论文部分内容阅读
在现有的基于移动窗口函数模型和随机模型系统误差自适应拟合方法的基础上,提出一种基于移动窗口动态导航模型系统误差的随机加权拟合法,在相同的窗口内给出了相应的状态预报向量协方差阵的随机加权拟合。由于动力学模型系统误差难以直接修正,采用修正状态估计误差向量及动力学模型误差向量的方法,实现对动力学模型系统误差的修正,然后利用修正后的动力学模型及相应的协方差阵进行导航滤波计算,有效地抑制动力学模型系统系统误差的影响,提高导航解算的精度。仿真结果证明,采用随机加权拟合后的算法精度优于未进行拟合的卡尔曼滤波和自适应卡尔曼滤波算法。
Based on the existing adaptive fitting method based on moving window function model and stochastic model system error, a random weighted fitting method based on system error of moving window dynamic navigation model is proposed. In the same window, a corresponding Stochastic Weighted Fitting of Covariance Matrix of State Forecast Vector. Due to the difficulty of direct correction of the systematic error of the dynamic model, the error of the state estimation error vector and the dynamic model error vector can be corrected to correct the systematic error of the dynamic model, and then the corrected dynamic model and the corresponding covariance matrix The navigation and filtering calculation can effectively suppress the influence of systematic errors in the dynamic model system and improve the accuracy of navigation solution. The simulation results show that the accuracy of the algorithm after random weighted fitting is better than the Kalman filter and adaptive Kalman filter without fitting.