Lightning Laboratories | Strategic Innovation & Consulting for Startups and Enterprise

Open AR: what’s the point?

Tuesday, September 8th, 2009

Like many other folks involved in augmented reality, I’d like to see the mobile AR community embrace open standards for AR experiences. And just to be clear, by “embrace” I mean “create and implement”. Now, I know this discussion is eventually going to take us into deep waters, but let’s just start off with the simplest possible thing. I’d like to see the mobile AR community agree on how it represents a point in space. If we could do that, we might be able to create some simple, public AR experiences that work across platforms and in the various competing AR browsers. And the positive example of one agreed open standard, arrived at by an open community process, might lead to additional good things. So let’s talk about points.

Geographic AR Points

Geographic AR systems like Layar, Geovector, Wikitude, Robotvision, Gamaray etc, use a spheroid-based coordinate system of latitude, longitude and (sometimes) altitude to specify the point locations of the observer and georeferenced content. POIs (points of interest) consisting of a single (lat,lon,alt) coordinate tuple plus various metadata, are commonly used to represent physical entities such as restaurants, monuments and attractions. Unfortunately even in this extremely simple case, there is no agreement on specifications for a single point in space. For example, if altitude is used, is it the height of the point above the topographic surface at that location, the height above the observer’s location, or the height above the WGS-84 reference ellipsoid approximating mean sea level, as a GPS would measure it? Does a point also have accuracy metrics? And what metadata are required or optional for each point? Each of the companies mentioned above is doing something a bit different, and so are their upstream POI data providers. So far, and despite recent announcements, openness is not really happening yet.

3D AR Points

AR has its roots in computer graphics & vision technologies, and these approaches primarily use 3D cartesian (xyz) coordinate systems. A 3D model of a teapot might have a local xyz coordinate system; the teapot rests on a 3D model of a table which in turn has its own reference coordinate system; the observer of the scene has their own reference coordinate system; the screen that the scene is displayed on has its own 2D pixel coordinates, and a set of mathematical transformations (e.g., translation, scaling, rotation & projection) ties them all together. A 3D graphics scene is not inherently tied to any physical world reference point; in marker-based AR, the fiducial marker provides an anchor that binds the 3D augmented scene to a physical world location. However, the data structure for the scene’s location is entirely relative, which makes the location of 3D models fairly portable.

Simple Geo + 3D AR

Of course, one simple and obvious thing we want is to enable 3D graphics models to be placed in geographic locations. If we truly think open AR is important, we are going to want to agree on which kinds of coordinate systems to use. This is not a trivial question. Do we want the 3D model to be on a local or global coordinate system? A fixed position relative to the world and regardless of viewpoint, or always located relative to the observer? What if the model and the observer are on boats? What if the model is something like an entire city? Different choices for coordinate systems and schema will impact computational costs and accuracy. In Google Earth, KML allows use of static COLLADA models which are then imported/transformed to the GE geographic coodinate system. Planet9’s virtual cities have a single reference coordinate system for the entire city, and use UTM WGS-84 in order to keep their building models square. The Web3D Consortium’s X3D framework supports georeferencing models in geodetic, UTM and geocentric reference frames, appropriate for a variety of use cases. What approach(es) makes sense for mobile AR? Can we leverage & extend existing standards, or will we have to create new ones from the ground up?

Start simple, but start now

Okay, so clearly things can get messy, even for the simple case of specifying a point in space. And it is also clear that multiple constituencies are going to be very interested in the geographic and 3D graphic aspects of AR. I think it’s time to have serious discussions about open standards for mobile AR, starting with the basic question of representing POIs and static 3D objects. I realize it is hard for small, fast moving teams to spend precious energy on this kind of discussion, but to me it seems like a critical thing for the community to establish a common foundation for the mobile AR experience. Do you agree? If not, why not? If so, then where should this discussion happen and who should be involved? Perhaps the recently formed AR Consortium can play a role here? Maybe it is already happening somewhere?

I’m very interested in your thoughts on this topic. Please share in the comments below, link here from your own blog, or respond @genebecker. YMMV as always.

For further reading

* Augmented Reality Should Be Open by Joe Ludwig
* Augmented Reality: Open, Closed, Walled or What? by Robert Rice
* Wikitude API
* Layar API
* Gamaray formats
* Garmin GPX POI schema
* WGS-84
* UTM
* A Discussion of Various Measures of Altitude
* GeoRSS
* GeoJSON
* W3C Geolocation API
* KML
* COLLADA
* X3D
* CityGML
* OGC GML

thinking about design strategies for ‘magic lens’ AR

Tuesday, September 1st, 2009

I love that we are on the cusp of a connected, augmented world, but I think the current crop of magic lenses are likely to overpromise and underdeliver. Here are some initial, rough thoughts on designing magic lens experiences for mobile augmented reality.

