I had the honor of leading a plenary debate at this year's Indoor Positioning and Indoor Navigation Conference in Sydney on 14. November. The panel had representatives from Nokia, Google, Locata, Aeroscout (Stanley Healthcare), Smarttrack RFID, and Siemens.
Indoor positioning covers a vast range of applications and technologies such as:
- Tracking and guiding of rescue workers in smoke filled and damaged buildings
- Precise surveying with mm precision
- Navigation of robots in harsh environments such as a nuclear facility
- Ambient Assisted Living: Aid to the elderly so they can live longer in their homes without having to move to an institution
- Logistics in the car manufacturing industry
- Tracking of valuable equipment in hospitals and other institutions
- Localization of distress calls to 911 or 112 even when the caller is indoors
- Precise navigation indoors and outdoors if GPS has been jammed
- Tracking of nurses or patients
- Tracking of prisoners and giving alarms if they happen to be on the wrong side of the wall
- Consumer applications where a smart phone can aid you in finding your way at IKEA, an airport, or a shopping mall
The industry in the panel represented in particular the last 5-6 applications and some of the questions that were discussed were:
- How can one measure and quantify performance? One way is of course in terms of precision in meters, but also latency, or how much power which is consumed in a handheld device is important. It is also quite obvious that in an application such as tracking in a prison you cannot have anything but 100% accuracy in pinpointing the room or floor, while in other applications such errors have fewer consequences.
- Much indoor positioning takes advantage of the already existing wireless network (WLAN) access point infrastructure. But it has been known for some years that this gives an average error of about 3 meters, but also that there is on the order of a 5% chance of getting errors larger than 10 meters. The latter can be a big problem. The panel was very much in agreement that one has hit the wall in terms of performance and that these limits are fundamental. In order to progress one has to combine WLAN with other technologies such as smart phones with barometers (in order to resolve floors) and if possible also a compass, accelerometers and gyroscopes. This is the approach that Google is taking as they develop Indoor Google Map. Alternatively one can combine WLAN with more infrastructure such as Aeroscout does with for instance ultrasound from Sonitor in Oslo.
- 2.5D vs. 3D positioning. It is often not so interesting to know at which height in a room an object or a position is located. On the other hand it may of vital importance to resolve the floor. Therefore indoor positioning is not fully 3D, but rather 2.5 D. But this is still more than the 2 dimensions that one usually finds in most academic papers on the subject.
The conference and the efforts of Google were featured in an interesting article in the Sydney Morning Herald: Indoor GPS': Every step you take, every move you make, Google's got maps for you.
Thanks to participants in the panel and their contributions: Lauri Wirola (Nokia), Waleed Kadous (Google), Nunzio Gambale (Locata), Daniel Aljadeff (Aeroscout), Gary Miller (Smarttrack RFID), and Alejandro Ramirez (Siemens).
This is one of a few posts in English on this blog.