GPS Part I - GPS and DGPS Navigation

The USAF's NavStar Global Positioning System (GPS) satellite navigation system has taken the world by storm, and together with the Internet is probably one of the best ever examples of dual use military/civilian technologies to emerge in the last decade. GPS promises revolutionary changes in civilian aviation, both in RPT and GA operations, and with the proliferation of Differential GPS (DGPS) will provide accuracies of several feet to its users.

What has been less publicised is that GPS is becoming the technological foundation for a new generation of guided munitions, which promise a significant reduction in the cost of hitting point targets under any weather conditions. What has been even less publicised is that GPS will also provide precision weapons capabilities to nations which have historically lacked the ability to hit anything smaller than a football field.

To fully appreciate the implications of this technology, we will first take a closer look at the strengths and weaknesses of GPS and Differential GPS, review the principle technical and strategic aspects of GPS munition guidance, speculate on other possible applications for GPS guidance and finally review current GPS based munition programs.

NavStar GPS - a Technical Perspective

The GPS system traces its origins to the sixties. In 1960, Aerospace Corporation was founded for the purpose of applying then advanced technology to space and ballistic missile problems. In 1963, the company started work on Project 621, the Global Positioning System, a scheme for replacing strategic aircraft astro-navigation systems with satellite navigation. Whereas astro-navigation systems needed clear sky to track stars, the satellite navigation scheme would use microwaves and a satellite distributed master clock, thereby providing all weather operation and superior accuracy.

The Operational GPS Constellation uses 24 satellites, of which 3 are spares, orbiting in precise 12 hour orbits. The orbit geometry is adjusted so that these orbits repeat the same ground track once per day, and at any point on the Earth's surface at any given time the same configuration of satellites should be seen. The satellites are grouped, nominally in sets of four, into six orbital planes, each of which is inclined at approximately 55 degrees to the polar plane. A user at any point should be able to see between five and eight satellites at any time.

The USAF's constellation of 24 NavStar GPS satellites will revolutionise navigation as we know it, with a wide range of commercial applications as well as its intended military applications. A GPS receiver will measure time of signal propagation from four or more satellites, and use this information to calculate the receiver's position in three axes, using the WGS-84 earth model (Rockwell).

The satellites are controlled via a worldwide network of tracking stations, with the Master Control situated at Falcon AFB in Colorado. The Master Control station measures signals from the satellites to incorporate into precise orbital mathematical models, which are then used to compute corrections for the clocks on each satellite. These corrections, and orbital (ephemeris) data are then uploaded to the satellites, which then transmit them to GPS user's receivers. A GPS receiver can then use these signals to compute its geographical coordinates, measure time, and also then calculate velocity.

The GPS system provides two navigational services, the military Precise Positioning Service (PPS), and the civilian Standard Positioning Service (SPS). PPS provides nominally 17.8 m horizontal accuracy, 27.7 m vertical accuracy and time accurate to 100 nanoseconds. SPS provides nominally 100 m horizontal accuracy, 156 m vertical accuracy and time accurate to 167 nanoseconds, and is available to civilian users. The degraded accuracy results from the use of Selective Availability. In practice, achieved accuracy can significantly better the nominal figures.

The GPS constellation transmits two microwave (D band) carrier signals, L1 at 1.57542 GHz and L2 at 1.2276 GHz. The L1 carrier is modulated with the Coarse/Acquisition (C/A) code and Navigation Message, used for PPS and SPS, and the military P-code, used for PPS only. The L2 carrier is modulated only with the military P-code.

The central idea behind GPS is that of precisely measuring range to several satellites, the positions of which are known. It is then possible to calculate the position of the receiver. The simplest geometrical model to use is the the sphere model - knowing the range to any given satellite places the receiver on the surface of a sphere centred upon the satellite, with a radius equal to the measured range. Knowing the range to two satellites places the receiver on the curve where the two respective spheres intersect. Knowing the range to a third satellite places the receiver at the intersection point common to all three spheres. In practice, however, a fourth range measurement to yet another satellite will be required to compensate for the inaccuracy in the receiver's clock. The result is a set of equations, which if solved yield the position of the receiver and the time.