The magic lens

The magic lens metaphor [1] for mobile augmented reality overlays graphics on a live video display from the device’s camera, so that it appears you are looking through a transparent window to the world beyond. This idea was visualized to great effect in Mac Funamizu’s design studies on the future of Internet search from 2008. Many of the emerging mobile AR applications for Android and the iPhone 3GS, including Wikitude, Layar, Metro Paris, robotvision, Gamaray and Yelp’s Monocle, are magic lens apps which use the device’s integrated GPS and digital compass to provide location and orientation references (camera pose, more or less) for the overlay graphics.

The idea of a magic lens is visually intuitive and emotionally evocative, and there is understandable excitement surrounding the rollout of commercial AR applications. These apps are really cool looking, and they invoke familiar visual tropes from video games, sci-fi movies, and comics. We know what Terminator vision is, we’re experienced with flight sim HUDs, and we know how a speech balloon works. These are common, everyday forms of magical design fiction that we take for granted in popular culture.

And that’s going to be the biggest challenge for this kind of mobile augmented reality; we already know what a magic lens does, and our expectations are set impossibly high.

Less-than-magical capabilities

Compared to our expectations of magic lenses, today’s GPS+compass implementations of mobile AR have some significant limitations:

* Inaccuracy of position, direction, elevation – The inaccuracy of today’s GPS and compass devices in real world settings, combined with positional errors in geo-annotated data, mean that there will generally be poor correspondence between augmented graphical features and physical features. This will be most evident indoors, under trees, and in urban settings where location signals are imprecise or unavailable. Another consequence of location and orientation errors is that immediately nearby geo-annotations are likely to be badly misplaced. With typical errors of 3-30 meters, the augments for the shop you are standing right in front of are likely to appear behind you or across the street.

* Line of sight – Since we can’t see through walls and objects, and these AR systems don’t have a way to determine our line of sight, augmented features will often be overlaid on nearby obstructions instead of on the desired targets. For example, right now I’m looking at Yelp restaurant reviews floating in space over my bookshelf.

* Lat/long is not how we experience the world – By definition, GPS+compass AR presents you with geo-annotated data, information tied to geographic coordinates. People don’t see the world in coordinate systems, though, so AR systems need to correlate coordinate systems to world semantics. The quality of our AR experience will depend on how well that translation is done, and today it is not done well at all. Points Of Interest (POIs) only provide the barest minimum of semantic knowledge about any given point in space.

* Simplistic, non-standard data formats – POIs, the geo-annotated data that many of these apps display, are mostly very simple one-dimensional points of lat/long coordinates, plus a few bytes of metadata. Despite their simplicity there has been no real standardization of POI formats; so far, data providers and AR app developers are only giving lip service to open interoperability. Furthermore, they are not looking ahead to future capabilities that will require more sophisticated data representations. At the same time, there is a large community of GIS, mapping and Geoweb experts who have defined open formats such as GeoRSS, GeoJSON and KML that may be suitable for mobile AR use and standardization. I’ll have more to say about AR and the Geoweb in a future post. For now, I’ll just say that today’s mobile AR systems are starting to look like walled gardens and monocultures.

* Public gesture & social ambiguity – Holding your device in front of you at eye level and staring at it gives many of the same social cues as taking a photograph. It feels like a public gesture, and people in your line of sight are likely to be unsure of your intent. Contrast this with the head down, cradled position most people adopt when using their phone in a private way for email, games and browsing the web.

* Ergonomics - Holding your phone out in front of you at eye level is not a relaxed body position for extended viewing periods; nor is it a particularly good position for walking.

* Small screen visual clutter – If augmented features are densely populated in an area, they will be densely packed on the screen. A phone display with more than about 10 simultaneous augments will likely be difficult to parse. Some of Layar’s layer developers propose showing dozens of features at a time.

Design strategies for practical magic

Given these limitations, many of the initial wave of mobile AR applications are probably not going to see great adoption. The most successful apps will deliver experiences that take advantage of the natural technology affordances and don’t overreach the inherent limitations. Some design strategies to consider:

* Use augments with low requirements for precision and realism. A virtual scavenger hunt for imaginary monsters doesn’t need to be tied to the exact geometry of the city. A graphic overlay showing air pollution levels from a network of sensors can tolerate some imprecision. Audio augmentation can be very approximate and still deliver nicely immersive experiences. Searching for a nearby restroom may not need any augments at all.

