Indoor navigation comprises the collection of 2D mapping data, building information models (BIM) and 3D visualization. Presentations delivered at COM.Geo this week nicely framed these issues.
Carl Smyth, director at MobileGIS Ltd, characterized the situation by stating that small scale spaces are inherently 3D, but for true navigation applications, human cognition is challenged. Applications must consider that the obfuscation of objects is prevalent.
Perfect knowledge models break down because there might be too much detail; too many sources; no consistency; and details are incomplete. In contrast with large spaces, most small space features are "areal" and not "linear." So, how does the paradigm change for indoor navigation? Smyth suggested these principals:
- Pay attention to human cognition.
- Incorporate mobile device sensors.
- Offload "hard" conversion, mining, rendering, image processing, or line of sight remotely to a cloud service.
- Agree to work with partial and contradictory information from multiple sources.
- Build on a self-sustaining web of consumed and created resources.
Not surprisingly, the company that has invested so much in vehicle navigation, NAVTEQ, is looking closely at facilitating navigation to objects as well, or what Paul Bouzide of NAVTEQ calls "highly-context-focused 'around me' use cases." Citing the not always obvious issue that geospatial features and properties are dynamic. He mentioned that in the case of building models, walls and fixtures change, color schemes and décor change and signage can be altered both purposely and otherwise. Any change can matter crucially to the contextual behavior of the application and to maintaining believability and realism. Most changed features or properties exist in relation to others and can depend on prior changes for consistency. NAVETQ also considers adjuncts to road navigation such as parking spaces, road edge, or sidewalks as a component of small space navigation
Geoff Zeiss, Autodesk's director of the Utility Industry Program, discussed how BIM has significant cost savings not just in pre-construction visualization but also during the entire constricution process so that project managers can keep the project on schedule. But Zeiss pointed out that the biggest use is in operating and maintenance management after construction where, as mentioned above, objects properties change and keeping track of these changes is critical to good building maintenance. Zeiss truly feels that BiM may represent an inflection point in engineering design.
Dr. Eyal Ofek, a principal researcher at Microsoft on the Bing Maps team demonstrated how the problem of "data freshness" and currency can be mitigated by using crowd-sourced information like photos. Microsoft Photosynth has been used to illustrate changes to both indoor and outdoor environments.
Michael Loushine, a senior scientist with Telcordia Applied Research presented information about how to "Move E911 Indoors" thereby extending the scenarios we use for emergency response to indoor situations. He said that standards are being used to Improve situational awareness in e911 prototypes such as those from the OpenMobile Alliance (OMA), 3GPP, Internet Engineering Task Force (IETF) Presence Information Description Format Location Object extension (or PIDF-LO) and the OMA Secure User Plane (SUPL) to perform network-based position determination and context. However key to any implementation would be also based on OGC specifications and the use of CityGML would be important.
Summarizing this session that was organized by the OGC was George Percival suggesting that from here, much collaborative development needs to occur to expand the Geoweb to an "Internet of Things." While the session focused more on the platform development rather than the positioning device environment (i.e. RFID tags or other RF sensors), sensor web enablement (SWE) was a central theme to the entire discussion.
See also OGC's resources.