Safety arguments for next generation location aware computing

Abstract

Concerns over the accuracy, availability, integrity and\ud continuity of Global Navigation Satellite Systems (GNSS)\ud have limited the integration of GPS and GLONASS for\ud safety-critical applications. More recent augmentation\ud systems, such as the European Geostationary Navigation\ud Overlay Service (EGNOS) and the North American Wide\ud Area Augmentation System (WAAS) have begun to address\ud these concerns. Augmentation architectures build on the\ud existing GPS/GLONASS infrastructures to support locationbased services in Safety of Life (SoL) applications. Much of the technical development has been directed by air traffic management requirements, in anticipation of the more extensive support to be offered by GPS III and Galileo. WAAS has already been approved to provide vertical guidance against ICAO safety performance criteria for aviation applications. During the next twelve months, we will see the full certification of EGNOS for SoL applications.\ud This paper identifies strong similarities between the safety\ud assessment techniques used in Europe and North America.\ud Both have relied on hazard analysis techniques to derive\ud estimates of the Probability of Hazardously Misleading\ud Information (PHMI). Later sections identify significant\ud differences between the approaches adopted in application\ud development. Integrated fault trees have been developed by\ud regulatory and commercial organisations to consider both\ud infrastructure hazards and their impact on non-precision\ud RNAV/VNAV approaches using WAAS. In contrast,\ud EUROCONTROL and the European Space Agency have\ud developed a more modular approach to safety-case\ud development for EGNOS. It remains to be seen whether the\ud European or North American strategy offers the greatest\ud support as satellite based augmentation systems are used\ud within a growing range of SoL applications from railway\ud signalling through to Unmanned Airborne Systems. The key\ud contribution of this paper is to focus attention on the safety\ud arguments that might support this wider class of location\ud based services.