* Design for context. The context of use matters tremendously. Augmenting a city experience is potentially very different from creating an experience in an open, flat landscape. Day is a different context than night. Alone is different than with a group. Directed search and wayshowing is different from open-ended flaneurism. Consider the design parameters and differences for a user who is sitting, standing, walking, running, cycling, driving and flying. It seems trivially obvious, but nonetheless important to ask who is the user, what is their situation, and what are they hoping will happen when they open up your app?

* Fail gracefully and transparently. When the accuracy of your GPS signal goes to hell, reduce the locative fidelity of your app, or ask the player to move where there is a clear view of the sky. When you are very close to a POI, drop the directional aspect of your app and just say that you are close.

* Use magic lens moments sparingly. Don’t make your player constantly chase the virtual monsters with the viewfinder, give her a head-down tricorder-style interaction mode too, and make it intuitive to switch modes. If you’re offering local search, consider returning the results in a text list or on a map. Reserve the visual candy for those interactions that truly add value and enhance the sense of magical experience.

* Take ownership for the quality of your AR experiences. Push your data providers to adopt open standards and richer formats. Beat the drum for improved accuracy of devices and geo-annotations. Do lots of user studies and experiments. Create design guidelines based on what works well, and what fails. Discourage shovelwARe. Find the application genres that work best, and focus on delivering great, industry-defining experiences.

We are at an early, formative stage of what will eventually become a connected, digitally enspirited world, and we are all learners when it comes to designing augmented experiences. Please share your thoughts in the comments below, or via @genebecker. YMMV as always.

[1] The idea of a metaphorical magic lens interface for computing was formulated at Xerox PARC in the early 1990’s; see Bier et al, “Toolglass and Magic Lenses: The See-Through Interface” from SIGGRAPH 1993. There is also a substantial body of previous work in mobile AR including many research explorations of the concept.

augmenting Nokia’s reality: quick thoughts on Point & Find

Tuesday, August 25th, 2009

From the “Who asked you anyway?” department…

I hadn’t really looked at Nokia’s Point & Find concept before, even though it has been shown around at least since Ubicomp 2007 and the public beta has been running since May09. After last night’s Nokia AR talk I thought I would have a peek.

point-find

OK, it’s a beta with limited functionality, and undoubtedly there are plenty of enhancements to the system and the business plan going on behind closed doors, so take this with a large scoop of salt because I only know what’s been said in public. But if I was advising Nokia on Point & Find, I would start with this:

1. The business opportunity is in the service for building and managing the image database, and managing all of the associated meta-behaviors like hyperlinking, bookmarking, purchasing, search, customer intent discovery, collaborative recommendations and so on. By building this physical-to-digital gateway between individual users and companies and institutions, Nokia can accelerate the development of the connected world platform and profit nicely from it as well.

2. The business opportunity for this is NOT in the handset. That should actually be great news for Nokia, because their core hardware business is highly commoditized and that’s a cold reality their employees have to live every day. In the long run, selling commodity hardware is not going to be the best strategy for Nokia (yes, that includes stuff like windows-based netbooks).

3. Point & Find should go on the iPhone and Android platforms, as soon as possible. Because the opportunity is in growing profitable services, not in fighting the low margin handset wars. Because the current AR hype is a swell marketing tool, and right now mobile AR is all about iPhones and gPhones. Because if Nokia doesn’t own image-based physical hyperlinking on those phones, someone else will, and that means the strategic momentum will shift. I realize this is totally countercultural, maybe even unthinkable, but that’s why it would be such an interesting play.

4. Nokia should open up an API to its backend service, so that any developer can make services that tap into that (hypothetical?) database of millions of images of products, printed ads, books, artworks and on and on. Developer momentum advances the entire market, and solidifes the platform’s competitive position. This can be one of those big web squared opportunities, if played right.

How about you, have you tried Point & Find? If you were Nokia, what would you do?

a brief response re: web squared

Friday, July 10th, 2009

Tim O’Reilly and John Battelle recently proposed the term “Web Squared” to describe the next phase of the web, where “web meets world” in a melange of collective intelligence, data utilities, pervasive sensing, real time feedback, visualization, emergent semantic structure, and information infusing the physical world. For what it’s worth, I quite like it. We needed a new handle for the remarkable confluence of technologies we are experiencing, and I think Web Squared nicely captures the exponential expansion of possibilities while reaffirming that the web is the only plausible distributed systems infrastructure to build the new world on.

