利用粒子滤波进行动态精密单点定位观测异常影响控制

Influences Control of Outlying Observations in Dynamic PPP by Particle Filtering

采用粒子滤波控制观测异常影响,提高动态精密单点定位精度。粒子滤波是一种非高斯噪声分布的动态滤波,通过重点概率密度进行随机采样以获取高精度状态参数;根据观测噪声概率密度、状态噪声概率密度以及重点概率密度等因素确定粒子权值,降低受污染粒子对定位结果的影响;采用Kalman滤波进行重点采样,减缓粒子退化;采用单差无电离层固定模糊度,减少状态参数维数,进而减少粒子的选取个数。实测数据结果表明,粒子滤波有效控制了观测异常影响,提高了动态精密单点定位的精度。

The precision of dynamic precise point positioning using Kalman filtering will be degraded,even be divergent when the outliers exist.Particle filtering is applied to control the influences of the observational outliers,and improve the accuracy of positioning.Particle filtering is a kind of nonlinear filter with non-Gaussian distribution,and it can obtain accurate parameters by random sample.The weight of each particle is defined based on the probability densities of the observational errors,predicted state errors as well as the important distribution in order to control the influences of contaminated particles to the positioning results.Kalman filtering is employed to get the important sampling to slow down the degeneracy of the particle.The free-ionosphere ambiguities are fixed before data processing to reduce the number of parameters in the state vector.An actual dynamic GPS data set is employed to test the particle filter procedure.The procedure of the particle filtering can efficiently control the influences of the observational outliers,and improve the accuracy of the dynamic precise point positioning.