GPS Signals, Messages and Error Sources

Whilst the basic idea behind GPS is straightforward, implementation becomes somewhat more complex. Both carriers are modulated in phase (conceptually similar to FM radio) with Pseudo-Random Noise (PRN) codes. The C/A code is a 1023 bit 1 MHz PRN code which is unique for each satellite, and is used by military receivers to acquire and lock on to the P-code, whilst in civilian receivers it is the navigational reference signal. The military P-code is a seven day repetition cycle 10 MHz PRN modulation which is imposed upon both the L1 and L2 carriers. It is usually encrypted to P(Y) code, and can only be used if the user has both a military GPS receiver as well as the classified key to decode it with. The P-code modulation on the L2 carrier is used by military PPS receivers to measure ionospheric transmission delays. The third code, the Navigation Message, is a 50 bits/s digital signal which contains six second duration frames comprised of five 300 bit subframes of data.

The Navigation Message is broadcast by each satellite. It contains encoded clock corrections, precise orbital data, correction parameters for an ionospheric model, and Almanacs, which describe approximate satellite orbital data over extended periods of time. A receiver will extract this data from the NM signal, and use it to correct its clock to within 100 (PPS) or 167 (SPS) nanoseconds of UTC time, as well as to calibrate its internal model for the satellite orbit, and its internal model for ionospheric delays.

The C/A and P-codes are used for measuring range to each respective satellite. A receiver will use an internal PRN code generator to produce a PRN code for each of the satellites. This code is then compared to the received satellite signals using a circuit termed a correlator, and if the PRN codes match, the receiver can lock on to the satellite to measure range. When a receiver's PRN code generator is in lockstep with the satellite's transmitted PRN code, the time at which the repeating PRN code starts is extracted. This time is termed the Time Of Arrival (TOA), and the difference between the TOA and receiver internal time, adjusted for the offset between receiver time and GPS network time, is a measure of the distance to the satellite. The range thus calculated is termed Pseudo-Range.

A GPS receiver will use the four or more Pseudo-Range measurements to compute position in Earth-Centred, Earth-Fixed (ECEF XYZ) coordinates. These are then converted by the receiver into geodetic latitude, longitude and height above the surface of an ellipsoid (the Earth isn't round after all !), typically using the WGS-84 Earth model, although other models may be used. As the GPS system assumes the WGS-84 model, use of other models without correction can produce significant positioning errors.

GPS receivers can measure platform velocity by differencing consecutive position measurements, or by measuring the Doppler shift of satellite carrier signals and using this with computed direction to each satellite, to calculate velocity in three axes (like aircraft Doppler Nav inside out). Some receivers may use both methods to improve accuracy.

There are a number of error sources in GPS navigation. Electrical noise in the receiver, a well as phase noise in the PRN code modulation will degrade accuracy by about 2 metres. Each satellite uses four atomic clocks (two cesium and two rubidium) which are highly accurate, but drift in time nevertheless. If satellite clock errors are not corrected by the ground station, this will degrade accuracy by about one metre. Errors in orbital position estimation will also lose about one metre. As well unmodelled signal propagation delays in the troposphere, due changes in humidity, temperature and pressure changing the refractive index, will lose about one metre. Multipath,the effect of satellite signals bouncing off obstacles and arriving from several directions each with different time delays, will degrade accuracy by about 0.5 metre.

The biggest single natural source of error is unmodelled ionospheric signal delay, the model broadcast by the satellites can only compensate for about one half of the possible error, with the resulting error being up to 10 metres. In addition, another effect comes into play, Geometrical Dilution Of Precision (GDOP). Where the angles to the satellites in view are very similar, GDOP will result in inaccuracy in solving the coordinate equation, which will further degrade the solution. Because all of these sources of error will fluctuate in time, users may experience substantially better accuracy at some times, and worse accuracy at other times, depending on the geometry of the satellites in view and ionospheric conditions (the latter a Jindalee problem as well).

