Method and device for detecting and excluding satellite malfunctions in a hybrid INS/GNSS system

A data fusion device (1) comprising a bank (3) of Kalman filters including:

• a main Kalman filter generating a principal hybrid navigation solution based on intertial measurements calculated by a virtual platform (2) and raw measurements of signals emitted by all the satellites of a satellite positioning system (GNSS),
• for each satellite, a secondary Kalman filter generating a secondary hybrid navigation solution based on inertial measurements calculated by a virtual platform (2) and raw measurements of signals emitted by all the satellites of the satellite positioning system except for said satellite,
• and a satellite failure detection and rejection module configured so as to implement a process consisting in, for each satellite in the constellation:
• calculating a cross-innovation of the satellite representative of the distance between:
• an observation corresponding to a raw measurement issuing from said satellite, and
• an a posteriori estimate of said observation generated by the secondary Kalman filter generating a secondary hybrid navigation solution based on inertial measurements calculated by a virtual platform (2) and raw measurements of signals emitted by all the satellites of the satellite positioning system except for said satellite,
• performing a statistical test comparing the cross-innovation of said satellite to a threshold, and declaring a satellite to be faulty when its cross-innovation exceeds a threshold.