Article - Issue 7, February 2001
Planes, trains and GPS satellites
Professor Vidal Ashkenazi FREng
GPS: From military system to civilian utility
The Global Positioning System of navigation satellites has been around for nearly 25 years. It started life as a US ‘force enhancer’, Pentagon-speak to emphasise the role of GPS as a dedicated military system, to be used for navigation and guidance of aeroplanes, ships, land vehicles and missiles. Using pseudoranges (ranges affected by measurement errors) to four satellites, a user with a GPS receiver could navigate to within 50 metres using the Standard Positioning System (SPS), and within 10 metres if the user had privileged access to the Precise Positioning System (PPS).
In the early 1980s, two radio astronomers from MIT (Counselman and Shapiro) suggested getting rid of the pseudoranges in the GPS signal, and instead using the two carrier frequencies to make phase measurements. This had already been done successfully for many years in very-long-baselineinterferometry measurements of radio signals from distant extra-galactic radio sources. This suggestion opened up a window to greatly increased accuracy – centimetric GPS – provided one could resolve the so-called integer ambiguity, i.e. the integer number of whole wavelengths in the measured ranges from satellites to receiver. Several receiver manufacturers responded to this opportunity and developed highprecision GPS receivers, capable of relative positioning accuracies from a few centimetres to a few millimetres.
Meanwhile, civilian GPS navigation users were seeking to overcome the large errors in pseudorange point positioning caused by atmospheric delay effects and Selective Availability (SA), the intentional downgrading of the GPS SPS signal’s quality by the US military. They came up with the concept of Differential GPS (DGPS), based on the principle that a user is affected by satellite ephemeris, atmospheric propagation, SA and clock synchronisation errors to the same extent as a nearby Reference Station. By pre-determining the Reference Station’s position to a high accuracy, one could create a system which could be used to calculate corrections to the pseudoranges measured to the different satellites, and transmit these corrections to a multitude of users.
DGPS positioning accuracy is generally of the order of 1 to 2 metres. In fact, not only are there now a large number of commercial DGPS service providers, but the US military, which had introduced SA to downgrade the SPS service available to civilians, uses its own DGPS system whenever higher positioning accuracies are required. In May 2000, the US decided to stop Selective Availability altogether, so that anyone with a simple GPS receiver can now obtain a position to within 5 to 10 metres. In this context one should, of course, mention that the GPS positions are obtained in the WGS84 (World Geodetic System 1984) datum which, although highly consistent on a global basis, is not the same as the national mapping coordinate systems used in the various countries. This difficulty is, however, fast disappearing as more and more countries, especially in Europe, North America and Australasia, are computing and publicising highly accurate coordinate transformation parameters between the GPS datum WGS84 and their respective National Datums.
Further advances by civilian users have led to the development of realtime kinematic (RTK) GPS, enabling a receiver to make carrier-frequency phase measurements while in motion, compute the required integer ambiguities, and provide ‘centimetric GPS navigation’, as opposed to DGPS which only provides metric accuracies. However, this method, which is continuously being improved, is not 100% reliable and is therefore only used in a number of non-navigational engineering applications, such as the monitoring of structural deformations of reservoirs and tall buildings.
GPS user communities
Prior to the 1990s, GPS users fell into two broad categories. The military used GPS for air, marine and land navigation, based on pseudorange measurements Civilians used GPS for high-precision scientific engineering and commercial applications, based on carrier-frequency phase measurements. There was hardly any use of GPS for civilian navigation, other than a few individuals guiding small craft at sea or light aeroplanes in general aviation. GPS was considered too new, too complex and too different a technology for it to be used in conjunction with the more traditional and trusted technologies used in civil aviation, marine and land navigation. After all, it was said, how could one depend on signals coming from a system of satellites, which could go wrong or be turned off at any time? This point of view changed dramatically in the late 1980s and early 1990s when the various civilian navigation communities, most notably Marine Navigation and Civil Aviation, tried the GPS system and became familiar with it. Indeed, GPS changed the whole concept of navigation in the air and at sea, in ways which have little to do with satellites. Concepts of coordinate reference systems, datums, accuracy and integrity, which are applicable and essential with or without GPS, have only been introduced to civil aviation and marine navigation because of the imminent introduction of GPS to civilian navigation, and the need to derive the maximum benefit from it.
