基于采样的非线性滤波算法比较

在处理目标跟踪等动态系统实时估计问题中,通常采用EKF作为状态估计方法提高估计精度。由于EKF进行非线性估计存在一些缺陷,将系统进行线性化近似存在估计误差,从而影响目标跟踪的精度。为了获得更高的估计精度,介绍了几种非线性滤波算法,包括unscented卡尔曼滤波算法、简单粒子滤波算法以及无味粒子滤波算法（UPF）。分析了这几种算法的原理和实现,对各种算法的适应性进行了比较。通过目标跟踪仿真实验,表明UKF、PF较EKF估计精度和收敛速度有所提高。

Comparison of Nonlinear Filtering Algorithms Based on Sampling

In dealing with real-time estimation of dynamic system,such as target tracking.The extended Kalman filter（EKF） is used as a state estimation method to improve the estimation accuracy.However,there is estimation error in linearizing system due to the defects of EKF in nonlinear estimation,which affects the accuracy of target tracking.Three new nonlinear filter algorithms are presented in order to yield higher estimation accuracy.The three methods are unscented Kalman filter（UKF） and particle filter（PF） and UPF.the algorithms are analyzed.The applications of the algorithms to the state estimation models are compared.Finally,the algorithms are compared through a tracking model simulation.Experiment results show that the proposed algorithms outperforms EKF at convergence speed,consistency and tracking precision.