At the time of a Live Sky test, there are several unknowns. The unknowns include:
Satellite clock errors
Over time, these errors should be accounted for in the navigation message and corrections broadcast, but because this message is updated infrequently, it is possible for a clock error to exist, which is not being corrected for.
Simulator benefits: Using a satellite simulator, there are no errors on the satellite clocks, unless you wish there to be, and then they are precisely known and can be applied at known times.
Satellite orbit errors
The position of each satellite as declared in the navigation message is different to its exact physical position in orbit. This is due to several orbital errors that are caused in part by the gravitational effects of the Sun, Moon and Earth, which serve to add perturbations to the satellites position.
Simulator benefits: With a simulator, it is possible to either remove all orbital errors and use a ‘perfect’ constellation, or allow fully quantifiable errors to exist in a controlled manner.
Navigation data errors
As with any data transmission system, errors occur in the data as a result of the modulation, demodulation and transmission processes. There is robustness built in as, for example with the GPS system, the last 6 bits of each word of the navigation message are parity bits, and are used for bit error detection. However, errors can still occur, and these will not be accounted for.
Simulator benefits: With a simulator, it is not possible for navigation data errors to occur, unless they are deliberately applied.
The GPS signals have to pass through the layers of the atmosphere, which in its two main parts comprises the Ionosphere and the Troposphere. Free electrons in the ionosphere (70 to 1000km above the earth’s surface) cause the modulation of a GPS signal to be delayed in proportion to the electron density (its speed of propagation through the ionosphere is referred to as the group velocity). The same condition causes the RF carrier phase to be advanced by the same amount. (Its speed of propagation through the ionosphere is referred to as the phase velocity)
Simulator benefits: With a simulator, it is possible to completely disable the atmosphere, thereby removing the errors. Alternatively, errors can be applied to a known model, and are therefore fully accounted for.
GPS signals are line-of-sight, and can be regarded in the same way as rays of light. If a signal ray falls upon an RF-reflective surface at an angle less than the critical angle of internal reflection, it will be reflected, with some attenuation. Therefore, it is possible for a receiver to not only receive the direct line-of-sight ray, but also the reflected version. The receiver has no way of knowing which one of the two is the true LOS signal, so it uses both, and inherits the delay error present on the reflected signal.
Simulator benefits: With a simulator, it is possible to eliminate multipath completely, or to apply multipath to signals using various multipath models. In this way, multipath can be applied in a known, controlled manner enabling its effects on receiver performance to be accurately analyzed, and the appropriate design alterations or mitigations to be applied. With Live Sky, it is impossible to quantify the multipath conditions present at any one time, and therefore impossible to analyze and improve a receiver’s performance in its presence.
GPS signals are very weak when they reach the receivers antenna, due to the fact that they have travelled a long way from the satellites. This makes them vulnerable to interference from external sources. Interference can be deliberate (known as jamming or spoofing) or unintentional. The vulnerability of GPS to interference has been well documented and the discussion is beyond the scope of this page.
Interference not only introduces errors in a receiver’s position computation, but can stop it navigating altogether. The problem this causes if interference is present (and cannot be stopped) during a Live Sky test is obvious.
Simulator benefits: With a simulator, thankfully, no such interference exists by default, but if required, it is possible to simulate it in a controlled and repeatable manner. Interference which changes as a function of the proximity of its source to the receiver can be applied using an interference simulation system such as Spirent’s GSS7765.
Download your free eBook - Effects Of Interference On GNSS Signals.
When you perform testing on a GPS receiver, and it highlights weaknesses in the design, the normal process is to make changes to the design with a view to improving it. To confirm if improvements have been made, you need to repeat the same tests exactly. If Live Sky is being used, it will be impossible to ensure subsequent tests are subjecting the receiver to the same conditions as the original test.
The most obvious difference is the fact that time has progressed, and the constellation visible to the receiver will be completely different. These are factors that by themselves will ensure the test conditions cannot repeat. The other characteristics that will not remain fixed are atmospheric influences and satellite performance.
Therefore, Live Sky is unsuitable as a method for testing with a view to making design improvements.
Simulator benefits: With a constellation simulator, every time a test scenario is run, the signals produced are identical. The scenario will start at the same time on the same date, and the satellite positions will be identical – even down to the relative phase offsets between the different signals. In this way you can guarantee that the receiver is being stimulated with the exact same signals every time the test is run.
With any comprehensive testing, finite and accurate control of the test conditions is essential. Fine-tuning of a design or system parameter can often demand very small, closely-controlled manipulation of the test conditions.
Simulator benefits: With a Live Sky test method, there is little that you have control of. With the exception of the physical location of the test antenna, there is in fact nothing else that you have any control over. You cannot wind back time, disable the atmosphere, adjust the satellite signals, errors, data, orbits – all of which are parameters you need to have complete control over.
A GPS RF Constellation Simulator is a precision piece of test equipment and if properly maintained, its performance is accurately specified and controlled. The fidelity of a simulator’s signals is much better than the signals from a real GPS system, which not only allows advanced testing of a receiver’s true ‘laboratory’ performance, but means that signal noise contributions due to the simulator are well below the level of thermal noise, and therefore will not contribute any noise errors to the test.
Simulator benefits: Two parameters closely related to accuracy are quality and reliability. The precision engineering employed in the simulator’s design and construction, and the quality control processes governing these disciplines ensure that the equipment gives reliable service for many years.
Record & Playback Systems do have a role to play
Thorough evaluation of receiver performance requires that the impact of these various sources of previously described impairments is assessed. An emerging technique for performing this testing is by recording the RF signal for subsequent playback in the lab
Simulator benefits: Simulation allows absolute control of the test environment where individual sources of impairment can be added or removed at will. Simulation also allows the evaluation of signals not yet available from space or extremes of vehicle motion which may be expensive or difficult to trial. Indeed the generation of synthetic signals derived from mathematical models represents the ultimate in control.
No project survives without a sound business case. Those responsible for managing projects and setting budgets will have to take this into account. It is often wrongly assumed that simulation only saves money over real field trials for applications involving high-dynamics on sophisticated platforms. For example, it is very obvious that there is no way a space-grade receiver can be flown in orbit purely in order to test how well it works, but what is often not so obvious is the fact that simulation can prove to be more cost effective for much less sophisticated applications. A few months of drive testing will pay for a simulator and in many cases makes its choice over real field trials academic.
Simulator benefits: A leading European automotive manufacturer calculated that the total cost of performing a real drive test is in the order of £5k per day. Notwithstanding the technical issues with real-world tests already discussed, the financial cost-benefits alone are enough to demonstrate the viability of simulation.