Alignment Estimation for Star Trackers and Inertial Measurement Units Based on Attitude Increment Measurements

Star trackers, due to their high precision and drift-free characteristics, can be used to evaluate the attitude accuracy of airborne inertial navigation systems (INS) during flight. The evaluation process requires high-precision calibration for the alignment between the star tracker and the inertial measurement unit (IMU). However, the alignment may change due to factors such as vibrations of airborne platforms and variations in operating temperature, necessitating real-time estimation. Existing estimation methods rely on attitude information from the INS and stellar observations from the star tracker, but the former has the problem of cumulative error, and the latter is affected by the atmospheric refraction of starlight under terrestrial observation conditions. This paper proposes an alignment estimation method based on attitude increment measurements. Since the attitude increments of the IMU only contain minor cumulative errors, and the attitude increments of the star tracker can eliminate the effects of starlight atmospheric refraction, we utilize the two attitude increments to establish a measurement equation for the misalignment. Considering the uncertainty in the sensitivity matrix of the measurement equation, a misalignment estimator is designed using the regularized robust filtering framework. By avoiding the introduction of cumulative attitude errors and starlight atmospheric refraction, the proposed method enables high-precision alignment estimation. Additionally, because the attitude increments are established in the sensor coordinate system, complex celestial-to-ground coordinate transformations are unnecessary. Finally, simulations and experiments were conducted. The results indicate that the accuracy of the proposed estimator is superior to 1 arcsecond.