论文部分内容阅读
对非线性非Gauss系统,粒子滤波是一种有效的状态估计方法。粒子滤波的关键是建议分布的选择,好的建议分布会改进粒子贫化和样本耗尽等粒子滤波存在的普遍问题。该文用迭代无迹Kalman滤波产生粒子滤波的建议分布,提出了一种新的粒子滤波算法——迭代无迹Kalman粒子滤波。给出的建议分布将最新的观测融入样本过程并修正该过程,从而改进了滤波性能。数值模拟结果表明,提出的算法与常用的无迹粒子滤波、扩展Kalman粒子滤波相比,具有数值稳定、估计结果精确的优点。
For nonlinear non-Gauss systems, particle filtering is an effective state estimation method. The key of particle filter is the choice of recommended distribution, and the good suggestion distribution will improve the common problems of particle filtering such as particle depletion and sample exhaustion. In this paper, Iterative unscented Kalman filter is used to generate the recommended distribution of particle filter, a new particle filter algorithm is proposed - iterative unscented Kalman particle filter. The proposed distributions incorporate the latest observations into the sample process and correct the process, improving the filtering performance. The numerical simulation results show that the proposed algorithm has the advantages of numerical stability and accurate estimation compared with the commonly used non-trace particle filter and extended Kalman particle filter.