基于奇异值分解的鲁棒容积卡尔曼滤波及其在组合导航中的应用
发布时间:2018-07-14 21:43
【摘要】:为了提高GPS/INS组合导航系统对异常观测值的鲁棒性,引入H∞滤波思想提出一种新的非线性鲁棒滤波.分析H∞鲁棒容积卡尔曼滤波中不同约束水平对滤波结果的影响,指出在一定范围内当约束水平越小时,系统的鲁棒性越强,但容易造成Riccati不等式无解,导致滤波发散.采用奇异值分解代替容积卡尔曼滤波中的Cholesky分解,改善了滤波的稳定性,放宽了H∞鲁棒容积卡尔曼滤波器对约束水平的要求.GPS/INS组合导航实验验证了该滤波方法的正确性和优越性.
[Abstract]:In order to improve the robustness of GPS / ins integrated navigation system to abnormal observations, a new nonlinear robust filter is proposed by introducing H 鈭,
本文编号:2123043
[Abstract]:In order to improve the robustness of GPS / ins integrated navigation system to abnormal observations, a new nonlinear robust filter is proposed by introducing H 鈭,
本文编号:2123043
本文链接:https://www.wllwen.com/kejilunwen/wltx/2123043.html