摘 要：粒子滤波算法是信号与信息处理和统计模拟理论学科之间的交集学科，因而对其研究具有重要的理论意义和长远的实践价值。它在雷达跟踪，语音信号的处理以及传感器等领域都有着广泛的应用前景。交互多模型粒子滤波是一种新的粒子滤波算法，该算法采用多个模型跟踪机动目标，较好地克服了粒子滤波算法由于重采样而造成的滤波多样性减少的困难，为粒子滤波算法提供了更加广阔的应用前景。

本文首先研究了粒子滤波算法，并采用正弦、曲线、直线三种轨迹仿真实现了该算法，从仿真结果可以看出增加粒子数目会提高粒子滤波算法的跟踪性能。为进一步提升粒子滤波算法的应用性，本文还将交互多模型与粒子滤波算法相结合，实现出一种新的算法，该算法通过多个模型不断递推来估计后验分布，仿真结果表明，交互多模型粒子滤波算法相比于粒子滤波算法具有更好的跟踪性能。最后，基于MATLAB GUI将以上两种算法的仿真程序进行了集成，运行结果表明，设计的粒子滤波与交互多模型粒子滤波的用户图形界面具有较好的交互性能。

关键词：粒子滤波，目标跟踪，交互多模型粒子滤波，GUI

Abstract：Particle filter algorithm is an intersection between the signal and information processing and statistical simulation theory disciplines, and thus has important theoretical and long-term practical value for its research. It has a wide range of applications in radar tracking, speech signal processing and sensor fields. Interactive multi-model particle filter is a new particle filter algorithm, which uses multiple models to track maneuvering targets, which can overcome the difficulty of reducing the persity of filtering caused by resample, which provides a wider application prospect for particle filter algorithm

In this paper, we first study the particle filter algorithm, and use the sine, curve, straight three trajectory simulations to achieve the algorithm, from the simulation results can be seen that increasing the number of particles will improve the particle filter algorithm tracking performance. In order to further enhance the applicability of the particle filter algorithm, this paper also combines the interactive multi-model and the particle filter algorithm to realize a new algorithm. The algorithm recursively estimates the posterior distribution by multiple models. The simulation results show that, the interactive multi-model particle filter algorithm has better tracking performance than the particle filter algorithm. Finally, the simulation program of the above two algorithms is integrated based on MATLAB GUI. The results show that the designed particle filter has better interactive performance with the user graphical interface of multi-model particle filter.

Keywords：Particle filter, Target object tracking, Interacting multiple particle filtering, GUI

目  录

第一章 绪论 1

1.1 选题的目的与背景 1

1.2 课题的内容 1

1.3 国内外的研究现状 1

1.4 本文的主要结构 3

第二章 Matlab GUI简介 4

2.1 GUI的基本概念 4

2.2 GUI的结构 5

2.3 利用GUIDE创建GUI 6

2.4 编写程序创建GUI 8

2.5 Matlab GUI的组成 9

2.5.1 nimenu菜单及设计 9

2.5.2 GUIDE界面 9