "Front" door, because few plans would
label a door "Front", but I expect a couple of simple conventions from that
would work.
Elevators are marked in most cases, although sometimes the signage on them
is non-existent, but "Elevator A" is meaningful in most circumstances on
most plans.
That would be my goal: if you had a 3d model or 2D floor plan, in any
reasonable machine readable form, you should be able to locate the reference
point. We need to be able to align the building axis (or really, the floor
plan or model axis) with the earth's in order to be able to specify the
offset in cardinal form. All that takes is a north arrow on the plan.
We are then primarily limited by the fact that few floor plans are accurate
("as built"). My architect friends say less than 1% of plans are as built.
They may well be good enough however. In most hotels, for example, there
are good enough floor plans.
While you can use a geo offset from a geo reference point, you can't usually
use a floor plan with that without a benchmark AND an axis alignment. I do
think it's more likely we can get existing plans with just the north arrow.
I don't think I've ever seen a floor plan with an accurate benchmark
(building floorplates as part of a land survey would work I guess, but then
you have to overlay the floorplan on the footplate, and we get back to "as
built" problems).
Brian
On 10/20/09 11:06 AM, "Carl Reed" <creed@opengeospatial.org> wrote:
> Gabor -
>
> This is a follow on from yours and Martin's emails.
>
> First, let me say I agree that we need a common, well understood mechanism
> for describing an arbitrary location inside a building.
>
> That said, how does the application creating the Interior Location
> Extensions (ILE) payload "know" where the elevator is (or for that matter
> any fixed asset in the building)? I understand that semantics etc are
> outside the scope of this group. However, is there an assumption that I am
> missing? Is the actual compilation of all the reference information
> happening in some external app and then being used in an ILE payload?
> Relative location means to locate a place relative to other landmarks. One
> has to know what and where the landmarks are for relative location to be
> useful. Perhaps at the very least, any document on IL needs to provide some
> guidance about where the reference locations (landmarks) come from.
>
> Sorry if I am being dense here.
>
> As to the CRS, an arbitrary Cartesian coordinate system for a building may
> or may not be oriented to some real world projected system. If one considers
> the building to be a singular self contained feature with no external
> references, then your assumption is correct - no need to worry. However, if
> we wish to plan ahead (as Martin also points out), then I would suggest that
> we need to make sure that any interior positioning approach can easily be
> integrated into a larger content infrastructure.
>
> For example, within about 2 or 3 years, every building in Germany will be
> represented in a high quality 3d urban model in which the buildings are
> georeferenced (lat, long, and altitude) to a common CRS. Currently the work
> is for the exteriors of the building but they are rapidly moving into the
> interiors of the building. And in Korea, there is a massive government led
> program called the u-City. In the u-City, sensors permeate the information
> fabric of the entire city, including complete integration of in-building
> sensors and sensor networks. These sensors are all network connected (and IP
> enabled). The Koreans are doing considerable work in defining how to map
> from one spatial reference system (such as outside building WGS-84) to
> another spatial reference system (such as an inside building local
> coordinate system). This is the group that Richard Barnes was so kind to
> meet with in Germany a few weeks ago. Similar work is also going on in
> Taiwan. In both cases, they are using absolute coordinates to describe the
> location of the ip enabled devices.
>
> I just want to make sure that this work integrates well with other, strongly
> related activities - such as navigation, routing, augmented reality,
> information fusion for decision support, and so forth.
>
> The integration of in-building and outside building information has been a
> weakness in the AEC and GIS communities for decades. Nice to see that
> perhaps many of these integration problems are being addressed and hopefully
> solved!
