End-user Localization Platforms

Posted: March 9, 2011 in Mapping, WiFi

2.1 End-user Localization Platforms

Let us start with two previous works that have resulted in end-user applications. The technical details of both of them are not published.

2.1.1 Examples of End-user Localization Platforms

Google’s My Location My Location is a GPS-like location service that works without a GPS receiver. My Location is published as a feature within Google Maps for Mobile 2.0 (GMM) in the late 2007. The ”my location” feature exploits the GPS location of the mobile device, if it is available. If the GPS location is not available, the software determines the nearest wireless networks and cell sites. The software then  nds the location of the cell site from a database of known wireless networks and cell sites. This localization method is working by triangulating the different signal strengths from different cell transmitters and with the knowledge of their locations, which of course are retrieved from the online database, this helps My Location to determine the user’s current location. Similar approach is applied in wireless network location method . It discovers the nearby WiFi hotspots and using their location property, which is as well retrieved from the online WiFi database, to help discovering the user’s location. The order in which these services take precedence is[4]:

• GPS-based services

• WLAN-based / WiFi-based services

• Cell transmitter-based services

7Skyhook Wireless Technology The other end-user product that proposed non-GPS localization scheme is Skyhook Wireless. Skyhook Wireless technology developed by Skyhook Wireless company. It helps to determine the geographical location using Wi-Fi as the underlying reference system. It mainly works similarly to Google’s My Location wireless network location method. It uses the MAC addresses of audible wireless access points and proprietary algorithms to determine the position of a mobile device within 20-30 meters according to their reports.

2.1.2 Discussion

The main difference between Google’s My Location and Skyhook Wireless Technology is how they build their databases. Google makes use of their widely deployed Google Maps application. The application collects the surrounding access point and cell sites identification information and associates them with the GPS location if available, and then sends all such information to Google servers. Of course, millions of records are submitted daily to Google servers as a result of running Google Maps application on mobile all over the world. Google runs complicated algorithms on the returned data to aggregate them together and builds the database that correlates each access point or cell site to its location which would be returned later on when the location of a non-GPS device is requested through My Location service.

Skyhook Wireless uses a different way of building the data base holding the location information of Wifi access points identified by their MAC addresses. It uses a methodology called wardriving. Wardriving is the act of searching for Wi-Fi wireless networks by a person in a moving vehicle, using a portable computer or PDA. For example, you install Skyhook’s application on your GPS-enabled laptop, you get into your car, run the application and start driving. While the application is running it collects the audible access points as well as the current location of the car from the GPS device. All those information are sent back to Skyhook severs to run their proprietary algorithms and build their own database. They also made a technology named XPS, Figure 2.2: a hybrid positioning engine that integrates the Wi-Fi Positioning System from Skyhook Wireless with other complementary location systems. . XPS is configured to integrate/synthesize the location output of the Wi-Fi Positioning System and other positioning systems. It uses several different techniques to determine which source system is most reliable at any given moment. In some cases, XPS estimates the error of each of its source systems and favors the one with the highest degree of confidence at that particular moment. In other cases, the usage scenario or the confidence history of each source are leveraged. In every case, XPS uses this data to decide which source to favor or how to combine the various inputs into a superior overall calculation.

While the previous two techniques seem really promising, especially they ended up into real application for the end-user and are not just couple of research papers, they have very critical drawbacks. The  first drawback of those techniques is the coverage. While GPS provides 100% coverage by definition, My Location’s and Skyhook’s coverage totally depends on the information in their own databases. In countries like United States, Canada and many countries in Europe coverage is acceptable and they claim to be covering 70%. Nevertheless, in the rest of world, the coverage ratio is much lower than that. This can be improved by time of course, but still at the moment, it is a stopping issue. Another point is that in urban areas it is probable to find high density of wifi access points that can help in finding a relatively accurate location. But on highways, they would use the cell site technique which results in location with error in terms of kilometers.

Another drawback is their high dependence on Internet connectivity. They need to communicate with the appropriate servers to find location estimates of the heard wifi hotspots and/or cell sites. This introduces mutli-dimension problem; the availability, the cost, and the power consumption. Relying on Internet connectivity that heavily results in total failure if the connection is lost for any reason. Both techniques will be unable to  find location estimate without Internet connectivity unless by using GPS. Also this introduces additional cost to the localization process. One now needs to pay Internet communication fees each time he tries to find his own location. Although caching and pre-fetching can partially solve those problems, but they will introduce other technical problems that may affect the freshness of data. Radio communications consume relatively high amount of energy. That is why depending on how frequent the technique is performing Internet communications, the scheme would consume a significant amount of power. Another major problem is the freshness of the data. Those databases that hold the location information need to be updated from time to another. For example people moving from apartment to another mostly would result in moving at least one access point to another location. If the database is not updated, a wrong location would be calculated for devices querying their location in the proximity of this access point. This needs repeating the process of wardriving from time to another which an expensive operation that adds to the cost of the localization technology.

 

Reference:

http://www.mamir.net/masters/mamir_thesis.pdf

 

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s