Waze Beacons and the problem of geolocation without GPS

Today, whether you want to know where you are or where you want to go, it always imply location technology. With the decline use of paper maps we gradually became dependent to those technologies, the first one being the GPS. This dependence is especially noticeable when we cannot have access to it : a good example is when we lose the signal in tunnels.

It’s to address this problem that Waze has intervened by developing a new geolocation technology independent from GPS : the Waze Beacons. This time the society doesn’t rely on it’s engaged community but try to broaden it with tunnels operators. The strategy is to partner with the later to install the famous beacons on tunnel walls: they are small devices that replace the GPS signal with Bluetooth to provide the location information. There is only one limit to that: the need of a Bluetooth-enabled smartphone, but that’s not a big problem because this feature is widely spread nowaday.

Waze Beacon – What it looks like (Source)

This is a relatively affordable program, as it is recommended to install 42 beacons (at $28.5 each) per mile so $1200 per mile. Being open source, the technology will be incorporated by GPS providers and delivered to all users, Waze drivers or not. There is ongoing experiments in 4 cities, including Paris with tests being held in the 4 tunnels of the A14.

But that is a very specific answer to a “small” problem, because we can lose GPS signal almost everywhere, particularly indoors or in rural areas. Usually there is two main methods that can be used to geolocate a device when the GPS is not available :

Wi-Fi-based positionning is one of the most efficient and has greatly developed in the past few years. It takes advantage of the growing number of access points in dense urban areas and is particularly efficient indoors. In this category exists different techniques which I’ll not describe, except one of the most used, which is called “fingerprinting”: it records in a database the signal strength from the nearby access points along with the known coordinates of the device. So during the tracking, the unknown location is compared to those stored in the database, the closest one being considered as the estimated device location. It can be really precise except in very changing environments because modifications in urban furniture and infrastructures changes the “fingerprint” linked to each location.

The second method is the good old one, used when others are not available. It is network-based and measure the signal strength received from access points in range. Because those points location are known, we can determine the position of the device with the technique of trilateration (or multilateration with more than 3 points). It is not the most accurate technique and require to work closely with service providers as it doesn’t work if they don’t install the needed components in their infrastructure. There is often progress in this field thanks to legislative developments: for example an important breakthrough has come with the system “Enhanced 911” with which when the 911 is called the location of the caller is automatically determined (a similar system exists in Europe with the 112).

Even if there is geolocation solutions working without GPS, today the trend is not to chose one method or another, but to combine them. A good illustration is the system A-GPS (for Assisted GPS) which is used by modern smartphones which combine internet network and GPS signal. In fact each time we don’t use the GPS for 4 hours, the device need to download the new ephemeris, a table of the position of each satellites. The time to get a first position with a “plain” GPS can be very long (few minutes) as the download speed via satellite signal is relatively low. With the A-GPS the download is made via internet, which can takes only few seconds. So the combination of different methods improve the results.

To conclude I want to broaden this subject, because geolocation is not only a matter of technology to use. When we only have photos of an unknown area, we need a deductive geolocation method. For example the journalists group Bellingcat, specialized in social media investigation has used in 2014 crowdsourced efforts to locate ISIS training camps based on militants photographs. The method is to detect important features in the photographs or videos and to compare them with freely available satellite imagery such as Google Earth.

To illustrate, the below photograph is containing precious information in it’s background: a long bridge that cross the river and a large building in the middle of smaller ones. After some researches, Bellingcat group has located this photograph in the north side of the Tigris river in Mosul (the camera pointing south), and so participated, whith their own means, to a better knowledge of ISIS groups location: