More on GPS Systems

Global Positioning Systems

GPS is an _aid_ to navigation, and does not free the navigator from the need to know and use more traditional piloting and navigation techniques.

“The prudent navigator will not rely solely on any single aid to navigation” (USCG Notices to Mariners)

  • The Space Segment:(How GPS works and accuracy)
  • Navigation (user) Receivers: (DGPS, waypoints, variance and where GPS will work)
  • Survey Systems

The Global Positioning System consists of three interacting components:
1) The Space Segment — satellites orbiting the earth
2) The Control Segment — the control and monitoring stations run by the DOD
3) The User Segment — the GPS signal receivers owned by civilians and military


The Space Segment

The space segment consists of a constellation of 24 active satellites (and one or more in-orbit spares) orbiting the earth every 12 hours. Four satellites are located in each of six orbits. The orbits are distributed evenly around the earth, and are inclined 55 degrees from the equator. The satellites orbit at an altitude of about 11,000 nautical miles. (Earlier plans for the system called for 18 or 21 active satellites.)

How does it work? 
Each satellite transmits two signals: L1 (1575.42 MHz) and L2 (1227.60 MHz). The L1 signal is modulated with two pseudo-random noise signals – the protected (P) code, and the coarse/acquisition (C/A) code. The L2 signal only carries the P code. Each satellite transmits a unique code, allowing the receiver to identify the signals. When a feature called “Anti-Spoofing” is active, the P code is encrypted, and known as P(Y) or Y code.

Civilian navigation receivers only use the C/A code on the L1 frequency (although some high-end civilian surveying GPS receivers can utilize the carrier frequency of the L2 band for more precise measurements).

The receiver measures the time required for the signal to travel from the satellite to the receiver, by knowing the time that the signal left the satellite, and observing the time it receives the signal, based on its internal clock. If the receiver had a perfect clock, exactly in sync with those on the satellites, three measurements, from three satellites, would be sufficient to determine position in 3 dimensions.

Unfortunately, you can’t get a perfect clock that will fit (financially or physically) in a $300 (or even $3000) receiver, soa fourth satellite is needed to resolve the receiver clock error.

Each measurement (“pseudorange”) gives a position on the surface of a sphere centred on the corresponding satellite. Due to the receiver clock error, the four spheres will not intersect at a single point, but the receiver will adjust its clock until they do, providing very accurate time, as well as position.

Since the receiver must adjust its clock to be precisely in sync with GPS time, a GPS receiver can be used as a precise time reference. Some receivers provide a 1 pulse per second output for this purpose.

What accuracy can I expect? 
The Standard Positioning Service (SPS) available to civilian users should give 20 meter horizontal accuracy, however it is normally degraded to 100 meters (95% of the time) due to Selective Availability (SA). (That is, the reported position will be within 100 meters of the true position 95% of the time.) The vertical accuracy is about 1.5 times worse than horizontal, due to satellite geometry. (Satellites are more likely to be near the horizon, than directly overhead.)

Example error budget for commercial navigation receivers:

  • Satellite clock error 2 ft.
  • Ephemeris error 2 ft.
  • Receiver errors 4 ft.
  • Atmospheric/ionospheric 12 ft.
  • Selective Availability 25 ft.
  • Total (root-sum-square) 15 – 30 ft depending on SA

The predicted accuracy is calculated by multiplying the above figure by the PDOP (Position Dilution of Precision) which typically will range from 4 to 6. This gives accuracies of 60 – 100 ft (30 m) without SA, up to 350 ft (100 m) with SA. The accuracy can be improved by averaging readings over some time. When taking readings for this purpose, there is apparently no point in taking the readings more often than every 15 min, or so.

What (and why) is Selective Availability?
Selective Availability is an intentional degradation of accuracy intended to prevent “the enemy” from making tactical use of the full accuracy of GPS. SA is normally on.

Military receivers can use the encrypted P code to get 20 meter accuracy, or better, regardless of the state of SA. In early Feb. 96, the US government passed a law that appears to require the military to turn SA off by May 1, 96.
Apparently the first bill that covered this was vetoed, but the language was added to another bill that did pass, and was signed by the president. However, that bill is worded such that the military could (and did) find a way to legally leave SA on. On March 29, 1996, the White House announced that SA would be removed in four to ten years (i.e. somewhere between 2000 and 2006).

