A quick overview of positioning methods

February 23, 2009

Location Based Services (LBS) require us to know where we are to differing degrees of accuracy, so how can we best obtain this information electronically in different scenarios? Below is a list of common technologies we can use today to help determine our location:

  1. Point Of Interest (accuracy of points of interest database and identification by human)
    We can enter a known point of interest, e.g. a church to let an algorithm search through a database of known places and thus find where we are.
  2. Traditional Maps (accuracy dependant on reading of map)
    If we know where we are on a map, we can enter a latitude/longitude by reading scales along the side of the map or ordinance survey grid coordinates.
  3. WiFi Positioning (accuracy <100m)
    Many mobile devices now have WiFi built in, by reading the SSID or MAC Address of a WiFi hotspot, we can search through a database of hotspots with known locations to find out where we are by virtue of WiFi hotspots having a range of <≈ 100m.
  4. Cell Tower ID (accuracy 200m – 32km)
    Using a similar method to WiFi positioning a mobile phone will be connected to one GSM cell transceiver at any one time by knowing the ID of the transceiver and comparing that with a database of known locations we can obtain a location with varying degrees of accuracy. Within urban areas where transceiver density is high, we can expect accuracy in the order of hundreds of metres. Whereas further out in the countryside where there are less users and fewer black spots accuracy will be in the order of kilometres.
  5. Cell Tower Triangulation (accuracy – unreliable)
    As GSM mobile phones are designed to be able to hop between different cell transceivers as the user moves across the land/air, they must be looking out for other transceivers that they may be able to jump to in case they lose signal with the cell they are currently connected to. Because of this the phone will keep a regularly updated list of cell towers it can ‘see’, the relative signal strength of each and their IDs. By comparing the signal strengths of known cell towers it is possible to triangulate the position of the user. This method is rarely used due to the extremely difficult nature of comparing signals with confidence, for example transmitter A may have a signal which is 10dB weaker than transmitter B and therefore we may deduce that transmitter B is closer to us. However, it may be that transmitter A is actually much closer to us, but is hidden behind a hill and thus displays a lower signal than B which has direct line of sight to us but is further.
  6. Inertial Navigation Systems (accuracy losses of > 0.6 nautical miles per hour)
    An Inertial system comprises of motion-sensing devices (often gyroscopes or accelerometers) which are used to determine forces exerted on the moving object over a period of time. By measuring these forces, we are able to determine that the object has moved a certain distance. Inertial systems therefore require the user to calibrate the system prior to it being used and thus are not particularly practical for personal navigation. However, for personal applications it can be used in conjunction with a different positioning method to provide a backup in case the primary method fails. For example, this is sometimes used on high-end car GPS based navigation systems where it is used to keep a ‘lock’ on the vehicle’s position when inside a tunnel. Inertial Navigation Systems have been used predominantly on aircraft but are being replaced by GPS systems in the US and similarly will be the world over when Europe’s Gallileo and China’s Compass are launched.
  7. GPS – Global Positioning System (accuracy <≈ 10m)
    GPS was established in the early nineties, although it was not until the middle of the year 2000 when the accurate positioning channels of the system were available for civilian use. Since then, there has been an explosion in satellite navigation options and devices for consumer use which has helped spur growth in LBS. In simple terms, GPS consists of a constellation of around thirty satellites, each orbiting the earth twice a day. Each of these satellites transmits the time according to its on-board atomic clock. Your GPS device receives these time signals and compares the time sent on each signal. The difference in time between each of the signals lets us calculate our relative distance to each satellite (the time signal will be delayed according to the distance between the satellite and the receiver according to the speed of light). By knowing the distances of the known locations of the satellites we can triangulate our location. For best accuracy GPS requires that the receiver have an open, unobstructed view of the sky as it is easily susceptible to various distorting effects e.g. obstructions in direct line of sight, ‘urban canyon’ effect, ionospheric disturbances etc.
  • Consumer methods of improving GPS accuracy (more to follow):
  • AGPS
  • DGPS

Which point is right?

February 17, 2009
Our new GPS location recording device

Our new GPS location recording device

In the run up to our new OxPoints system being launched, we’ve started to update some of our data and at the same time we’re attempting to make it just a little bit more accurate. Our basic data will be comprised of the main pedestrian and disabled (if different) entrances to each of the buildings we choose to record. Unfortunately recording the location of entrances is fairly difficult using standard GPS devices and our experience with a variety (more on that in another post) showed that nothing we already owned did the job particularly well. So we set our sights on aquiring a new device, namely a Garmin eTrex Vista HCx which would hopefully give an appropriate level of accuracy by means of having a high sensitivity chipset, being able to automatically average multiple readings and being EGNOS/WAAS enabled.

So far the results have been relatively consistent and a fair bit easier to record over previous devices but we now have a different problem – which map is correct? Below are screenshots of the same latitude and longitude (51.75982, -1.25724)  shown on different mapping sources. I was actually standing on the red cross as shown in the first picture.

Read the rest of this entry »