Advaced RAIM User Algorithm Description: Integrity Support Message Processing, Fault Detection, Exclusion, and Protection Level Calculation

Advanced RAIM (ARAIM) for vertical guidance has attracted considerable attention both from integrity providers and receiver manufacturers, due to its potential to achieve worldwide coverage of vertical guidance with a reduced investment on the ground segment compared to Space-based Augmentation Systems. Several user algorithms have been published, mostly variants of solution separation and possible optimizations. These descriptions have mostly focused on the definition of the Vertical Protection Level (VPL), because that is what was needed to simulate ARAIM availability as a function of different parameters and constellation configurations. However, an ARAIM user algorithm has many more elements that need to be defined. The purpose of this work is to describe an Advanced RAIM user algorithm step-by-step including: the Integrity Support Message (ISM) processing, the fault detection and exclusion, and the Protection Level calculation – including the Horizontal Protection Level. In this description, we attempt to clarify areas that have remained undefined. We give the contents of the ISM contents, and a clarification of the interpretation of its parameters. These parameters describe both the nominal error behavior and the probability of fault of one or more satellites. The nominal error is characterized by two sets of standard deviation and maximum bias, the first one for integrity purposes and the second one, less conservative, for accuracy and continuity evaluation purposes. We show how to compute the nominal error model as a function of the ISM content, and how to determine which faults must be monitored – including which subset solutions must be computed and compared against the all-in-view solution. In this paper, we make explicit under which conditions a fault must be declared. In addition to the solution separation statistics, we show why it is prudent to include a chi-square test on the residuals as well. We also describe the actions that follow the detection of a fault or faults, and under which conditions fault exclusion can be performed. Although this is not expected to be fundamentally different than the current approaches in horizontal RAIM, there are differences that arise. As mentioned above, the Vertical Protection Level has been defined in several publications (each with small variations). In this paper we address the implementation details in both the VPL and the HPL. First, in case a large number of fault modes need to be monitored, a large number of subset solutions must be computed. We show how to efficiently compute the subset solutions. Second, the PLs that provide good availability typically require an iterative halving algorithm. We describe a method to compute a tight upper bound with very few steps. In addition, we provide the formulas for the Effective Monitor Threshold, the fault free 10-7 error bound, and the 95% bound on the accuracy. A concrete numerical example is given to facilitate the verification of the provided formulas and algorithms