I was also intrigued by the authors’ conclusion, which moves the discussion beyond the realm of technology and into “the stuff that matters”:

All of this is in many ways a preamble to what may be the most important part of the Web Squared opportunity. The new direction for the Web, its collision course with the physical world, opens enormous new possibilities for business, and enormous new possibilities to make a difference on the world’s most pressing problems.

As a techno-optimist by nature, I’m pretty susceptible to visions of enormous new possibilities. I’ve even generated a few of those lovely consensual hallucinations myself, and they can be very exciting to be in the middle of. And it’s almost certainly true – the potential implications are huge. However, I think we also need to examine this vision more critically as part of the ongoing discussion, for example giving serious attention to Adam Greenfield’s design principles for Everyware, and to John Thackara’s concerns when he writes:

Connected environments…and the Internet of Things as a whole, are not a step forwards if they guzzle matter and energy as profligately as the internet of emails does

and echoes Patricia de Martelaere’s caution against

“wasting our lives by continuously watching images of world-processes, or processes of our own body, and desperately trying to interfere – like a man chasing his own shadow.”

After all, in the era of Web Squared we are not just creating new business opportunities; we are talking about cyberspace seeping out of the very fabric of reality. I’m thinking that we don’t want to screw that up.

what is ubiquitous media?

Friday, June 26th, 2009

In the 2003 short paper “Creating and Experiencing Ubimedia“, members of my research group sketched a new conceptual model for interconnected media experiences in a ubiquitous computing environment. At the time, we observed that media was evolving from single content objects in a single format (e.g., a movie or a book), to collections of related content objects across several formats. This was exemplified by media properties like Pokemon and Star Wars, which manifested as coherent fictional universes of character and story across TV, movies, books, games, physical action figures, clothing and toys, and American Idol which harnessed large-scale participatory engagement across TV, phones/text, live concerts and the web. Along the same lines, social scientist Mimi Ito wrote about her study of Japanese media mix culture in “Technologies of the Childhood Imagination: Yugioh, Media Mixes, and Otaku” in 2004, and Henry Jenkins published his notable Convergence Culture in 2006. We know this phenomenon today as cross-media, transmedia, or any of dozens of related terms.

Coming from a ubicomp perspective, our view was that the implicit semantic linkages between media objects would also become explicit connections, through digital and physical hyperlinking. Any single media object would become a connected facet of a larger interlinked media structure that spanned the physical and digital worlds. Further, the creation and experience of these ubimedia structures would take place in the context of a ubiquitous computing technology platform combining fixed, mobile, embedded and cloud computing with a wide range of physical sensing and actuating technologies. So this is the sense in which I use the term ubiquitous media; it is hypermedia that is made for and experienced on a ubicomp platform in the blended physical/digital world.

Of course the definitions of ubicomp and transmedia are already quite fuzzy, and the boundaries are constantly expanding as more research and creative development occur. A few examples of ubiquitous media might help demonstrate the range of possibilities:

An interesting commercial application is the Nike+ running system, jointly developed between Nike and Apple. A small wireless pressure sensor installed in a running shoe sends footfall data to the runner’s iPod, which also plays music selected for the workout. The data from the run is later uploaded to an online service for analysis and display. The online service includes social components, game mechanics, and the ability to mashup running data with maps. Nike-sponsored professional athletes endorse Nike-branded music playlists on Apple’s iTunes store. A recent feature extends Nike+ connectivity to specially-designed exercise machines in selected gyms. Nike+ is a simple but elegant example of embodied ubicomp-based media that integrates sensing, networking, mobility, embedded computing, cloud services, and digital representations of people, places and things. Nike+ creates new kinds of experiences for runners, and gives Nike new ways to extend their value proposition, expand their brand footprint, and build customer loyalty. Nike+ has been around since 2006, but with the recent buzz about personal sensing and quantified selves it is receiving renewed attention including a solid article in the latest Wired.

A good pre-commercial example is HP Labs’ mscape system for creating and playing a media type called mediascapes. These are interactive experiences that overlay audio, visual and embodied media interactions onto a physical landscape. Elements of the experience are triggered by player actions and sensor readings, especially location-based sensing via GPS. In the current generation, mscape includes authoring tools for creating mediascapes on a standard PC, player software for running the pieces on mobile devices, and a community website for sharing user-created mediascapes. Hundreds of artists and authors are actively using mscape, creating a wide variety of experiences including treasure hunts, biofeedback games, walking tours of cities, historical sites and national parks, educational tools, and artistic pieces. Mscape enables individuals and teams to produce sophisticated, expressive media experiences, and its open innovation model gives HP access to a vibrant and engaged creative community beyond the walls of the laboratory.

