A High-Accuracy SINS/RCNS Integrated Navigation Algorithm

This paper addresses the issues of SINS/RCNS (Strapdown Inertial Navigation System/Refraction Celestial Navigation System), specifically in areas where navigation accuracy degrades in virtue of failing to establish the stellar atmosphere refraction model accurately. In this scheme, the atmospheric model is established in the form of a combination of exponential model and atmospheric density error coefficient, which represents the difference from the actual atmospheric density. Thus, the atmospheric density error coefficient can be extended to the SINS/RCNS state equation as the state quantity, which is to be estimated. And the relationship among the stellar refraction angle, the position error and the atmospheric density error coefficient is derived. The atmospheric density error coefficient can be estimated and compensated by the adaptive Kalman filter. The typical orbit simulation results indicate that, compared to the traditional SINS/RCNS, the method proposed in this paper effectively overcomes the problem of inaccurate atmospheric modelling and significantly improves the accuracy of SINS/RCNS integrated navigation.