作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2008, Vol. 34 ›› Issue (14): 185-187. doi: 10.3969/j.issn.1000-3428.2008.14.066

• 人工智能及识别技术 • 上一篇    下一篇

基于自调整粒子滤波的组合导航方法研究

崔平远,郑黎方,裴福俊   

  1. (北京工业大学电控学院,北京 100022)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2008-07-20 发布日期:2008-07-20

Research on Integrated Navigation System Based on Self-adjust Particle Filter

CUI Ping-yuan, ZHENG Li-fang, PEI Fu-jun   


  1. (School of Electronic Information & Control Engineering, Beijing University of Technology, Beijing 100022)
  • Received:1900-01-01 Revised:1900-01-01 Online:2008-07-20 Published:2008-07-20

摘要: 在非线性模型非高斯噪声条件下,标准粒子滤波在组合系统的观测精度较低时能取得较好的滤波效果,但在高观测精度情况下会导致滤波发散。该文针对这一问题,提出一种自调整粒子滤波方法,根据观测噪声的统计大小,自适应调整似然分布的形状,使之与先验分布重叠的区域更大,有效提高滤波稳定性。将自调整粒子滤波算法应用到组合导航系统中,并在非高斯噪声、观测信息由低观测精度跳变到高观测精度条件下进行了仿真研究,结果表明,该自调整粒子滤波算法在组合导航系统具有高观测精度的情况下依然保持了滤波精度和稳定性。

关键词: 组合导航系统, 粒子滤波, 自调整粒子滤波, 非高斯噪声

Abstract: In nonlinear and non-Gaussian integrated navigation system, the standard particle filter is effective when the observations are not accurate, but it is ineffective and diverges more greatly when the measurements are accurate. In this paper, a self-adjust particle filter is proposed to improve the stability of the filter. The self-adjust particle filter changes the likelihood distribution adaptively according to the statistical characteristic of the observation noises to increase the overlap area between the likelihood and prior distribution. The self-adjust particle filter is applied in non-linear and non-Gaussian integrated navigation system when the observation precision changes drastically from low to high. Simulation results show that the self-adjust particle filter is steady and accurate when the observation precision is high.

Key words: integrated navigation system, particle filter, self-adjust particle filter, non-Gaussian noise

中图分类号: