Advanced RAIM Architecture Design and User Algorithm Performance in a Real GPS, GLONASS and Galileo Scenario

I. Martini, M. Rippl, M. Meurer

Abstract:	In the near future of multiconstellation and multifrequency GNSS signals, the users will experience a significantly increased accuracy, thanks to the larger number of satellites in view and to the ionosphere free observations. Integrity, continuity and availability requirements will mainly drive the system performance. In this scenario Advanced Receiver Autonomous Integrity Monitoring (ARAIM) has attracted significant interest, since it can provide aviation users with worldwide vertical guidance. This paper aims to contribute to the on-going discussion on the ARAIM design proposing an integrity concept and an Integrity Support Message (ISM) architecture satisfying the aviation integrity risk requirement and at the same time maximizing availability and continuity. Several integrity risk allocations to the satellite, the GNSS ground, the ISM ground and the user are analysed and discussed. Caution is required because integrity requirements for vertical guidance in aviation are strict: a continuous trade-off between conservatism, necessary to protect the user against small integrity risk, and efficiency, that provides sufficient availability and continuity performances, has to be performed. A significant amount of conservatism also originates from an intrinsic limitation of new GNSS systems under deployment, like Galileo: the lack of history to characterize the system. The intent of future GNSS would be to have integrity information, e.g. ISM, provided through the navigation message. An even more ambitious goal would be to verify and certify such service with an intermediate Early Version, like the Initial Operation Capability of Galileo planned for the year 2014. Unfortunately, the lack of historical data risks invalidating the attempt. In fact, trying to reach the required confidence through too conservative margins puts extremely strict requirement on the ground monitoring design with consequently unaffordable costs. The idea behind ARAIM is to bring a third monitoring service into play and shift part of the risk of the GNSS ground monitoring to that. This introduces additional degrees of freedom and relaxation on GNSS system design. Another constraint is represented by the necessity of receiver manufactures to have a simple algorithm to be implemented. In fact the user monitoring algorithm cannot be too heavy in terms of computational cost, implementation complexity and consequently certification requirements (each line of code requires an additional certification procedure with consequent costs). Integrity concept must be kept simple and easy to be implemented. This paper addresses first of all the criticism of the overbounding methodology and presents a method for the ground monitoring to assess the ISM. The ground monitoring must assess standard deviation values overbounding the tails of the signal in space error distribution. For this purpose it has mostly a limited sample size, which is significantly smaller than the minimum required by the integrity risk probability. This is also applicable to GPS, which is going to face a modernization phase where the system error distributions might be different from the past. The issue becomes even more critical when considering verification and certification phases, where the minimum required confidence levels are even higher. To solve these problems different overbounding definitions have been proposed in the past. Some of them are constrained to the assumption of unimodal and symmetric distributions. But the ground monitoring must protect any user in the satellite visibility area, that is it must overbound the worst user location error distribution, which in general does not satisfy unimodality and symmetry properties. Furthermore, as suggested in the recent years by the SBAS experience, small biases must be included in the ranging error model instead of using unbiased Gaussian distribution. In this case the usual overbounding definitions cannot be applied without introducing excessive margins. The paper proposes a different approach. It defines the mismodelling risk related to the overbounding as an additional GNSS threat and allocates the relative integrity risk in the satellite failure probability. The overall integrity risk allocation tree will be investigated, providing sensitivity analysis with respect to the mismodelling risk, the sensor stations network coverage, the ISM dissemination strategy and the ISM update rate. The optimum integrity risk tree trade-off able to reasonable relax the GNSS and ISM design requirement will be presented. Furthermore the ISM ground monitoring algorithms will be investigated. The proposed design foresees a short term monitoring addressing high dynamic large error and a long term monitoring. For the short term monitoring the proposed algorithm is based on instantaneous error monitoring instead of standard deviation and mean estimation. For the long term monitoring a modified quantile method is proposed instead of usual standard deviation estimation. The proposed method takes into account also the test accuracy related to the sample size. Comparison with common approaches will be provided, to show the improvement in terms of availability performance. The paper suggests to use complementary and independent methods in the ISM ground monitoring, which lead to a multiplication of integrity risk probability, i.e. to an effective integrity risk share. Finally a design of the ISM information data is performed. In particular it is proposed to provide the user with a three dimensional information instead of a single scalar value. The user receives information in the satellite domain (radial, along-track and cross-track components) instead of in the range domain. In this way the error projection from the ranging domain to the position domain is performed by the user. There is no need for the ground monitoring to add conservative margin to protect a hypothetic worst user. To validate the proposed approach the EVnet global monitoring network of the German Aerospace Center will be used. GPS and Galileo data from around 15 multiconstellation multifrequency reference receivers will be collected and processed to characterize the GPS and Galileo nominal performances. The overbounding values for the satellite orbit and clock errors will be assessed. The ISM data will be generated with the estimated values and disseminated locally with experimental data link. The overall user performance improvement in terms of integrity, availability and continuity will be assessed in particular for aviation users.
Published in:	Proceedings of the 26th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2013) September 16 - 20, 2013 Nashville Convention Center, Nashville, Tennessee Nashville, TN
Pages:	2624 - 2636
Cite this article:	Martini, I., Rippl, M., Meurer, M., "Advanced RAIM Architecture Design and User Algorithm Performance in a Real GPS, GLONASS and Galileo Scenario," Proceedings of the 26th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2013), Nashville, TN, September 2013, pp. 2624-2636.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In