Michael McFadyen's Scuba Diving - How GPS Works
One of the few good things that ever comes out of the burgeoning defence industry is the development of new forms of technology. One such development that has moved from the sole use of the military over the past decade or so is the Global Positioning System (GPS). Originally developed for the United States military, it had an abundance of applications for its first users, the Navy, Marines, Air Force and Army. Some of these include providing easy navigation for individual Army units to complex cruise missile navigation systems.
In simple terms, a GPS unit works by triangulation from three or more satellites orbiting the Earth. These satellites are the property of and maintained by the US military, primarily for their original purpose. They cover the greater part of the world, although not necessarily for the full 24 hour period of a day in any one location, although this appears to be correct for NSW. The satellites put out coded signals that include their location and very precise time signals. The GPS units receive these signals and by calculating the time difference between the signals from three satellites (due to knowledge of the speed of such signals and the location of the satellite), it can calculate the unit's location on the Earth's surface (or above it if in a plane or helicopter). The GPS unit then provides a digital reading of the location using one or more methods. The most common is latitude and longitude, although Australian Map Grid References (AMGs) are better for land based
Virtually all GPS units not only record locations, they can be programmed with known locations and used to navigate to those locations. In these cases, the GPS unit provides readouts of current heading, correct heading to travel, distance to location, estimated time of arrival and current speed. Some can even provide the times of sunrise and sunset and moon phases!
In reality the GPS system could always provide an extremely accurate method of location. However, the US Military deliberately downgraded the accuracy of the signals from the satellites so that hostile foreign governments could use the system to guide missiles to small objects in other countries. This downgrading, called Selective Availability, was (typically) stated to be a maximum of 100 metres, although in 95% of occasions the accuracy was far greater than this. Countries allied to the US (eg Australia) can obtain GPS receivers for their military that are able to receive other coded signals and thus remove the downgrading. These are reputedly accurate to less than one metre. The downgrading has considerable impact on how GPS could be used by the general public as it could never be guaranteed that the accuracy was acceptable for some uses. For example, an accuracy of 100 metres out in the bush or open ocean is more than adequate. However, 100 metres when sailing into a port through a narrow channel may be dangerous and using it to find a very small site impractical.
There are GPS units that can be made more accurate (say a maximum of 20 metres out when the Selective Availability was switched on). These are called differential GPS units. In brief, these work by comparing a known site with the GPS unit's reading of the site and producing a signal of the variation. This is transmitted by radio from the known site and received by your differential GPS unit. As the variation will be identical (so long as both GPS units are in the same large area and using the same satelites), your GPS unit then calculates the revised location by adding or subtracting the variation. This gives the new reading and is a more precise location than the normal GPS unit. In the major capital cities and other larger urban areas, different ial GPS is possible as the variations are being transmitted as part of the transmission of JJJ-FM broadcasts, although you need a receiver and may have to pay to use it.
GPS received a real boost around the time of the Gulf War when production increased dramatically to meet the demand created by parents purchasing GPS units for their children so they would not get lost in the desert. Since then, the price has dropped and the quality of the product increased (early models had problems in forested areas and used batteries very quickly).
After the Gulf War, the increased production meant that the cost of GPS units came down to an extent where they came within the price range of the average person. As many people are aware, I work in a NSW Government department. We were one of the first organisations to adopt the GPS unit to help manage our field based resources. As well as recording specific locations in the bush (to enable mapping and easy relocation), the GPS unit has been put to considerable uses, including locating helicopter landing pads in remote locations. The GPS unit has meant
considerable savings to our department. For example, when fighting bushfires, a major job has been to accurately record the extent of the fire and transfer this to maps. Typically this was
done by flying over the fire edge in a helicopter with two or three people drawing the edge on a map (at a cost of $750 per hour plus staff costs). Even a smallish fire could take nearly an
hour to record. The three maps then had to be compared and a master one prepared. This meant that it could take two hours to get an accurate update of the fire out to the people involved.
Using a GPS unit connected to a laptop computer (via sophisticated software), it is possible to complete the task in 15 minutes or so and have the results printed out on overhead transparency material for laying on the map in minutes. A dramatic saving in dollars and, more importantly in this case, time and freed up resources.
