Search - matlab particle filter

Search - matlab particle filter - List

[File Format] TheApplicationResearchofImprovedParticleFilterAlgo DL : 0: 本文的题目是改进的粒子滤波在组合导航中的应用研究。文档可用caj打开。本课题首先研究了GPS/DR车载定位系统的组合模型，然后在分析了非线性滤波的基础上，引入了粒子滤波。粒子滤波是一种基于递推计算的序列蒙特卡罗算法，它采用一组从概率密度函数上随机抽取的并附带相关权值的粒子集来逼近后验概率密度，从而不受非线性、非高斯问题的限制。虽然粒子滤波存在诸多优点，然而它仍然存在诸如粒子数匿乏、滤波性能不高、实时性差等问题。-The title of this article is to improve the particle filter in the navigation of the applied research. CAJ can be used to open the document. This issue initially on the GPS/DR Vehicle Location System portfolio model, and then the analysis of nonlinear filtering based on the introduction of a particle filter. Particle filter is a recursive calculation based on Sequential Monte Carlo algorithm, it uses a set of probability density function from random samples and weights attached to the relevant set of particles to approximate a posteriori probability density, and thus not subject to non-linear, the issue of non-Gaussian constraints. Although there are many advantages of particle filter, yet it still exists, such as particle number Punic poor, filter performance is not high, real-time poor.
Date : 2026-01-07 Size : 4.93mb User : 阳关
[File Format] ADPF DL : 0: 基于统计决策规则提出自适应采样数粒子滤波算法, 在定义综合性能风险函数的基础, 推导出粒子数与滤波误差方差之间的关系式, 使得在跟踪过程中, 可以根据目标的机动情况在线调节粒子数, 以使跟踪性能达到最优。在Matlab仿真平台下进行了闪烁噪声下的机动目标跟踪实验, 结果表明, 自适应采样数粒子滤波算法是一种有效的机动目标跟踪方法, 跟踪性能较基本粒子滤波算法提高了3.17倍。-Based on statistical decision rules of the number of adaptive sampling particle filter algorithm, in defining the basis for comprehensive performance risk function derived particle number and the filtering error variance of the relationship between the type, makes the tracking process, it can be depending on the target-line adjustment of the motor case particle number in order to achieve optimal tracking performance. In the Matlab simulation platform, carried out the flicker noise under the maneuvering target tracking experiment results show that the number of particle filter algorithm for adaptive sampling is an effective method of maneuvering target tracking, tracking performance than the elementary particle filter algorithm increases 3.17-fold.
Date : 2026-01-07 Size : 283kb User : 胡瑾秋