Google Maps Floor Plans
maps.google.com
maps.google.com
My question is at 28:30 http://audio.sxsw.com/2011/podcasts/GoogleMarrisaMayer.mp3
One of my fellow panelists from Google discussed ways for developers to use Google's existing map tools to create their own indoor map content. This uses the Google Maps slippy map API, but doesn't require the developer to submit indoor map data to Google: http://www.google.com/events/io/2011/sessions/map-your-busin...
I would also like to see indoor map data, especially for public spaces, be more freely (libre) available. There are a few contributors at OpenStreetMap working on indoor maps, but privacy/data structure/data acquisition pose significant challenges. I'm sure they'd be interested in your help!
[0] https://plus.google.com/u/0/111626127367496192147/posts
[1] http://www.youtube.com/watch?v=9c6W4CCU9M4&feature=youtu...
http://googleblog.blogspot.com/2011/11/new-frontier-for-goog...
I want indoor maps because they can help reduce healthcare costs. It costs us all money when people get lost in hospitals. While environmental signage helps, it's often not sufficient. Visitors end up interrupting doctors and nurses to ask for directions. This directly costs the hospital money (leads to higher personnel costs), and leads to a poor customer experience (affects the recovery process for patients and families). It is surprising how many people get lost in hospitals: http://twitter.com/lostinabuilding
In addition to navigation, there are uses for facilities management, inventory management, real estate sales, and analytics on how people use buildings. I'm a big fan of using data to improve human efficiency, and these can help save time/energy/money.
For retail environments, indoor maps present targeted advertising opportunities, which can lead to more efficient advertising spending for brands and retailers, and time savings for consumers. This is part of why Google is so interested in indoor maps.
Really? Like how much money? Seems like it needs a huge "citation needed" sign.
If Google provided bulk downloads I'd support their service, though, even if the online API required payment, because I do think it's reasonable to charge people who're using Google's infrastructure as basically their webapp backend.
/sarcasm (you never known)
If I'm contributing such stuff, I'll go with open street maps.
But there's also the user generated content argument, where there are plenty of for profit companies that completely depend on content given for free by their users (youtube, reddit, etc), where the users also benefit from having better content on a site they visit and more attention to the content that they post. From that way of looking at things, it makes perfect sense for property owners to want to have their floor plans up on google maps.
In any case, I disagree that paying to access the maps API is necessarily bad...any OSM host will also charge over any trivial amount of traffic because none of them have the (free) infrastructure to handle it. I do agree with other posters, though, that there should be Data Liberation Front coverage for things like this and data contributed through map maker, and that it should have been that way for years now.
The suggestion that people just upload them is somewhat problematic in so far as it is akin to Napster's approach to music a decade ago.
Standard AIA contract language allows the owner to utilize the plans to construct their project. The author (architect) retains copyright.
I guess it's slightly more hairy because architecture is considered to be artistic as well as functional, whereas street layouts are generally only considered to be functional, but I still think that would stand.
On the other hand, architectural drawings(aka "blueprints") are automatically protected by copyright as soon as they are published.
The advanced version of this game would use Google Street View for first-person racing!
Ceiling heights can likely be extrapolated simply from photographs with the right programming. Find a photo with 5 people in it and approximate the building height from the peoples photographs.
I personally can approximate the height and width of a house simply by counting bricks and shingles, because I work in construction I need to know if my scaffolding poles or ladders will reach and you're lucky if you get a tape measure to extend past 7ft without it buckling. (a sheet of shingles is 2ft wide, so every 3 shingle tabs is 2 ft. A brick and mortar roughly make 4", so 3 courses of brick make roughly 1ft. vinyl siding is normally double 4.5", double 5" or double 6" per sheet in height, one sheet is 12' long. Concrete board is generally 3ft by 8" high. Aluminum is always 8" high, generally 12' or 12.5' wide; If I can do this without a pencil and paper inside 30 seconds, it shouldn't be difficult to make an algorithm to work this out.)
Extrapolating the interior of a building would likely be much harder, however pictures of these often come with scale references (IE people), but as with outside everything has to meet building codes and everything is made out of standardized parts. However, one great clue is handrails. First floor railings have to be a minimum of 36" high, second floor railings have to be a minimum of 42" high (standard is 42" by code for commercial spaces, regardless of height). Maximum 4" gap between vertical members, (generally made 4" on centre as a precaution), and no horizontal members between 4" in height and 35" in height as to make a railing climbable by a small child.
Turning a rendering of an entire mall level in a game into simply picking colour swatches would be huge. Yes a pro studio is going to invest more time in polish, but for the guy making an iPhone app and is never going to see a huge return on the time he invests. Then accepting a bit of 'cookie cutter' genericness in level design won't be a huge compromise if it's in a great game.
The national building codes of countries would give you all the information you need to extrapolate the size of rooms and buildings.
There is no way to get feedback about the process and no indication on where in the queue our submissions are.
There are probably 40 buildings on our campus that I was going to submit floor plans for. However, I was using this first building as a guinea pig of sorts. I can't imagine spending the time to get the other 39, multi-floor, buildings aligned in their system if there is little likelihood they will ever end up in the system for public consumption.
Your problem is something we are very interested in. If you are interested in solving this, please let me know by email at thomasdn@google.com.
