平滑变结构滤波在航天器组合导航中的应用

针对系统模型不准确或噪声无法建模等导致的组合导航精度下降问题,在闭环间接滤波方案中引入一种基于变结构和滑模控制概念的平滑变结构滤波器(SVSF),该滤波器对于建模的不确定性和给定上限但无法建模的扰动噪声具有较好的稳定性和鲁棒性。将提出的方案应用于航天器SINS/GNSS组合导航系统中,进行仿真验证,并与EKF滤波器进行比较。仿真结果表明:SVSF尽管不是"最优"滤波器,但其精度与EKF相当,可实现位置精度优于10 m,速度精度优于0.01 m/s;在模型误差与测量误差增大和系统突变等情况下,其较EKF更加具有抗干扰能力,可提高导航系统的鲁棒性和稳定性。

Application of SVSF in spacecraft integrated navigation system

In view of the precision degradation ofGNSS/SINSintegrated navigation system due to the fact that the system model is not accurate or the noises can not be modeled, a smooth variable structure filter (SVSF) based on variable structure and sliding mode control concept is introduced in the closed loop indirect filtering scheme. The filter is stable and robust for the modeling uncertainties and the disturbance noise when with an upper bound on the level of un-modeled dynamics or knowledge of the level of noise. By applying the proposed scheme into the spacecraft SINS/GNSS integrated navigation system, the simulation is carried out and the comparison with the EKF filter is made. The simulation results show that the SVSF, though suboptimal, has equivalent precision to EKF, and the position and velocity accuracies are better than 10m and 0.01m/s, respectively. The SVSF's anti-interference ability is stronger than that of EKF in cases such as the model and the measurement have larger errors, the system has a mutation, etc., which can improve the robustness and stability of the navigation system.