Stellar Refraction-aided SINS/CNS Tightly-Integrated Navigation Method

In order to solve the navigation accuracy degradation due to inaccurate accelerometer bias estimation of the traditional SINS/CNS integrations, a constrained SINS/Refractive CNS (SINS/RCNS) integrated navigation method is proposed for long-range ballistic vehicle applications. The SINS/CNS tight integration model is established by directly utilizing the star sensor image point coordinates and improved SINS dynamic equations. The constraint in the free flight phase is used to derive the virtual accelerometer observations, and the refraction measurement equations are also derived to apply stellar refraction into the ballistic missile. The extended Kalman filter (EKF) is devised to fuse the multi-rate observations of stellar refraction, virtual observation and star sensor image points, in order to estimate the accelerometer and gyro biases accurately on-line. Without adding new extra sensors, the simulation results indicate that there is a significant accuracy improvement in position, velocity and attitude by about 97.84%, 98.61% and 78.70%, compared with traditional gyro-drift-corrected SINS/CNS method. There is also a significant improvement in position, velocity and attitude accuracy by about 28.09%, 67.72% and 79.10%, compared with the traditional SINS/RCNS method. To sum up, the accuracy degradation problem encountered by traditional methods can be addressed by the proposed constrained algorithm with remarkable accuracy improvements. © 2020, Editorial Department of Journal of Chinese Inertial Technology. All right reserved.