Thanks again. Tom
Compare: http://g.co/maps/8e9xp
Is there some poor UI incantation you need to do to get this to appear on google maps?
Where that "level of detail" line is drawn is an important discussion as the availability of indoor maps from Google, Bing, OSM, Micello, etc. continues to expand.
But I'm in agreement, their UI to set the GCPs is terrible. :-P
Sort of. At the level of indoor floor plans, affine transformations are just fine.
I guess, then, that begs the question of why the Google floorplan submission UI uses that weird double-pane thing?
[1] http://blogs.esri.com/esri/arcgis/2011/03/24/tissot-s-indica...
If it's just for Android I'm not submitting my building.
Another likely factor for Google deploying on Android only is the ability to use wifi signal strength data from the phone to do approximate indoor positioning (the you-are-here "little blue dot"). Apple chose to lock developers out of the API for reading wifi signal strength, which limits the iPhone's ability to do indoor positioning (or show you nearby wifi routers).
In Apple's defense, the indoor positioning experience on Android does not meet the high expectations that GPS has set for users. Instead of a precise little blue dot, you get a 10-30m wide big blue circle. Some of the positioning hardware announced from chipset makers like CSR and Broadcom may improve this situation in the next generation of handsets.
If Google is able to bring it all together, this could be pretty epic!
It's a pretty interesting development for me to watch as the app I'm currently working on named Mashupforge (http://mashupforge.com) lets people easily create interactive maps out of image floorplans.
On the other hand, in the US, using the "blueprints" is probably going to be a violation of the copyright of the design professional who prepared them.
Typical contract language restricts the Owner's use to "this project" and such rights terminate "upon completion of this project." [AIA Document B105-2007]
Upon execution of this Agreement, the Architect grants to the Developer-Builder a nonexclusive license to use the Architect’s instruments of service solely and exclusively for purposes of constructing, using, maintaining, altering and adding to the Project or the Development, provided that the Developer-Builder substantially performs its obligations, including prompt payment of all sums when due, under this Agreement [AIA Document B107 2010].
When it comes to Google, this isn't civil disobedience.
It's simply taking advantage of the fact that the copyrights it is interested in ignoring are mostly held by many different people unlikely to have the resources to enforce them vigorously.
BTW, it's Ice Cream Sandwiches Google wants everyone to carry.
Or Roosevelt Field Mall, in NY
But I'm a little mixed on this. Google maps, while great and useful, isn't an open source project. The data is closed and proprietary and not useful to the community at large. I'd much rather this be done under the umbrella of OSM or the like, or if nothing else published as a separate data set subject to open licensing.
A floor plan goes against the self interest of the store.
They are a startup from my town (Rotterdam) and have been doing online floorplans longer then anyone else.
http://www.pointinside.com/blog/2011/12/what-google-indoor-m...
Triangulation (trilateration) the way it's used in GPS will not work indoors because it's impossible to build accurate signal propagation model – model would be extremely complicated.
On the other hand signal propagation model can be learned via sampling actual signals in different spots on site. There are two ways to sample signals: 1) direct 2) indirect (via statistical models like SLAM)
Positioning can be done by comparing signal strengths from model and actual device and feeding them to some statistical positioning algorithm like particle filter or LSE. These algorithms use not only signal strengths but accelerometer, gyro and magnetic sensors for movement model. Fingerprinting techniques can provide up to 30 cm accuracy.
Huge win of WiFi fingerprinting is that it works on existing consumer-level hardware. But the downside is that you cannot scan WiFi in iOS public API.
There is less accurate way to position user indoors using supersonic marks – and it works on any existing phone.
Companies who do indoor positioning: WiFiSlam, Qubulus, WalkBase, Google, Nokia and we (applying to yc :). Some provide APIs, some are in stealth mode and Google has indoor positioning in some US and Japanese malls and airports.
Personally I think that there is huge opportunity for location-aware apps for malls, airports, parking lots, hospitals, etc and it has not been tackled much by anyone yet.
However, if crowdsourcing is allowed this could become major. I and 20 other people could all map out the Walmart (for example), upload the data, and then whatever matches in 80% of the submissions would become part of the "map".
If enough people did that it would almost force the major chains to participate, if for no other reason than to maintain control of their maps.
Honestly, I've never seen stores arranged for impulse buying except to have certain items promoted above others. Eye-level shelves, displays at the front of aisles, etc. Otherwise, the fruits and vegetables are together, the canned items are together, the milk and cheese are together, etc. Pretty much every grocery store I've ever been in has been laid out in approximately the same way.
So I'm worried that you may be propagating an urban legend.
Thanks for the interesting perspective. I've actually opened, run, and sold a retail store before (mobile phones and accessories) so, while the type of layout logic I referred to in my comment above did not apply to my store (because it wasn't large enough) I did read ALOT on the subject and was part of the local Chamber of Commerce so I can assure you that impulse buying considerations are a big part of why stores get laid out the way they do.
There are even consulting firms that make a tidy income by advising stores on product placements within the store for maximum sales.
On the other hand, why even bother going to your store when groceries can be ordered online and delivered? Then you don't need a map at all.
On the other hand, milk is invariably at the back of the store to require the shopper to pass areas containing impulse buy items.