Implementation and Testing of Clustered ARAIM in a GPS/Galileo Receiver

This paper introduces a method developed by Honeywell named Clustered ARAIM, which drastically reduces the number of sub-solutions that needs to be computed, compared to the baseline ARAIM algorithm developed by the EU/US Working Group C, without degrading the algorithm performance (in terms of computed protection levels). It consists in grouping individual satellites into so-called clusters. Removing a cluster (or their combination) then covers multiple fault modes at once. Depending on the number of satellites, constellations, and clusters per each constellation, the reduction in the sub-solution count is at least an order of magnitude. However, the performance in terms of the protection levels remains comparable. The paper provides a summary of the algorithm, including certain technical aspects such as the distribution of individual satellites into clusters and optimization of sub-solution characteristics computation, as well as a comparison in terms of computational demands and performance with the standard ARAIM algorithm. The proposed algorithm was implemented in Honeywell dual frequency GPS/Galileo prototype. This paper will show the results from a large number of tests conducted in a representative environment with the aforementioned prototype. These tests include lab tests using a GNSS simulator and live signals as well as flight tests. Large number of simulated scenarios were created in order to assess the capability of the algorithm to detect different type of faults: single satellite faults, multiple simultaneous satellites faults, and constellations faults. The results show that the Clustered ARAIM algorithm provides equal performance compared to the standard ARAIM while requiring considerably less computational power.