How do some users get centimeter accuracy? 
The 20 to 100 meter accuracy mentioned above applies to single frequency navigation receivers, which are capable of updating the position every second or so. The high accuracy measurements are achieved with much different equipment covered below under “Survey Systems”. These systems use both frequencies, and differential measurements, comparing the data from a roving receiver with that from a fixed receiver at a known location. They may also average the measurements over some period of time. These measurements actually determine the _difference_ in position between the fixed and roving receivers to great precision, rather than determining the absolute position of either one.


Navigation Receivers

Why are my GPS positions consistently wrong?
The chart probably uses a different horizontal datum than the GPS.

What is a horizontal datum, and which should I use? 
A horizontal datum in effect defines where on the earth the lines of latitude and longitude are drawn. In earlier times, surveys were based on points determined by astronomical observations, and by physical measurements on land. This resulted in many slightly different regional Lat/Long grids. The GPS system forces us to use a consistent, world-wide grid.

Positions reported by GPS are based on a horizontal datum called “World Geodetic System of 1984” (WGS84). In the US and Canada, most older (but still current) charts and maps are based on the North American Datum of 1927 (NAD27). Newer nautical charts (including Canadian charts prepared since mid-87) are on NAD83 which is, for all practical purposes, identical to WGS84.

The difference between NAD27 and NAD83/WGS84 varies across the continent. In the Pacific Northwest, an NAD83 position plotted on a NAD27 chart will be about 0.65 seconds (65 ft) south and 5 seconds (330 ft) west of its true position.

In some areas of the world, the local datum may differ from WGS84 by a mile or more. Many GPS receivers can be set to display positions in a local datum rather than WGS84. Most Garmin receivers can display positions in more that 100 different datums. It appears that Garmin receivers store waypoint positions as WGS84 co-ordinates, and convert between the currently selected datum and WGS84 as needed, so that the physical location of a waypoint should not change as you change the current datum on the receiver.

Why does the reported altitude vary so much? 
Primarily due to satellite geometry. To get the most accurate altitude and location, you should use satellites that located at close to right angles to each other and one directly overhead. However, the satellites are more likely to be nearer the horizon, and a receiver will likely choose satellites nearer the horizon in the interests of getting a more accurate horizontal position, since that is what most navigators are interested in. The error in altitude is typically about 1.5 times the horizontal error.

The altitude may also _appear_ to vary more than the horizontal position, since it is given in “normal units” (feet or meters). Also, particularly for those of us at sea level, the real altitude is probably known better than the lat/long, making the error more obvious. (I know the altitude of my boat is 0 +/- tides. When the GPS shows an altitude of 200 ft (or -150 ft) I _know_ it is wrong!)

What is DGPS? 
Differential GPS (DGPS) is a means of correcting for some system errors by using the errors observed at a known location to correct the readings of a roving receiver. The basic concept is that the reference station “knows” its position, and determines the difference between that known position and the position as determined by a GPS receiver. This error measurement is then passed to the roving receiver which can adjust its indicated position to compensate. Unfortunately, the error depends on the particular satellites used to compute the position, so the reference station can’t just say “move all positions 100 meters south”.

The differential reference station computes the errors in the pseudorange measurements for each satellite in view separately, and broadcasts the error information, and other system status information, by some means. A differential beacon receiver receives and decodes this information, and sends it to the “differential ready” GPS receiver. The GPS receiver combines this information with the individual pseudorange measurements it makes, before calculating the position.

For marine use, the US and Canadian Coast Guards (and corresponding agencies in other countries) have established DGPS reference stations that broadcast the correction data over the existing 250 – 350 KHz marine radio beacons. This marine service is available free of charge in the US and Canada, but may only be available by subscription in some countries. Commercial DGPS providers use subcarriers on FM Broadcast stations, and other means, to distribute the correction data.

DGPS will eliminate the error introduced by Selective Availability, and errors caused by variations in the ionosphere, resulting in reported postions within about 10 meters (33 ft.) of the true position 95% of the time, for typical marine DGPS systems using inexpensive navigation receivers. Better receivers can get within 3 meters, or so. The DGPS correction data can be used as far as 1500 KM from the reference station depending on the DGPS setup — if the DGPS is part of a larger monitoring network.

(Note that the advertised range of the marine radiobeacons is only 50 – 200 miles, so other means of data transmission must be used at greater distances.)

