arXivDaily arXiv每日学术速递 周一至周五更新
arXiv math/0602488math.PR
收藏

连续-离散滤波器的分支粒子近似的速率

Rates for branching particle approximations of continuous-discrete filters

Michael A. Kouritzin, Wei Sun

摘要:

在这里,我们分析了一类连续-离散滤波问题的渐近解的有效分支粒子方法。假设$t\to X_t$是一个马尔可夫过程,我们希望计算测度值过程$t\toμ_t(\cdot)\doteq P\{X_t\in \cdot|σ\{Y_{t_k}, t_k\leq t\}\}$,其中$t_k=kε$和$Y_{t_k}$是$X_{t_k}$的扭曲的、损坏的部分观测。然后,构造一个具有观测依赖分支和$n$初始粒子的粒子系统,该初始粒子在时间$t$,$μ_t^n$的经验测量非常接近$μ_t$。每个粒子根据观测时间$t_k$之间的信号规律独立于其他粒子演化,并且在观测时间以小概率分支。对于$ε$非常小的过滤问题,使用本文中考虑的算法需要比其他算法少得多的计算,无论$ε$的值如何,分支或交互所有粒子。我们分析了Lévy稳定信号上的算法,并给出了$E^{1/2}\{\|μ^n_t-μ_t\|_γ^2\}$的收敛速度,其中$\Vert\cdot\Vert_γ$是Sobolev范数,以及相关的收敛结果.

英文摘要:

Herein, we analyze an efficient branching particle method for asymptotic solutions to a class of continuous-discrete filtering problems. Suppose that $t\to X_t$ is a Markov process and we wish to calculate the measure-valued process $t\toμ_t(\cdot)\doteq P\{X_t\in \cdot|σ\{Y_{t_k}, t_k\leq t\}\}$, where $t_k=kε$ and $Y_{t_k}$ is a distorted, corrupted, partial observation of $X_{t_k}$. Then, one constructs a particle system with observation-dependent branching and $n$ initial particles whose empirical measure at time $t$, $μ_t^n$, closely approximates $μ_t$. Each particle evolves independently of the other particles according to the law of the signal between observation times $t_k$, and branches with small probability at an observation time. For filtering problems where $ε$ is very small, using the algorithm considered in this paper requires far fewer computations than other algorithms that branch or interact all particles regardless of the value of $ε$. We analyze the algorithm on Lévy-stable signals and give rates of convergence for $E^{1/2}\{\|μ^n_t-μ_t\|_γ^2\}$, where $\Vert\cdot\Vert_γ$ is a Sobolev norm, as well as related convergence results.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