Non-military users will also experience an artificially produced error, resulting from Selective Availability. The SA mechanism introduces a time varying bias in the C/A signal, which is designed such that it is virtually impossible to remove. The potential C/A code accuracy of at least 30 metres is thus reduced to the nominal 100 metres.

GPS Receivers

With a system as complex as GPS there are a multiplicity of ways in which a receiver can be built, and this results in a wide range of achieved accuracies and costs across receiver types. The simplest receivers are single channel receivers, which time share a single channel of receiver hardware across the satellites in view. Whilst this saves in hardware costs, it is slow and as a result such receivers do not usually deliver spectacular performance, and are usually ill suited to fast moving platforms such as aircraft. Most high performance receivers today are five channel receivers, which dedicate a channel of receiver path and correlator hardware to each of the five or more satellites they are tracking. Such receivers can also accommodate platform motion more readily, indeed most airborne military receivers use at least five channels.

A typical strategy used in a five channel receiver is for four channels to track satellites, and one to hunt for the next satellite to come into view, so that there is no loss in continuity when switching one of the four channels (eg the IEC SEM-E receiver tracks five, the Collins GEM-III receiver tracks four with a fifth hunting). High cost, high performance military receivers may use up to eight channels to provide best possible accuracy when eight satellites are in view. Typical receivers will use an antenna, a frequency downconverter and receiver hardware. Antennas come in all shapes, sizes and levels of performance. The usual requirement is for upper hemispherical omnidirectional coverage, and antennas will use schemes based upon monopoles, dipoles, volutes, spiral helices or microstrip patches. Military receivers with directional antennas are becoming popular, as this provides improved resilience against jamming and interference.

The key to success is affordability, and affordability is very much a function of complexity. This Rockwell 5-channel commercial GPS receiver fits on a 4 x 2.5 in. printed circuit board, with all receiver functions performed by the chipset on the board. Because GPS receivers are built from mass producible electronic components, they can be relatively cheap, and this is vitally important in both commercial and munition guidance applications (Rockwell).

The Magnavox MX-8000 Anti-jam GPS Receiver (AGR) was specifically designed for operation in heavily jammed environments, and was to be used in the cancelled Northrop AGM-137 TSSAM missile. This receiver uses adaptive nulling techniques to suppress jammers, and beam steering to boost the satellite signal. The receiver will acquire a GPS signal with a 70 dB Jam/Signal ratio (jam power 10,000,000 times higher than GPS signal) and once acquired, track a GPS signal with a 100 dB Jam/Signal ratio (jam power 10,000,000,000 times higher than GPS signal). It is worth comparing the complexity of this receiver with the simplicity of commercial receivers, which are highly susceptible to interference and hostile jamming (Hughes-Magnavox).

GPS Vulnerabilities and Countering Them

For all of its technological splendour, GPS has its weaknesses. The principal of these is the low power level radiated by the satellites, which introduces vulnerability to both interference and jamming. The power level to be detected by a GPS receiver is -160 dBm (decibels wrt one milliWatt, or 10 exp -19 Watts), which is by radio broadcast standards miniscule. A USAF source acknowledged that this was about 1/1000 the received power from a small FM broadcast station.

In practice, this vulnerability has been observed in some parts of the US, where GPS signals have been jammed by harmonic interference from commercial TV stations, operating in the VHF band, and mobile telephone transceiver towers, operating in the UHF band. Even the small amounts of energy leaking from these transmissions into the 1.5 GHz band were found to produce volumes of space, miles across, where airborne GPS receivers were unable to maintain lock and dropped out. This produced much debate in the US, and as a result GPS reception performance will be monitored across the country to determine which radio transmitters may be interfering. These would then be assigned to different channels and frequencies.

In the military context, this vulnerability is a major concern and has produced some heated debate in the US trade press. Even low powered jammers radiating pseudo-noise signals against the GPS carriers could cause typical receivers to either break lock, or fail to acquire satellites from distances of tens of miles. A one Watt transmitter (comparable to a mobile phone) at a distance of 60 km (32 NMI) can in theory prevent a common GPS receiver from acquiring the C/A code. Military receivers locked on to the encrypted P(Y) code are more resilient, and cca 100 W of jam power at 20 km (10.7 NMI) is required to break lock. Significantly, a jammer radiating hundreds of Watts can foil satellite C/A code acquisition at ranges of several hundred nautical miles. The Saddams of this world could potentially disrupt attacks by weapons using many current generation receivers by hoisting such jammers to several thousand feet altitude on devices as simple as tethered balloons.

