基于不同等价权函数的抗差粒子滤波对比研究  

The Comparative Study Based On Robust Particle Filter of Different Weight Function

作　　者：隋玉成[1] 施昆[1] 王宇鸿[1] 熊俊华[1] 

机构地区：[1]昆明理工大学国土资源工程学院,云南昆明650000

出　　处：《江西科学》2010年第6期744-747,共4页Jiangxi Science

摘　　要：粒子滤波与其他非线性滤波相比,具有更好代表性,因而在实际中得到广泛应用。在状态空间中寻找随机样本来对概率密度函数进行近似,以样本均值代替积分运算,来获得最小方差状态估计。在此过程中,含粗差的观测值对赋权产生干扰影响粒子滤波效果。在已有抗差粒子滤波的基础上,讨论了基于不同权函数的抗差粒子滤波。模拟计算并对数据分析表明,IGGⅢ抗差粒子滤波提高了粒子滤波的鲁棒性。Compared with other nonlinear filtering,particle filter has a better representative and is widely used.Approximating probability density function through finding the random samples in the state space,it obtains minimum variance state estimator using the sample mean instead of integral operation.In this process,observe-value which contains crude error can affect the result of particle filter through influencing the Weight.On the basis of the submitted robust particle filter,this paper discusses robust particle filter based on different weight function.Simulation and analysis of data show that IGGⅢ improved the robustness of particle filter.

关 键 词：抗差粒子滤波 等价权函数 鲁棒性 IGGⅢ 

分 类 号：P207[天文地球—测绘科学与技术]