基于半参数回归的联合定轨误差估计仿真算法

Combined Orbit Determination Error Estimation Simulation Algorithm Based on Semi-parametric Regression

通过对联合定轨非线性影响因素特征的分析,提出了一种参数建模与非参数分量表示相结合的联合定轨非线性半参数回归建模方法。首先讨论了联合定轨模型及常值系统误差估计算法,在此基础上建立了一种广义非线性半参数回归联合定轨模型,设计了相应的卫星轨道参数、系统误差参数及模型误差估计算法,并从理论上证明了半参数模型轨道估计精度优于经典最小二乘轨道估计精度。仿真计算结果表明,广义非线性半参数回归联合估计方法能将模型误差和随机误差有效分离,同时联合定轨精度也得到了进一步的改善,从实际应用角度验证了半参数回归联合定轨建模方法的合理性和可行性。

A modeling method of combined orbit determination （COD） nonlinear semi-parametric regression which associated parametric modeling with non-parametric components expression was put forward by analyzing the characteristics of COD nonlinear influencing factors. The COD model and constant system errors estimation algorithm were discussed firstly, and then a generalized nonlinear semi-parametric regression COD model was constructed and estimation algorithm of satellites trajectories parameters, system errors parameters and models errors was designed. Finally the conclusion that semi-parametric model estimation precision was superior to classical LS orbit estimation precision was proved from theoretical angle. Simulation computation results show that the method of generalized nonlinear semi-parametric regression COD estimation can separate models errors from stochastic models effectively, and ameliorate COD precision ulteriorly, which validates rationality and feasibility of semi-parametric regression modeling method from practical angle.