There are a number of Electronic Counter CounterMeasures (ECCM) which may be used to improve the resilience of GPS receivers to jamming. The first technique is the use of Controlled Reception Pattern Antennas (CRPA), which can electronically form antenna beams in the direction of satellites, thereby boosting the signal relative to the jammer signal. This typically improves Signal/Jammer power ratios by 30 dB (1000 x). Further improvement can be provided by adding a Nuller to the receiver antenna. A Nuller will suppress antenna sensitivity in the direction of a detected jammer, and this will together with CRPA beamforming techniques provide a 50 dB improvement in resilience against jamming. If the receiver is locked on to the P(Y) PPS code, and uses these techniques, jamming power levels of hundreds of kiloWatts at several miles of distances will be required to break lock. It is worth noting that the RAAF's Rockwell MAGR GPS five channel receiver being fitted to the F-111 uses CRPA techniques, unlike many other military receivers currently in use. The USAF has at least two test programs under way to develop intelligent nulling GPS antenna technology.

Differential GPS Systems

Systematic GPS errors as well as the unavailability of GPS P-code to civilian users was seen as a challenge by many in the civilian technical community, and given the potential commercial payoff in using GPS to its full potential, it did not take very long for techniques to be developed to defeat the Selective Availability of the GPS system.

The central idea behind all Differential GPS schemes is that of broadcasting an error signal which tells a GPS receiver what the difference is between the receiver's calculated position and actual position. The GPS error signal can be most easily produced by siting a GPS receiver at a known surveyed location, and by comparing the received GPS position with the known actual position. The difference in positions will be very close to the actual error seen by a receiver in the geographical vicinity of the beacon broadcasting the error signal.

In reality, the successful implementation of DGPS requires somewhat greater sophistication than merely broadcasting differences in absolute position coordinates. This is because an airborne receiver may be tracking a different set of satellites, as well as being in a different position and thus experiencing a different GDOP error. To deal with these problems, DGPS stations will track all satellites in view and calculate corrections for the pseudorange measurements to each and every satellite. This allows compensation for the SA bias error and well as the systematic errors in the pseudorange measurement, particularly ionospheric delays. A DGPS receiver will then apply the correction factors to the pseudorange measurements its uses to generate its navigation solution. The broadcast updates must be several seconds apart to defeat both SA and other error sources.

Differential GPS schemes thus require a beacon to broadcast the local GPS error signal, as well as an airborne GPS receiver which can decode the broadcast, extract the error signal, and apply it to the position estimate which it has derived from the GPS constellation. Accuracies achieved by civilian C/A based DGPS have been as good as 1-3 metres, which has led to their application to areas such as Cat III Instrument condition approaches and landings. This level of accuracy is also more than adequate for the precision guidance of munitions, and DGPS schemes have thus become an area of major military interest.

There are numerous ways in which a DGPS scheme can be implemented. The earliest non-military DGPS applications saw local area beacons implemented by plugging a GPS card into a Personal Computer, wrapping some appropriately written software around it and broadcasting the DGPS error signal on a dedicated VHF radio channel. More sophisticated schemes are of course possible, such as piggybacking the DGPS signal on to a VOR beacon subcarrier, as well as broadcasting encrypted and coded signals to paying or authorised users only. The US FAA is currently looking at the implementation of the Wide Area Augmentation System (WAAS), which will see DGPS error signals broadcast over the continental US from geostationary INMARSAT satellites. Aircraft with suitable receivers will thus be able to exploit both wide area and local DGPS schemes to get the best possible positional accuracy.

