Localization refers to the process of determining an object's location in space. Although most often associated with modern technology, more primitive localization methods do exist. In fact, basic localization can be achieved without the use of any special instruments; sailors have been using celestial objects for sea-based localization for a few thousand years. Many specialized tools have been developed to help provide more accurate localization, most notably the astrolabe, sextant, compass, and chronometer, as well as detailed maritime charts and maps . In the 1920s, localization based on radio signals from shore-based transmitters further improved sea-based navigation . In the late 1960s, the U.S. Department of Defense (DoD) began developing a satellite-based localization system for military purposes, which eventually evolved into the Global Positioning System (GPS). The system saw its first use in combat during the Persian Gulf War in 1990 .
As far as back as 1983, however, GPS technology had begun to migrate into the public sector. In response, in 1990 the DoD activated Selective Availability (SA), a purposeful degradation in the civilian GPS signal which limited the accuracy of most civilian GPS units to about 100 meters. In 2000, the DoD de-activated SA in recognition of the increasingly important role played by GPS in numerous commercial activities. The de-activation, along with the development of other technologies such as Differential GPS, now allows civilian GPS units to obtain an accuracy of 10 meters or better. For specialized applications like surveying, technology has been developed allowing accurate measurements at the centimeter level .
For localization in an outdoor environment, GPS works extremely well. Unfortunately, the signal from the GPS satellites is too weak to penetrate most buildings, making GPS useless for indoor localization. Many other schemes have been envisioned for indoor localization, mostly based on machine vision, laser range-finding, or cell-network localization. Ladd et al present a novel technique whereby localization is achieved using the IEEE 802.11b standard, commonly known as wireless Ethernet .
The 802.11b wireless networking standard incorporates a mechanism by which a wireless network card can measure the signal strength of all base stations within its broadcast range . A mobile system can use this information to attempt to determine its distance from these fixed base stations. Based on these distances and the known locations of the base stations, the mobile system can then estimate its own current position. A significant change in the actual position of the mobile system will cause a change in the measured signal strengths and a corresponding change in the estimated position. Therefore, as long as the system remains in range of multiple base stations, it should be able to fairly accurately determine its location. The actual implementation is much more complicated than this basic summary suggests, but the results so far have been extremely promising: Ladd et al have implemented this approach to achieve an accuracy of about one meter.
The most difficult part of wireless Ethernet localization is the conversion from signal strength to distance, since signal strength is "nonlinear with distance... [and] has non-Gaussian noise, resulting from multipath effects and environmental effects, such as building geometry, network traffic, presence of people, and atmospheric conditions" . Adding to the difficulty is the fact that the IEEE 802.11b standard operates in the 2.4-GHz frequency band, meaning "microwave ovens, Bluetooth devices, 2.4-GHz cordless phones, and welding equipment can be sources of interference" . Although many attempts have been made to model the propagation of radio signals in indoor environments, none have been widely accepted. In an effort to get around these difficulties, Ladd et al came up with a slightly different approach. They first broke up the area of interest into cells, and then took signal strength readings in each cell, effectively training the system. A mobile system could then take signal strength readings, compare the measured data to the training set, and use Bayesian inference to determine the location that would most likely produce those measurements.
Ladd et al identify a number of areas for future research . They note that most of their experiments were conducted at night, when there was relatively little human or network traffic. Their experiments were also conducted in corridors, meaning their movement was restricted to relatively narrow straight lines. It would be interesting to study the behavior of the system in an environment that is either more dynamic or more geometrically irregular, or potentially even both. The researchers also suggest the development of techniques for determining the placement of base stations for improved localization accuracy.
The advantages to wireless Ethernet localization are clear. Unlike GPS, the system will work in any location with access to multiple wireless base stations, regardless of whether it is indoors or outdoors. Also, since most systems that would potentially use this technology already use a wireless network card, there is no associated hardware cost. The reliance on the existing wireless infrastructure can be a disadvantage, however, as the system would clearly not work in areas with no wireless base stations.
Many potential applications of wireless Ethernet localization exist. Perhaps the most obvious and of most interest to many researchers is the application to mobile robotics. An individual robot could determine its absolute position inside a building, while multiple robots could determine their relative position in an arbitrary outdoor environment. Since many robots already use wireless Ethernet for communication, using this localization method would often require no extra hardware.
Although most of the current research in localization seems to be focused on its application to robotics, there also exist a number of important applications of wireless Ethernet localization for people. For example, it would also be possible for a person to use a handheld device to determine his or her location inside of a building with an existing wireless infrastructure. The device could then use that localization information to provide contextual information to the user. This would be especially relevant in a location where visitors are given tours, such as a museum or a college campus. As wireless networks become more ubiquitous and the signal strength of base stations increases, this technology could also potentially be adapted for more general purpose localization. Especially in metropolitan areas where the number of wireless access points is rapidly increasing, a handheld device could potentially use the signals from these publicly accessible base stations to determine its location and, as before, provide contextual information such as nearby restaurants or stores. One unique application, mentioned in , is that of determining the location of a malicious user accessing an organization's wireless network.
Potential senior projects in this area include implementing a wireless Ethernet localization system for use at Union and comparing its performance to that of the system discussed in , , . Considering that the other system was created over a number of years by a team of graduate students, this project would probably be infeasible. A more practical option would be to use the previously-developed system and implement some of the future research suggestions discussed above. In particular, analyzing the effect of human and network traffic on the performance of the system would be interesting. It would also be possible to use the existing system to develop an application relying upon localization. The most likely application would be a hand-held tour guide system for the Union campus, or at least for the buildings on the campus with multiple wireless base stations.