基于EKF的多模型自适应组合导航算法

Multi⁃model Adaptive Integrated Navigation Algorithm Based on EKF

系统非线性以及噪声的不确定性是非线性系统滤波的两大问题,直接采用线性模型和固定的噪声模型会影响系统的估计精度。针对复杂环境下组合导航系统模型非线性和量测噪声统计特性未知多变而引起的定位精度下降问题,提出了一种将交互式多模型(Interactive Multiple Models,IMM)估计与扩展Kalman滤波(Extended Kalman Filter,EKF)相结合的交互式多模型扩展Kalman滤波(IMM-EKF)算法。通过模型集自适应调整策略,该算法能够快速估计出当前时刻量测噪声统计特性,以增强系统的鲁棒性;引入EKF,使得本算法能够更好地应用于非线性系统。以GNSS/SINS组合导航系统为例,通过仿真和实际道路测试,结果表明:本算法能够有效降低量测噪声统计特性未知多变对系统造成的不利影响,使得系统有更好的估计精度和鲁棒性。在实际道路测试中,所提算法相比于单模型结构,东向、北向的速度误差和位置误差均方根误差分别下降了56%、24%和63%、69%,水平定位精度明显提高。

System nonlinearity and noise uncertainty are two major problems of nonlinear system filtering,and direct⁃ly using linear model and fixed noise model will affect the estimation accuracy of the system.In complex environments,ai⁃ming at the problem of the decrease of positioning accuracy caused by the nonlinear model of the integrated navigation sys⁃tem and the unknown and variable statistical characteristics of the measurement noise,an interacting multiple models⁃ex⁃tended Kalman filter(IMM⁃EKF)algorithm is proposed,which combines interacting multiple models(IMM)estimation with extended Kalman filter(EKF).The algorithm can quickly estimate the statistical characteristics of the measurement noise at the current moment through the adaptive adjustment strategy of the model set to enhance the robustness of the system.The introduction of EKF enables the algorithm to be better applied to nonlinear systems.Taking the GNSS/SINS integrated navi⁃gation system as an example,through simulation and actual road tests,the results show that this algorithm can effectively reduce the adverse effects of the time⁃varying statistical characteristics of measurement noise on the system,it makes the system have better estimation accuracy and robustness.In the actual road test,compared with the single⁃model structure,the root mean square error of the east and north speed error and the position error are reduced by 56%,24%and 63%,69%,respectively,and the horizontal positioning accuracy is significantly improved.