针对非线性卡尔曼滤波算法发散的抑制方法研究

Research on methods of restrainning algorithm divergence for nonlinear Kalman filter

针对常用的非线性扩展卡尔曼滤波算法在工程应用中所存在的发散问题,文中分析归纳了导致该算法发散的主要原因,同时在目前两种抑制滤波发散非线性算法的基础上,探讨了一种既保证滤波精度又提高自适应能力的改进型自适应滤波算法。并通过捷联惯导系统的初始对准误差模型对3种滤波算法进行数学仿真。仿真结果表明,改进型非线性自适应滤波可有效地抑制滤波发散,并在提高误差估计精度的基础上具有较大范围的自适应能力。

For nonlinear extend Kalman filter algorithm divergence existed in the practice, this paper analyses the main reasons leading to filter divergence. Based on two nonlinear algorithms for restrainning algorithm divergence, a improved nonlinear adaptive fiter is adapted. By SINS（Strap-down Inertia Navigation System） initial alignment error model, these three filters are used to simulate the nonlinear system.Simulation results show that the improved nonlinear adaptive filter can effectively restrain the divergence, improve the accuracy of error estimation and have a large range of adaptive capacity.