Location-based services are reaching beyond navigation and E911. Social networking, advertising, and other emerging applications are driving research and development into better technologies in support of these services, starting with the location estimation process itself.
Today, there are several ways to determine the location of a wireless device in a network. As a rough estimate, the device knows the identification of the base station or access point with which it is associated. That’s a start, but a base station can cover a wide area. To refine that, devices can use GPS. Networks of cellular base stations can measure differences in the timing or signal strength of an uplink signal, and use the results to estimate location. MAC addresses of Wi-Fi access points can be sniffed and mapped by roving monitors. Sometimes, two or more of these techniques are used in concert.
Google adds another technique to the list as described in a January 28 patent application: mapping packet data rates to distances. Generally, the closer a wireless device is to a base station or access point, the higher the data rate due to the better quality signal. For a particular system and environment, one can create a table mapping different data rates to corresponding distances. The mapped results can then be used to estimate the location of devices relative to each other in the network. If the locations of some devices in the network are known absolutely (maybe a few have GPS), the locations of other devices can be determined absolutely.
As with the other techniques above, the results produced are not exact. The measured data rates can vary not only as a function of distance but also due to other factors including varying propagation conditions, varying transmit power, and interference. Part of the patent covers a method for estimating and assigning an accuracy and confidence level to the results.