These two examples demonstrate an essential point about ubiquitous media: in a ubicomp world, anything – a shoe, a city, your own body – can become a touchpoint for engaging people with media. The potential for new experiences is quite literally everywhere. At the same time, the production of ubiquitous media pushes us out of our comfort zones – asking us to embrace new technologies, new collaborators, new ways of engaging with our customers and our publics, new business ecologies, and new skill sets. It seems there’s a lot to do, so let’s get to it.

a few remarks about augmented reality and layar

Wednesday, June 24th, 2009

I genuinely enjoyed the demo videos from last week’s launch of the Layar AR browser platform. The team has made a nice looking app with some interesting features, and I’m excited about the prospects of an iPhone 3GS version and of course some local Silicon Valley layarage.

At a technical level, I was reminded of my Cooltown colleagues’ Websign project, which had the very similar core functionality of a mobile device with integrated GPS and magnetometer, plus a set of web services and a markup language for binding web resources (URLs) to locations with control parameters (see also: Websigns: Hyperlinking Physical Locations to the Web in IEEE Computer, August 2001). It was a sweet prototype system, but it never made it out of the lab because there was no practical device with a digital compass until the G1 arrived. Now that we have location and direction support in production platforms, I’m pretty sure this concept will take off. Watch out for the patents in this area though, I think there was closely related prior art that even predated our work.

Anyway I looked carefully at all the demos from Layar and the various online coverage, and wondered about a few things:

Layar’s graphical overlay of points of interest appears to be derived entirely from the user’s location and the direction the phone is pointed. There is no attempt to do real-time registration of the AR graphics with objects in the camera image, which is the kind of AR that currently requires markers or a super-duper 3D point cloud like Earthmine. That’s fine for many applications, and it is definitely an advantage for hyperlinks bound to locations that are out of the user’s line of sight (behind a nearby building, for example). Given this, I don’t understand why Layar uses the camera at all. The interaction model seems wrong; rather than using Layar as a viewfinder held vertically in my line of sight, I want to use it like a compass — horizontally like a map, and the phone pointed axially toward my direction of interest. This is most obvious in the Engadget video, where they are sitting in a room and the links from across town are overlaid on images of the bookshelves ;-) Also, it seems a bit unwieldy and socially awkward to be walking down the street holding the phone in front of you. Just my $0.02 there.
How will Layar handle the navigation problem of large numbers of active items? The concept of separate “layars” obviously helps, but in a densely augmented location you might have hundreds or even thousands of different layers. Yes this is a hard UI/UX problem, but I guess it’s a problem we would love to have, too much geowebby goodness to sort through. I suppose it will require some nicely intuitive search/filtering capability in the browser, maybe with hints from your personal history and intent profile.
Will Layar enable participatory geoweb media creation? I’d be surprised if they don’t plan to do this, and I hope it comes quickly. There will be plenty of official corporate and institutional voices in the geoweb, but a vibrant and creative ecosystem will only emerge from public participation in the commons. This will demand another layer of media literacy, and this will take time and experimentation to develop. I say the sooner we get started, the better.

In any case, good luck to the Layar team!

a remarkable confluence of technologies

Tuesday, June 23rd, 2009

I have used versions of this picture in many of my talks over the last 10 years, and it just keeps getting more interesting. The overarching message is that we are in the midst of a remarkable wave of innovation, with major advances coming to fruition across many different technology domains, at more or less the same time. This is creating fertile conditions for all manner of new products, services and experiences to emerge through what economist Hal Varian calls “combinatorial innovation” (a nicely refined phrase for an incredibly messy process). Another way to put it is, this is the ubiquitous computing + digital media + internet + physical world supercollider, and we are starting to see the results of a very big, long-running experiment.

It has become somewhat of a cliche to promise radical transformation of businesses through technology, but just because it’s a cliche doesn’t mean it’s wrong. Over the next several years, much of the physical world will become connected into an Internet of people, places & things. This will fundamentally change the nature of our experiences with products, services, environments, organizations, and each other. There is no industry or institution that will be untouched; in retrospect we will see that traditional media companies were simply the low-hanging fruit. Just as the Nike+ running system transforms a shoe into a networked sensor node and an apparel company into the host of a worldwide social media community, combinatorial innovation will plant seeds of opportunity and disruption widely.

(Click image for ridiculously large version, cc-by-sa)

Company

Services

Project & Program Management

Innovation Workshops

Custom Concept Investigations

Organization Design for Innovation

Resources