Overbounding non-zero mean Gaussian ranging error for navigation integrity of LAAS

Abstract

Navigation integrity is one of the most critical elements in the development of Local Area Augmentation System (LAAS) architecture and remains an unsolved key technical problem. For example, in LAAS, the navigation integrity is ensured through the computation of the aircraft position error bounds (i.e., Protection Levels, VPL/LPL) in which implicitly zero-mean, normally distributed fault-free error models are assumed. However, it is observed that true ranging error involved biases. For example, significant values of mean, in satellites ranging error, extracted from LAAS Test Prototype (LTP) data for ten satellites. Therefore, to ensure navigation integrity of LAAS, such true mean error in ranging must be bounded with zero-mean normal distribution. In our past work [1], theoretical bounds were generated and negligibility of mean was sought within Alert Limits. In this paper, these bounds are reexamined and quantified by using true mean values and also considering availability. In this regard, both simulation and experimental data are heavily used for forming inflations on broadcast sigma to bound and accommodate ranging mean error effects.