摘要
为降低GPS/DR组合导航系统在复杂导航环境下的定位精度及DR误差累积,提出一种性能较优的数据融合算法。在传统粒子滤波(PF)的基础上,考虑最新观测值的影响,使用基于平方根二阶差分的高斯混合(GM)模型给出粒子滤波的建议分布,采用基于蒙特卡罗的重要性采样和进化再采样方法减轻PF样本退化问题,增强样本多样性。实验结果表明,与PF算法、GMPF算法相比,该设计能提高组合导航系统的综合导航定位性能。
In view that the GPS/DR integrated navigation system in a complex environment has problems such as lower positioning accuracy,error accumulation of the DR and so on,a data fusion algorithm which has better performance is presented.Considering the latest observations of the traditional Particle Filter（PF）,the Gaussian Mixture（GM） model based the Square Root of Second-order Divided Difference（SRDD2） is introduced to produce proposal distribution.The importance sampling based on Monte Carlo and the evolution re-sampling are employed respectively that do not reduce sample degradation of the traditional PF but enhance the diversity of the samples.Experimental results show that this design has better comprehensive navigation and positioning performance for integrated navigation system than PF algorithm,GMPF algorithm.
出处
《计算机工程》
CAS
CSCD
2012年第12期268-271,共4页
Computer Engineering
基金
广东省自然科学基金资助项目(9451064101003233)
广东省科技计划基金资助项目(2011B010200011)
广东石油化工学院青年创新人才培育基金资助项目(2009YC04)
关键词
GPS/DR组合导航
误差累积
数据融合
平方根二阶差分
粒子滤波
GPS/DR integrated navigation
error accumulation
data fusion
Square Root of Second-order Divided Difference（SRDD2）
Particle Filter（PF）
