"We have built a development platform, defined an XML-based dialect for encoding the representation of complex environments and fielded a series of clients to deliver indoor location-based applications to a wide range of users," said Hoke.

Companies such as NearSpace, YDreams and Ekahau already recognize the importance and potential of such applications and are grounded in the proliferation of mobile devices and emergence of positioning systems able to locate users and objects indoors.

Vocal Surroundings

The applications serve diverse needs, including "way finding" indoors, querying objects and community building (relating people, events and objects based on location), entertainment (location-based games), and education (location-based media). Thanks to expanding use of technologies such as "always on" General Packet Radio Service, Wi-Fi and Bluetooth access points, and Radio Frequency Identification (RFID), new capabilities will help develop LBS applications resulting in "intelligent environments." Instead of being mute, your surroundings will tell you about themselves and what they offer.

Commerce likely will drive the adoption of intelligent environments with "smart shopping." Imagine arriving at a local supermarket, and, immediately after crossing the door, you receive a welcome message on your mobile phone that informs you about the day's promotions.

Automatically, a shopping list and route map is generated based on your refrigerator contents and consumer habits. At checkout, users just confirm the total and accept the electronic payment with an RFID-enabled mobile device acting as a credit or debit card. RFID chips will be important, because they're small, don't need a power supply, can have embedded documents containing location information, and can be read at a distance.

Such intelligent environments could be the "next big thing," with a paradigm shift from "anything, anytime, anywhere" to "right thing, right time, right place." Also, the LBS industry has been hindered largely because of privacy issues and closed cellular networks that restrict developers from creating more meaningful and innovative applications. Now, with embedded technologies like RFID tags, the idea of "smart places" is becoming attractive, because they can make people's lives easier by offering them beneficial LBSs while protecting user privacy.

The Micro-Geography Scale

Indoor LBSs require micro-detailed geo-referencing to satisfy users' growing needs. It's not enough to geo-reference a building if the position of users and other objects inside the building also is relevant. Objects are used as landmarks, and relationships among the objects are crucial for symbolic representation of the whole system.

Traditional latitude/longitude coordinate systems are no longer sufficient to provide effective and seamless interaction. System feedback must contain valuable information to guide users in their interaction adequate to scale, physical and social contexts. As a result, traditional geographic information isn't detailed enough to satisfy user needs in an indoor setting.

A "lingua franca" data format is needed for symbolic/contextual and geometric maps to achieve "seamless" map/content handoff.

According to Hoke, "the NearSpace Workshop contains various tools for importing appropriate data. They can be used to create NearSpace Frames (a combination of raster and vector information) based on other representations; they don't need to be explicitly spatial in nature. By separating the raster and vector data, we can make the visible representation as detailed or as 'symbolic' as required for the purpose at hand."

It's important to realize that symbolic doesn't need to be less accurate/preferred than geometric, because the real issue is whether the underlying data model and services are based on symbolic (adjacency/topology) or geometric reasoning. For example, a developer can use location to "snap" to a room or floor to enable services based on vicinity or containment vs. using geometry and Euclidian distance as the basis for judging what's close or far.

Whether a position is measured in absolute (i.e., x,y,z coordinates) or relative terms (i.e., "in room 101"), it alone almost certainly isn't the ultimate solution. Any robust and scalable location-sensing architecture needs to develop a hybrid data model to characterize location as a heterogeneous mix of sensors. This way, the LBS can understand the environment and provide users with the location of objects or places they're interested in finding.

Current Limitations

Positioning "handoff," which occurs, for example, when users go from outdoors to indoors, affects localization of data content. LBS providers have to figure out how to seamlessly provide location content using different (local) location model types and data formats. Indoor and outdoor positioning techniques need to automatically determine a user's location to provide a "seamless" experience.

One solution is to set up buildings so they overlay with world coordinates. Explained Michael Parkin, MIT Department of Facilities, "We figured out what the correct orientation is for each floor plan, which enabled us to construct a room model based on 'real-world' coordinates."

But such a CAD/GIS representation of space is a major investment. In most cases, a symbolic representation might suffice and be significantly cheaper.

Hybrid Location Data Model

A hybrid location data model would shield the details of underlying positioning sensors and support applications that need or could use symbolic and geometric location information. The symbolic name and geometric coordinates can convert to each other via pre-defined predicates.

In addition to being hybrid, a common (standardized) model may increase interoperability among applications and make new classes of applications possible (due to easier integration). The basic requirement for such an approach is a common language for describing and querying location information.

A modeling language is needed to exchange information that represents indoor settings. Building a new indoor LBS application is relatively easy, and gathering and organizing data isn't difficult. But if data from one application need to be integrated with data from another, a common data modeling language is essential.

So how will the indoor LBS market achieve "critical mass" user levels? Markets will materialize when vendors can build a rich profusion of commercial brands and proprietary products and services on top of an integrated standards infrastructure.

A main component of success for the aforementioned applications relies on real-time location or sensing. Spatial awareness has been mainly explored in outdoor environments, supported by an established GIS industry. In contrast, the field of indoor LBSs (or micro-GIS) is completely new.