基于模型概率的多模型融合定轨建模及仿真

针对实际定轨系统中存在的不确定性和非线性性,提出了一种基于模型概率的多模型融合定轨方法.通过多个线性模型的组合来逼近卫星定轨复杂非线性时变过程,将卫星状态的最优估计与多模型融合方法相结合,利用残差的大小来设计性能指标函数,给出了两种模型概率的表示形式,并建立了多模型融合估计相应的算法.仿真结果表明,与单一模型定轨方法相比,该方法不仅能大大提高卫星定轨精度和可靠性,而且还可以最终辨识和估计模型参数的真值,且对外界环境发生的变化有很强的自适应能力.

Modeling and Simulation on Multiple Model Fusion for Satellite Orbit Determination Based on Model Probability

A multiple model fusion method for satellite orbit determination based on model probability was researched. When the parameters or models in practical satellite orbit determination system are uncertain and nonlinear,the method used is to combine multiple linear models to approximate the complicated nonlinear time-change process,and then integrates optimal estimation with multiple model fusion method. Besides,a performance index function based on model observation residual was designed,and two kinds of expression modes for model probability were advanced. Moreover,the relevant estimation algorithm for multiple model fusion was also established. Simulation results show that as compared with single-model algorithm,this strategy not only can improve accuracy and reliability of satellite orbit determination greatly,but also can identify and estimate the model parameters in practical system,and has much stronger self-adaptability for external factor change.