In principle, the use of a Wi-Fi signal for location and tracking is simple. There's no need for the approach used in RADAR, generating and bouncing a signal off of the object to be located or tracked, because Wi-Fi-equipped devices are usually transmitting data, and with that unique information identifying a given station. Once a given environment is installed with sensors or access points, as required to meet the specifications of a given solution, and any calibration required to tell the infrastructure what to expect when signals are sent from a known location - a form of RF fingerprinting, if you will - you're good to go.
Several different techniques to calculate location are available in practice. The most important of these are:
" Signal strength - This technique examines the RSSI (received signal strength indication) of a given transmission and compares it to calculations of signal strength in a particular location obtained during a calibration process. Signal strength is, however, notoriously variable in wireless systems, and especially indoors. This variability is caused by the vagaries of radio propagation, including the natural exponential fading of signal strength and a wide variety of factors from radio-obstructive and even moving objects in the environment to the echoes and reflections of radio signals, known as multipath. However, with enough samples from a given object, signal strength can provide a very accurate measurement of location, and with rapid time-to-solution for any given object being tracked. These samples can be obtained very quickly, enabling, when coupled with a little artificial intelligence, the possibility of tracking moving objects, even in three dimensions.
" Time Difference of Arrival (TDoA) - Another approach is to use multiple reference sources transmitting what amounts to the value of a clock synchronized with all other transmitters. Assuming the location of transmitters is known with high accuracy and a reliable view (in radio terms) of these transmitters, the receiver can resolve its position, also with high accuracy, and also in three dimensions.
This is exactly, for example, how GPS works - the positions of the GPS NAVSTAR satellites are always known (with small corrective updates for the ephemeral variations in orbit because of simple physics and small atmospheric variations affecting the speed of the signal transmitted), so a receiver can simply measure the time difference in the arrival of each signal and do a relatively simple calculation to determine location with excellent accuracy. Many E911 systems also use this technique, but not via GPS, as GPS cannot usually be reliably received indoors. Instead, signals are transmitted from multiple cellular base stations.
Both of these techniques can be improved by increasing the number of transmitters and the application of other correlation techniques, such as adding a known fixed reference point for a differential calculation. But as resolution of a square meter or so is often all that is required in most RTLS applications, exotic or expensive additions are usually of little value in indoor situations. Hybrid GPS/Wi-Fi-based solutions, enabling the mapping of indoor coordinates to world coordinates, are also possible.
Just as in the case with RADAR, multiple samples and careful calculations are required in any given measurement so as to factor out errors and anomalous readings. In short, it would be wrong to think of Wi-Fi-based location and tracking technologies as an exact science - but they can be, as we learned through this series of exercises, regardless quite accurate and, depending upon the application, also quite valuable indeed.
Read more about anti-malware in Network World's Anti-malware section.