Аннотации:
The paper deals with the development of the multirate information fusion algorithm for the integrated navigation system. The system is designed for air-launch space launcher used within the conversion programs for non-military space exploration. The integrated navigation system includes a strapdown inertial navigation system (strapdown INS, SINS), CCD-based strapdown astronavigation system (ANS), multi-antenna combined GPS/GLONASS receiver i.e. SINS/ANS/GNSS is under consideration. Loosely coupled SINS/GNSS architecture is used. The innovation-based adaptive Kalman filter (KF) fuses navigation solutions feeded by the above systems to produce a more accurate and reliable navigation solution in the presence of non-stationary Gaussian noise. SINS/ANS/GNSS errors are the states of linear time-varying (LTV) state-space system. KF estimates the system state vector. First-order Gauss-Markov processes are used to model errors time-correlation. Correlation matrices of error shaping filters are adjusted according to body dynamics and environment. KF measurements are constructed as SINS, ANS, and GNSS navigation solution differences. Lagrange polynomials are used to interpolate data from each system. SINS/ANS/GNSS performance is compared to that of loosely coupled SINS/GNSS in numerical simulations. INS/ANS/GNSS implementations are reviewed.