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.