The USAF have been decidedly unhappy about this development, as they invested US$21 billion into developing and deploying the NavStar constellation, and expend US$600 million yearly to run it, only to have what they perceive to be civilian freeloaders exploit their system and defeat the built in safeguards against hostile use. As things stand, the deployment of the FAA's WAAS will allow anybody with a suitable commercial DGPS receiver to achieve blind bombing accuracies well in excess of what is provided by basic PPS P-code whilst in US airspace.

This is a nightmare for the USAF, responsible for defending US airspace, as the deployment of DGPS will very quickly lead to a virtual complete dependency of the civilian ATC and traffic management system upon DGPS. The option of shutting down the WAAS system, as well as local DGPS beacons would become extremely difficult, even in wartime, as the civilian infrastructure ever cost conscious will have dismantled much of its existing base of older navaids such as VOR/DMEs and NDBs. Even should much of the VOR/DME/NDB infrastructure remain in place and functional, the next issue to contend with is civilian pilot currency. The ease of using GPS/DGPS will see a steady erosion of the skills base and currency in the usage of conventional navaids. Thus shutting down the high accuracy component of the civilian GPS infrastructure would introduce serious operational hazards, certainly until the flying population regains its currency. The collapse of the VOR/DME/NDB infrastructure in the US will exacerbate this, as the Americans have become very spoilt with the density of navaids in the US. In this respect Australia should look very carefully at what fraction of the existing navaid infrastructure is dismantled with the introduction of GPS.

The military dimension to DGPS is of particular interest, both from an offensive as well as a defensive perspective. The SRI developed wide area DGPS network used for the USAF EDGE project trials (Part 3) demonstrated accuracies within 0.5 metres. The accuracy of DGPS allows both blind bombing and munition guidance with accuracies very similar to that achieved by using laser or TV guided bombs. Given the availability of a DGPS error signal, aircraft nav attack systems become pinpoint accurate under all weather conditions.

Because wide area DGPS beacons can be effective for hundreds of miles, an air force can position beacons within the theatre of operations and provide all suitably equipped aircraft within range of the beacon with DGPS updates. For deep penetration of hostile airspace, beacons which can be interrogated in burst mode by satellites could be planted in hostile territory by special forces, at presurveyed locations. Such beacons could be built to transmit encrypted position readings using low probability of intercept techniques (LPI) to avoid discovery, the interrogating satellite could then broadcast the derived error signal to penetrating aircraft.

What is even more important, is that GPS guided weapons can be fed DGPS derived positions prior to release from an aircraft, and should their flight time be relatively short, very little positional error will be accumulated enroute to the target. Many existing munitions, eg the BGM-109 Block III Tomahawk and the AGM-130/GBU-15 already exploit P-code GPS to improve the accuracy of the inertial midcourse guidance. Adding DGPS corrections will significantly improve the positioning accuracy of the weapon prior to transitioning to terminal guidance. To extend this model further, an aircraft could transmit via datalink both DGPS corrections as well as the updated position of a moving target to a weapon in flight, which would use these to adjust its aimpoint on the way to the target.

The RAAF's AUP Program will see the F/RF-111C fitted with a highly accurate 5 channel Rockwell MAGR GPS receiver, to provide precision velocity and position updates for the aircraft's dual RLG INS equipment. This will provide a significant improvement in the aircraft's accuracy, particularly over long distances. The MAGR receiver employs beam steering techniques for coping with jammed environments.

The task of equipping an aircraft to receive one or another form of DGPS signal update is not difficult, all that is required is a suitable beacon or datalink receiver with a Mil-Std-1553B bus interface, and a modification to the mission computer navigation software. Ideally, the receiver would be designed to accept DGPS signal broadcasts from civilian satellites (eg WAAS), local DGPS beacons, military satellites and UHF datalinks. This would also allow the receiver to identify intentional spoofing, as well as defeat jamming of any of the DGPS channels.

Furthermore, the accuracy of DGPS has spawned a new generation of all weather munitions which will rely wholly upon DGPS/GPS for their midcourse and terminal guidance. How these work will be the subject of Part 2 of this feature.

Special thanks to Dr Don Kelly then of the USAF EDGE Program for his review of the draft of this article.