What are waypoints and routes?
A waypoint is just a position stored in the GPS receiver’s memory. The receiver can calculate the distance and direction (and time-to-go) to the waypoint, and, if interfaced to an autopilot, will direct the autopilot to steer the boat to the waypoint. A route is a series of waypoints. When navigating a route, the GPS will automatically change the destination waypoint to the next waypoint on the list as it reaches each waypoint. The GPS receiver or autopilot normally sounds an alarm, and requires an acknowledgment, before making any course change.

Can I connect the GPS to a computer/autopilot? 
Most navigation receivers have NMEA-0183 data outputs to send data to autopilots and other instruments. Unfortunately, NMEA-0183 provides for several different “sentence formats” to transfer the same data, so it is possible to have two pieces of equipment which both legitimately claim NMEA-0183 compliance, but can’t communicate because they disagree on the specific sentences used. NMEA-0183 is a standard developed by the National Marine Electronics Association for data communications between marine instruments.

NMEA-0183 data is plain ASCII text sent at 4800 baud. The signal levels are not really RS-232 (as used on most computer serial ports) but will usually work when connected directly to an RS-232 port.

Many receivers also have proprietary data formats which are used (in the case of navigation receivers) to transfer waypoint lists, track logs and other data between the GPS and a computer, and also to pass data which is not covered by the NMEA standard.

Will my GPS work in a car/plane/forest/cave…?
The GPS signals are absorbed by most materials, so a GPS receiver needs a fairly clear view of the sky to operate. There are some reports that a true multi-channel receiver, such as the Eagle Accunav, will perform better in marginal conditions than a single-channel receiver, such as the Garmin 45.

Other reports seem to indicate that the difference between these receiver types is very small, or non-existent. Some of the varying reports may be due to different definitions of “forest canopy” or “dense tree cover”.

Many users have good results placing the GPS receiver on the dashboard of a car, right against the windshield. However, some cars (Pontiac was mentioned specifically) have a transparent metallic film embedded in the windshield as a heater for defogging. This film apparently also acts as an excellent shield against the GPS signals. In those cars, you will need an external antenna. Note that Garmin’s non-aviation models will not report navigation information if the speed exceeds 90 knots, so they will not be any use in flight.

*** WARNING ***

Government or airline regulations may prohibit the operation of radio receivers (including GPS) and other electronic equipment during some or all of a flight. Ask permission from the flight crew before using GPS on a commercial flight.

As indicated above, forest cover will block the signals to some extent. If you can receive some satellites, some reports indicate you may have a better chance of getting a fix if you keep moving, than if you stand still. Land masses such as a cliff (or a concrete building) will block the view of a major portion of the sky, and make getting a fix more difficult.


Survey Systems

GPS survey systems were one of the first uses of commercial GPS. These units are more accurate than the typical navigation units but rely on post-processing of the data collected by roving receivers and a fixed reference receiver, and on averaging the data collected over a period of time, by using carrier phase tracking, (and other techniques) to get the increased accuracy. These systems can have an accuracy of better than 1 cm for the very expensive models.Survey systems range in price from US$7,000 to $30,000, or more.

DGPS Survey Systems.
For some survey work short range DGPS systems are used. They operate over short distances and achieve accuracies of 0.5 to 1m. In this case the accuracy is mostly down to the quality of receivers and the short distance between the receiver and base station. Clearly the base station is usually owned by the user as well as the mobile which does increase the cost by a factor of 2!

Static Survey Systems.
Two GPS recievers can be placed in separate locations for a period of time (for short distances 2mins upto 1hr) and the raw pseudo range data can be collected. This data can then be post-processed and a baseline established i.e. range and bearing. The position can be as good as 1mm but may be less good under poor satellite geometry or larger distances. This method can then be used to transfer the knowledge of one accurate point (e.g. a trig pillar (that’s what we call them in UK) or in our case the GPS antenna on the roof of our office) to a new point. We use this regularly to survey in new DGPS stations.

Kinematic Post-processed Systems.
In the same way as static surveys, raw pseudo range data is recorded at a fixed site and a mobile. You can then post process the data to see accurately where you have been. Again the distances between the static and mobile systems determine the accuracy but generally 0.1m is possible.

Real Time Kinematic GPS. 
The newest thing on the block! RTK systems work in similar ways to DGPS short range systems but mathematically like Kinematic post processed and can achieve accuracies of around 7cm in real time.