No. There are ways to do indoor positioning indoors: WiFi fingerprinting and supersonic marks.
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.
There's such thing as inertial navigation (that's how missiles fly even with GPS assistance), and potentially you can use a phone's accelerometer for that.
We have a datacenter in downtown SJ and I was amazed when first parked there on the -3 floor underground. My car's in-dash nav works perfectly there :-) There's no even remote chance that GPS can work 3 stories down in a solid concrete building.
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.