Doubly filtered navigation method

Abstract

A method of navigating a vehicle (A) by means of a navigation algorithm arranged to determine spatial information (p(xr, yr, zr), v, a(Θxr, Θyr, Θzr), PL) on the basis firstly of inertial measurements (Mi) coming from a signal processor circuit () for processing the signals from an inertial measurement unit (I) and secondly of pseudo-distance measurements (Mpd) determined in response to receiving signals from positioning satellites, the measurement processor circuit () having a calibration input adjustable on a self-calibration value (VCal) in order to reduce the influence of an error of the inertial measurement unit (I) on the spatial information (p(xr, yr, zr), v, a(Θxr, Θyr, Θzr), PL) supplied by the navigation algorithm. The navigation algorithm makes use of at least: a Kalman-Schmidt filter () arranged to detect faulty measurements among the measurements and to prevent them from being used in determining the spatial information (p(xr, yr, zr), v, a(Θxr, Θyr, Θzr), PL), a Kalman filter () that is set to supply at least one estimate (E) of at least one error affecting the inertial measurement unit (I) and covariance (Cov) associated with said estimate, and a consolidation algorithm () programmed to compare the covariance with a first covariance threshold and to store the estimate if the covariance is less than the first covariance threshold in order to update the self-calibration value by means of the estimate of the error of the inertial measurement unit.