During research on GPS units for my employer, I carried out studies in 1993-96 using GPS units connected to laptop computers to record the downloaded information. As the unit moves, the software records the location on a diagram and eventually produces a map of the path travelled. Some of my work involved studying the accuracy of the GPS readout. For this, I set up the
GPS unit and laptop in a set location and left it to record what happened. During one such test over a 95 minute period, the GPS unit recorded 2393 strings of information (approximately one
every two seconds). Over this time the GPS reading ranged over an area of almost 190 by 50 metres. The total distance covered was 407 metres but of course, the GPS unit never actually moved location. What was recorded was the variation through the downgrading. The vast majority of the readings were in a small area 30 metres by 15 metres around the actual location with a
small diversion to the south of 30 metres and a very large diversion to the north of 150 metres (outside the stated accuracy of GPS). However, this diversion to the north only occupied about
four minutes of the total recording. Therefore, the GPS reading was quite accurate for about 95% of the time, with a couple of small variations and one major variation.
This sample was repeated in other tests. In one test the location stayed within a small 27 by 30 metre area for 30 minutes except for a 30 second period when it moved more than 200 metres to the north-west before returning to the actual location. It is assumed that differential GPS readings would be far more accurate with little variation.
President Clinton had made a commitment to turn off Selective Availablity for GPS before he retired. On 1 May 2000, the US Government finally switched off the Selective Availability. From this date, GPS units became even more accurate. See Some examples of increased accuracy from GME Web Site and a timeline showing accuracy change on 1-2 May 2000. My experience over the almost two years since then is that the accuracy is now about 10 metres at the worst for 95% or more of the time. This is excellent and a huge improvement on prior to 1 May. However, it should be noted that in the short period after the 11 September 2001 attack on New York, the Selective Availability appeared to be turned back on as GPS accuracy was again way out. It now appears to be back to normal.
How does this all relate to diving you might be asking? Well, as I indicated, the accuracy is now very good and as such is suitable for locating dive sites without any other help. Even when the system was less accurate and despite the above statements about accuracy then, it was possible to use GPS to find dive sites and for other diving related purposes. In reality, the accuracy of GPS readings (from a normal unit) was more likely to be plus or minus 20 metres and from a differential
unit plus or minus five metres. This was accurate enough to record and relocate a shipwreck (so long as it is a decent size, say 30 metres or more long). Of course, now it is really easy to use to find dive sites and shipwrecks.
What can be learnt from the above? Well, the first is that GPS units can be used to record and relocate shipwrecks and dive sites with fairly good accuracy. However, care must still be taken to ensure that the reading obtained (either when recording or relocating) is not the 5% that is a bit inaccurate. This is impossible to really monitor, other than by using a differential GPS unit or using a laptop computer to record the boat's path of travel. The laptop record would highlight any strange readings. Some GPS units and computer software include an indication of how accurate the recording is, although even this is not 100% correct.
The second is that if recording a site, use an average of the readings. Some GPS units have an averaging feature which can be set to record a certain number of readings over a period of time.
The unit then provides an average reading which should, in theory, be more accurate than one recording. If you do not have this auto matic feature, you can manually record a large number
of readings and later average them out mathematically.
In my experience, GPS can be used to locate a shipwreck. I have had my own GPS unit for use at work (during bushfires) and for diving since about 1992. I have used it to locate the SS Tuggerah off southern Sydney without a working depth sounder and without a view of the marks and numerous reef dive locations. I would suggest that the best way to actually relocate a shipwreck or dive site would be to first locate the wreck and take an average of readings of the location (many newer GPS's will do this automaticlly if asked). Then, select two points that when joined cross over the most prominent part of the wreck or site. Take average readings of these sites as well or work them out from the site's location on a chart. Then enter these two additional locations into the GPS unit's navigational section and plot a course to cross both locations. This should prove a more accurate way to relocate the site as if you are looking for one location right on the site, you may miss it if the accuracy is out a bit.
Some things to look for in a GPS unit. The first thing to decide is whether you want a normal GPS unit or a differential unit (usually more expensive). I do not think that differential is now needed for most purposes. I think that most GPS units now come standard with differential capability. The second would be to decide if you want to be able to download to a computer or autopilot (be warned, not all units use the NMEA standard - eg Sony Pixis). The third major thing is whether you would want AMGs as well as latitude/longitude. If you want to use it for bushwalking or four wheel driving, this is an important requirement. These will set the basic requirements for the type of GPS unit you need.
