优化抗差自适应滤波器在深组合导航中的应用

为了提高惯性/卫星深组合导航系统的滤波性能，在抗差自适应滤波算法的 基础上，研究了一种优化抗差自适应滤波算法。该算法通过比较实际预测残差协方差矩 阵和理论协方差阵的差值来生成自适应因子，从而优化抗差自适应滤波。将所研究的算 法应用于惯性/卫星深组合导航系统， 在高动态环境下进行仿真验证， 并与常规卡尔曼 滤波、抗差自适应滤波进行比较。结果表明，优化算法能有效地控制观测异常和动态模 型异常对状态参数估值的影响，所得组合导航位置误差和速度误差明显减小，提高了组 合导航系统的滤波精度。

Optimization of Robust Adaptive Kalman Filter and Its Application in Deep Integrated Navigation

In order to improve the filtering performance of the GPS/INS deep integrated navigation system, an optimized robust adaptive filtering algorithm is proposed on the basis of robust adaptive filtering algorithm. The algorithm generates an optimized factor by comparing the difference between the actual predicted residual covariance matrix and the theoretical covariance matrix. The algorithm is applied to the GPS/INS deep integrated navigation system, simulates in high dynamic environment, and with the conventional Kalman filtering and robust adaptive filtering for comparison. The results show that the optimization algorithm can effectively control the abnormal effects of observation model and dynamic model on the state parameter estimation , significantly reduce the speed error and position error, improve the filtering accuracy of the system.