GALILEO Integrity Performance Assessment Results And Recommendations

Within the frame of the European Space Agency (ESA) project Galileo integrity performance assessment (GIPA, ESA contract no. 15393/01/NL/DS), integrity performance of the Galileo system has been analysed in detail. The opinions expressed in this paper are those of the authors and do not necessarily reflect the official views or policy of ESA. Both, ground integrity channel (GIC) as well as user level performances have been investigated. Several critical issues related to Galileo system integrity performance have been identified and a set of ten use cases has been defined with each of the ten use cases tackling a different issue. Within these use cases not only the baseline Galileo integrity architecture but also alternative approaches have been considered. The main focus has been laid on algorithmic issues. Operational aspects like communication network or up-link issues have not been considered within this project. Among the critical issues are the signal-in-spaceaccuracy (SISA) representation (scalar versus vector SISA), the SISA refresh rate, SISA checking margin assumption, integrity flag (IF) spatial concept, integrity monitoring station (IMS) network, IMS measurement quality, navigation signal, satellite orbit and clock and IMS clock error feared events as well as elevation masks at user and IMS level. The sensitivity of the integrity performance to these critical issues has been investigated partially by Monte-Carlo simulation, near end-to-end simulation, or analysis. Results with respect to false-alarm rate and missed-detection probabilities at GIC and user level as well as availability at user level have been obtained for the use cases. The first part of the paper gives a brief introduction to integrity and integrity performance at GIC level and user level. The basic situation is visualised and the functionality and performance of the Galileo integrity ground segment, the safe detection and isolation of satellite clock and orbit errors on the basis of the measurements taken at IMS, are characterised. In the second part of the paper, the use cases and underlying assumptions are described. A userequivalent-range-error (UERE) budget was adopted from Galileo phase B1. Orbit and clock errors were modelled according to the latest available early trials study results. A priori fault probabilities have been set to obtain statistical significant results in an efficient way and with not too high computational effort. Results of the individual use cases are given in the third part of the paper together with a discussion and conclusions that could be drawn from the results. In the last part of the paper recommendations are given with regard to different integrity architecture options. The main conclusion that could be derived is that under the environmental conditions that have been assumed in the simulations the current Galileo baseline integrity architecture seems to be well in line with current integrity performance expectations. Among the main recommendations are: • Use a scalar SISA value for broadcast. A vector SISA approach is only marginally better (in case of global integrity approach), but computational effort and transmission bandwidth are higher. • The SISA refresh rate is not critical to the integrity checking performance at GIC level as far as a safe checking margin is provided on top of the real errors. However, there is a severe impact on receiver autonomous integrity monitoring (RAIM) availability at user level. • As a spatial concept use a global integrity approach. The achievable performance will depend mainly on the density of monitoring stations in a certain region and the quality of the pre-processed measurements. • Implement a depth-of-coverage (DOC) 4 network. While a DOC 3 network provides too less redundancy, the benefit of a DOC 5 network at ground segment level is only small, when systematic effects on measurements are low (not considering monitoring station faults). • Choose an IMS elevation mask of 15°. A too low elevation mask will introduce too much systematic effects, while a too high elevation mask significantly reduces the number of available measurements. • Choose a user elevation mask of 10°. For the user it is most important to have as much measurements as possible. Finally, a note on caution is given as every simulation – as well as any experimenting with real data – has its limitations that have to be kept in mind to understand or extrapolate the results and conclusions. These limitations are briefly presented.