Over the last 20 years, the number of civilian applications of GPS has grown continuously and inexorably, from a few in geodesy and mapping, to literally hundreds. These applications now cover structural monitoring in civil engineering, crustal dynamics in geophysics, sealevel rise in oceanography, water vapour content fluctuations in atmospheric physics and meteorology, environmental monitoring, river discharge measurements, control in robotics and, of course, leisure activities such as sport, walking and climbing. With the extensive development of mobile telecommunications, one is now contemplating the development and provision of new, position dependent, ‘info-mobility’ services to the citizen. These services will vary from the automatic notification of road accidents to the emergency services to automatic tolling on highways and cities, tourism, location-dependent billing and even sales promotion (where individual customers could be targeted on the basis of their personal preferences and relative position from certain shops).
However, without doubt, the most important single civilian community which will have benefited substantially from the introduction of GPS is the civil navigation community and, in particular, civil aviation. Close behind, if not an equal beneficiary, is the civil marine community. In both cases, the introduction of GPS has not only contributed significantly to the enhancement of ‘safety of life’ procedures, but also helped increase traffic volumes and hence make more efficient use of airways and marine passages.
GPS and rail transport
This only leaves the rail industry as the last transport community to consider the use of GPS. Broadly speaking, GPS would strongly impact on two important aspects of the industry. Firstly, it would help not only in positioning all the ‘fixed’ assets such as bridges, stations and trackside equipment, in unambiguous geodetic coordinates, but it would also help with tracking all the ‘mobile’ assets, i.e. the rolling stock. The latter function, while being similar to fleet tracking in road transport, differs in its requirements from those expected when tracking fleets of police cars, taxis or containers. GPS could also be used for passenger timetable and information services. Ultimately, GPS can help improve significantly the current and proposed signalling procedures, which are largely based at present on physical devices on rails or at trackside and which require expensive installation and maintenance.
The current block signalling technique is based on a series of multi-coloured (green, yellow and red) signals, which keep trains on the same track a safe distance apart from one another. The system requires the train driver to see the signals and listen to an audible warning, to slow down or stop the train in order to minimise, if not completely prevent, the number of SPADs (Signals Passed At Danger). The Automatic Train Protection (ATP) system, which is to be introduced in a few years from now, will stop the train automatically if the driver either ignores critical signal aspects and/or the train is moving at an excessive speed. The position and speed of the train will be derived from track-side balises, placed at frequent intervals, an odometer and (possibly) a Doppler-radar device or accelerometer. The calculation is carried out by a computer on board which can then initiate the instructions for slowing down or stopping the train automatically. In the short term, a less advanced Train Protection and Warning System (TPWS) has been proposed for introduction in the UK by 2004. This has now been brought forward, following the series of recent train accidents which led to a number of fatalities. The longer-term vision of the rail industry is a Europe-wide standard for ATP called the European Train Control System (ETCS) which could be implemented by around 2015. This would consist of three levels of sophistication, with the final level (ETCS- 3) involving a system whereby trains continuously inform a central computer of their positions, speed and status, which is then used to update ATP information on all trains.
The key point of argument is that at present and in the near future, the position and velocity of a train for ATP purposes is largely based on odometery and trackside equipment which not only requires expensive installation, but also meticulous maintenance. Furthermore, this has to be done not only along a number of important main lines, but along ALL primary, secondary and even local tracks, if the system is to be fully consistent. This is where a combined GPS and inertial navigation device (INS) (similar to that used in civil aviation and marine navigation) could prove its worth, in providing all the required continuous position and velocity data for the effective operation of a future ATP system. Such a combined GPS and INS unit would initially operate in conjunction with trackside devices, but eventually dispense completely of any inaccurate odometer inputs. Furthermore, and critically, it would allow the ATP system to operate not just along primary routes, with wellestablished ETCS-standard trackside systems, but also along secondary or rural lines lacking a significant trackside infrastructure. This trend of GPS coming in initially just to provide an added input, and ending up transforming the whole concept of positioning and navigation, has been in evidence whenever a discipline decided to use it. The trend is likely to continue.