Optimal GPS/GALILEO GBAS methodologies with an application to troposphere

• Main Author: Guilbert, Alize
• Format: Others
• Published: 2016
• Online Access: http://oatao.univ-toulouse.fr/16120/1/Guilbert.pdf

In the Civil Aviation domain, research activities aim to improve airspace capacity and efficiency whilst meeting stringent safety targets. These goals are met by improving performance of existing services whilst also expanding the services provided through the development of new Navigation Aids. One such developmental axe is the provision of safer, more reliable approach and landing operations in all weather conditions. The Global Navigation Satellite System (GNSS) has been identified as a key technology in providing navigation services to civil aviation users [1] [2] thanks to its global coverage and accuracy. The GNSS concept includes the provision of an integrity monitoring function by an augmentation system to the core constellations. This is needed to meet the required performances which cannot be met by the stand-alone constellations. One of the three augmentation systems developed within civil aviation is the GBAS (Ground Based Augmentation System) and is currently standardized by the ICAO to provide precision approach navigation services down to Cat I using the GPS or GLONASS constellations [3]. Studies on-going with the objective to extend the GBAS concept to support Cat II/III precision approach operations with GPS L1 C/A, however some difficulties have arisen regarding ionospheric monitoring. With the deployment of Galileo and Beidou alongside the modernization of GPS and GLONASS, it is envisaged that the GNSS future will be multi-constellation (MC) and multi-frequency (MF). European research activities have focused on the use of GPS and Galileo. The MC/MF GBAS concept should lead to many improvements such as a better modelling of atmospheric effects but several challenges must be resolved before the potential benefits may be realized. Indeed, this PhD has addressed two key topics relating to GBAS, the provision of corrections data within the MC/MF GBAS concept and the impact of tropospheric biases on both the SC/SF and MC/MF GBAS concepts. Due to the tight constraints on GBAS ground to air communications link, the VDB unit, a novel approach is needed. One of the proposals discussed in the PhD project for an updated GBAS VDB message structure is to separate message types for corrections with different transmission rates. Then, this PhD argues that atmospheric modelling with regards to the troposphere has been neglected in light of the ionospheric monitoring difficulties and must be revisited for both nominal and anomalous scenarios. The thesis focuses on how to compute the worst case differential tropospheric delay offline in order to characterize the threat model before extending previous work on bounding this threat in order to protect the airborne GBAS user. In the scope of MC/MF GBAS development, an alternative approach was needed. Therefore, in this PhD project, Numerical Weather Models (NWMs) are used to assess fully the worst case horizontal component of the troposphere. An innovative worst case horizontal tropospheric gradient search methodology is used to determine the induced ranging biases impacting aircraft performing Cat II/III precision approaches with GBAS. This provides as an output a worst case bias as a function of elevation for two European regions.The vertical component is also modelled by statistical analysis by comparing the truth data to the GBAS standardized model for vertical tropospheric correction up to the height of the aircraft. A model of the total uncorrected differential bias is generated which must be incorporated within the nominal GBAS protection levels. In order to bound the impact of the troposphere on the positioning error and by maintaining the goal of low data transmission, different solutions have been developed which remain conservative by assuming that ranging biases conspire in the worst possible way. Through these techniques, it has been shown that a minimum of 3 parameters may be used to characterize a region’s model.