>
> Regards
>
> Carl
>
> ----- Original Message -----
> From: <Gabor.Bajko@nokia.com>
> To: <creed@opengeospatial.org>; <br@brianrosen.net>;
> <dstanley1389@gmail.com>; <geopriv@ietf.org>
> Sent: Monday, October 19, 2009 11:29 PM
> Subject: RE: [Geopriv] Interior Location Extensions internet draft available
>
>
>> Hi Carl,
>>
>> Some answers below:
>>
>>> That said, how does the application "know" what the main elevator is and
>>> where is is location or what is really meant by Floor 4
>>
>> The above are needed for navigation, but not for positioning. The first
>> version of the draft was not intended to provide support for navigation.
>> That capability would however be possible to add by including a link to a
>> floorplan in a vector format, which would have objects and places tagged
>> with the names of the INT elements (eg SVG).
>>
>>> Further, what is meant by "south" and "east"?
>>
>> I thought that we can safely assume that in a relatively small object like
>> a building (relative to the dimentions of the Earth) , the axes of the
>> Cartesian coordinate system can be assumed as being N-S and E-W (N being
>> the true north) cardinal directions. Is this a wrong assumption?
>>
>>> How does the local CRS related to the real world CRS (such as WGS 84)?
>>
>> From positioning pow, it doesn't have to. If you are looking to navigate
>> in and out of the building, then you need it. But for that to work you
>> need to have the WGS84 coordinates of the floorplan's edges and in/out
>> gates.
>>
>> Let me know if I am terribly wrong.
>>
>> - Gabor
>>
>>
>> -----Original Message-----
>> From: geopriv-bounces@ietf.org [mailto:geopriv-bounces@ietf.org] On Behalf
>> Of ext Carl Reed
>> Sent: Monday, October 19, 2009 3:08 PM
>> To: Brian Rosen; Dorothy Stanley; GEOPRIV
>> Subject: Re: [Geopriv] Interior Location Extensions internet draft
>> available
>>
>> I also believe this draft has merit and am willing to help it progress.
>>
>> That said, I am working multiple other standards activities related to
>> indoor/outdoor navigation that have a role in a variety of domains, such
>> as emergency response.
>>
>> For example, there is a new activity just beginning titles "Open Floor
>> Plan Display". From the activity plan:
>>
>> "Conversations between industry specialists often break down in semantic
>> frustration for many reasons. A primary reason is the lack of a common
>> framework and strategic planning to evolve this framework. Secondly,
>> documentation of building elements and building service performance
>> requirements has exploded. Which parts of the living documents need to
>> comply with long term exchange requirements for wide spread re-use?"
>>
>> The work proposed envisions a next generation 9-1-1 using modern devices
>> and open exchange languages to depict buildings and alerts in simple,
>> interoperable formats via lightweight web services. One objective is the
>> ability to read and pass along sensor alerts before emergencies such as
>> fires are large enough to be observed by a by-stander phoning in. Data
>> identities center around building addresses, standard building data
>> representations, and alert classifications. The work will build on
>> existing standards and practices as well as previous collaborations
>> between the OGC, NIST, buildingSMART Alliance, GSA, and other
>> organizations in the AEC community.
>>
>> One reason I mention this work is that there are a variety of issues
>> related to expressing location within a building and then using that
>> information to say guide first responders.
>> From the draft-stanley-geopriv-int-ext-00.txt document
>>
>> "within X meters of the point which is 10 meters south and 15 meters east
>> of the main elevator on Floor 4"
>>
>> I understand that perhaps the focus of the document is not on semantics.
>> That said, how does the application "know" what the main elevator is and
>> where is is location or what is really meant by Floor 4. Does there need
>> to be a statement about where the building floorplan information comes
>> from?
>> Further, what is meant by "south" and "east"? In reference to the
>> cartesian coordinate system (actually, in the engineering world, these are
>> engineering coordinate systems), where and how is the metadata about the
>> local CRS accessed? How does the local CRS related to the real world CRS
>> (such as WGS 84)?
>>
>> These are the types of questions that the indoor navigation community is
>> dealing with. For example
>> http://portal.opengeospatial.org/files/?artifact_id=35334 is one example.
>>
>> So, expressing the location of IP enabled devices should probably be in
>> the context of the larger in-building information domain. This way,
>> location content provided in an interior extension encoding would "fit" or
>> fuse directly into a larger common operational picture.
>>
>> Sorry for the length of this posting, but this is a really important
>> topic.
>>
>> Regards
>>
>> Carl
>>
>> ----- Original Message -----
>> From: "Brian Rosen" <br@brianrosen.net>
>> To: "Dorothy Stanley" <dstanley1389@gmail.com>; "GEOPRIV"
>> <geopriv@ietf.org>
>> Sent: Monday, October 19, 2009 3:24 PM
>> Subject: Re: [Geopriv] Interior Location Extensions internet draft
>> available
>>
>>
>>> Note that this draft builds on my INT draft
>>> (draft-rosen-geopriv-pidf-interior-00).
>>>
>>> The authors of this draft and I have cooperated on this aspect. One of
>>> the
>>> requests that these authors have made is that we include a registry of
>>> ³N²
>>> names in the INT draft. While I¹m not too thrilled with that idea, I¹m
>>> willing to go along if the rest of the work group agrees it will be
>>> useful.
>>>
>>> I think this draft has a lot of merit, and I'd like to see it progress.
>>> I
>>> have had extended discussions with these guys. There is a significant
>>> use
>>> case, they have the right idea of how to solve it, and it works within
>>> the
>>> kinds of systems it's intended to be used with.
>>>
>>> I think the syntax proposed is very awkward, and I'd like to propose
>>> changes.
>>>
>>> I think the "Reference Point" this draft defines is a bad idea. I think
>>> that the INTs should define the reference point, and the offset should be
>>> relative to that.
>>>
>>> In the examples used, I think it might be expressed as:
>>> ...
>>> <HNO>123</HNO>
>>> <INT N="Suite">100</INT>
>>> <INT N="Door">Front</INT>
>>> And then something like
>>> <POINT_OFFSET>
>>> <X>100</X>
>>> <Y>200</Y>
>>> </POINT_OFFSET>
>>>
>>> At this time, however, I think the syntax issues are not important and
>>> the
>>> basic idea is.
>>>
>>> Brian
>>>
>>>
>>>
>>> On 10/19/09 3:51 PM, "Dorothy Stanley" <dstanley1389@gmail.com> wrote:
>>>
>>>> All,
>>>>
>>>> An internet draft for Interior Location Extensions is now posted, see
>>>> http://www.ietf.org/id/draft-stanley-geopriv-int-ext-00.txt .
>>>>
>>>> This document supports the request for Location by reference
>>>> extensions in 11-09-0718-01-000v-liaison-request-to-ietf-geopriv.doc,
>>>> (available at http://trac.tools.ietf.org/wg/geopriv/trac/wiki ).
>>>> Extensions to existing Civic definitions include reference point, offset
>>>> and
>>>> shape
>>>> components.
>>>>
>>>> I've asked Alissa & Richard for time to present this in Hiroshima.
>>>>
>>>> Thanks,
>>>>
>>>> Dorothy Stanley
>>>>
>>>>
>>>> ------------------------
>>>> Dorothy Stanley
>>>> Aruba Networks
>>>> dstanley@arubanetworks.com <mailto:dstanley@arubanetworks.com>
>>>> 630-363-1389
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Geopriv mailing list
>>>> Geopriv@ietf.org
>>>> https://www.ietf.org/mailman/listinfo/geopriv
>>>
>>>
>>> _______________________________________________
>>> Geopriv mailing list
>>> Geopriv@ietf.org
>>> https://www.ietf.org/mailman/listinfo/geopriv
>>>
>>
>> _______________________________________________
>> Geopriv mailing list
>> Geopriv@ietf.org
>> https://www.ietf.org/mailman/listinfo/geopriv
>>
>
_______________________________________________
Geopriv mailing list
Geopriv@ietf.org
https://www.ietf.org/mailman/listinfo/geopriv
