Star Centroid Positioning Error Correction Aided by Gyroscope Output in INS and CNS

In a traditional integrated inertial navigation system (INS) and celestial navigation system (CNS) setup, measurements from both systems are only fused at the data output phase. Navigation star centroid positioning error persists in the calculated celestial measurement, which inevitably affects the entire integrated navigation system's accuracy. This paper proposes a novel integrated INS and CNS navigation system that includes two filtering processes. The angular velocity derived from the star centroid positioning information provides measurements in the first filter. To correct the star centroid positioning error, the gyroscope's real-time output is used to fuse this measurement via the nonlinear least-square method. In the second filter, the CNS attitude measurement is calculated from the corrected star centroid positioning information. Then, the INS and CNS measurements are combined via a standard Kalman filter to estimate the spacecraft attitude. Comparing the INS/CNS integrated system with the traditional star centroid positioning method, the simulation results illustrate that the proposed method markedly reduces star centroid positioning error and has strong universal applicability to provide similar high-accuracy spacecraft attitude estimations regardless of star sensor specifications.