The next things to consider are the size of the unit, the simplicity of use (some are overly complicated to use), battery life (or ability to use 12V power), ruggedness (eg waterproof)
and warranty and back up service. Probably the best hand held units on the market at the present time are the Garmin units.
For computer software, you can use the software that usually comes with the interface (normally very basic). If you need a bit more, I would recommend Garmin MapSource (various versions). See the Garmin website
Another major thing you to need to know when using GPS units that they are only as accurate as the user's knowledge of how they work. You need to be aware that you must set the GPS to the correct Map Datum, not only for the country you are in, but also for the map or chart that you are using. Some GPS units may come out of the box set to datums for countries other than Australia (especially if you purchased it overseas) and needs changing to one for Australia. In addition, there are a number of datums applicable to Australia and you must also select the right one. This will depend on the map or chart you are using as well as what datum is being used by anyone who gives you GPS Readings. The following attempts to explain this a little.
Australian Geodetic Datum - AGD
Most maps and charts in Australia (as at 2001) were created using the Australian Geodetic Datum (AGD). Adopted in 1966, AGD is based on the Australian National Spheroid (ANS). Note some charts (eg the sea floor bottom charts for Sydney) are not based on this datum but it can be hard to tell as most do not have any reference to the datum used.
The following coordinate systems are based on AGD:
For AGD, most GPS units have map datums called AUS66 or AGD66 or AUS84. Some may use AMG66, AMG84 or AGD84. My Magellan 3000 portable GPS unit has AUS66 and AUS84. These are all basically the same and, for most purposes, any one of these can be used and the result will be the same.
Geodetic Datum of Australia - GDA
There is a new international datum which has been adopted in Australia (in 2000 I seem to recall). This is called Geodetic Datum of Australia (GDA). It was adopted internationally in November 1994 and GDA is based on an Earth-centred ellipsoid, the GRS80.
The Following coordinate systems are based on GDA:
For GDA, most GPS units use datums called GDA94 or WGS84. My Magellan has WGS84.
Differences between GDA and AGD
The difference between identical longitudes and latitudes if used with AGD and GDA map datums (very confusing names I know) is just over 200 metres. Therefore, if you use one of the coordinate systems but should be using the other, there will be a straightline difference of about 200 metres (at least in NSW).
All of the GPS Readings used in my Web Site (except possibly some on the NSW Shipwreck Page) have been taken using AGD (AUS66) as the Map Datum. If you use any other datum on your GPS (eg one not for use in Australia), then you need to change to one of the correct ones for Australia. If you use GDA (GDA94 or WGS84), you will need to modify the readings on my web site for use with your GPS by either using AGD (AUS66) or by subtracting 5.6' from the given latitudes and adding 4.2' to the given longitudes (approximately in Sydney and NSW). A better way is to download a free piece of software called Geodetics Transformations (V3.42 as March 2009) or GEOD for short). This is available from the NSW Department of Lands Web Site. Go to their web site at GEODand download this small but excellent program. This will enable you to convert readings easily.
An example of what would occur if you use a GPS Reading off my web site but used the wrong datum is as follows. This is the GPS Reading for the wreck of the SS Tuggerah off southern Sydney:
|Reading on my site using AUS66/AUS84||Actual location using GDA94/WGS84||Difference in minutes||Where you will|
be using GDA
|Latitude||34° 08' 21"S||34° 08' 15.3"S||-5.7"||105.1 metres south|
|Longitude||151° 09' 02"E||151° 09' 06.2"E||+4.2"||190.9 metres west|
If you used AGD as datum instead of GDA when using my GPS Readings for the Tuggerah, you would be 105.1 metres south and 190.9 metres to the west of the wreck. In a straight line, this would make you 218.2 metres south-west of the wreck. So you can see how important it is that you use the right datum when using GPS.
You may be able to short cut these calculations or using the software (it works on my Magellan as well as on my Garmin GPS 126 on my boat) by doing the following:
I hope this information explains it a bit for you.
Return to NSW Shipwreck Marks Page.
NSW Land and Property Information Web Site
Help section of GEOD software