弹载大视场星惯组合深度融合导航技术

To lower the requirement for attitude adjustment and star observation in the middle stage of missile-borne Stellar-Inertial Navigation System (Stellar-INS) combination flight, and to improve the accuracy of the Stellar-INS attitude, this paper proposes a deep fusion navigation method for the Stellar-INS with Large Field of View (LFOV) star sensors. The Narrow Field of View (NFOV) star sensor outputs the star vector, while the LFOV star sensor can output both the attitude and the star vector information. The observation equations based on the output attitude and the star vector information from the star sensors are derived, respectively, and the correlation of the Stellar/INS combined method with the LFOV/NFOV is analyzed. Considering the Stellar-INS installation error, gyro error and initial platform error angle, this study establishes a total error term model of the Stellar-INS, and proposes a deep fusion navigation method using the linear Kalman filter. Finally, different performance of the LFOV/NFOV is analyzed via mathematical simulation under the constraint of different attitude observation paths. The results show that the deep fusion navigation method for the Stellar-INS with the LFOV star sensor can not only reduce the constraint requirements for attitude adjustment and observation, but also improve the performance of the combined attitude.