抗差高斯-雅克比组合平差法
Robust Gauss-Jacobi Combinatorial Adjustment
-
摘要: 基于粗差出现的频率信息,建立了观测污染分布模型和粗差数目二项分布模型,讨论了超定方程组基础解粗差分群的性质。通过对基础解解集进行聚类分析,提取基础解解集的零粗差分群,并提出了抗差高斯雅克比组合平差法。以GPS伪距单点定位为例,利用观测方程基础解的零粗差分群进行了抗差高斯雅可比组合平差。算例表明,即使杠杆观测为粗差观测,抗差高斯雅可比组合平差法仍具有高效性和稳健性。Abstract: Traditionally,theprocessofgrosserrordetectioniscloselyassociatedwithparameteresti mator.Unreliableinitialparametersmaycausetheprocessofgrosserrordetectiontofail,andviceversa.Inthispaper,themixednormaldistributionofobservationandthebinomialdistributionofgrosserrornumberareintroducedbyusingtheprioroccurrencefrequencyofgrosserror;Thepaperdiscussesthepropertiesofgross errorclusterscomposedofbasicsolutions,andfurtherproposearo bustGauss Jacobicombinatorialadjustmentmethodonthezero gross errorcluster.Atlast,thepro posedmethodisappliedtoGPSpseudo rangingpositioning.showthattheproposedmethodisrobustandhighefficientevengrosserroroccursonleverageobservation.