基于逐步回归的模型误差估计及在天基测控系统中的应用研究

Model Error Estimation Based on Distributing Regression and Its Application Research on Space Based Tracking Telemetry and Command System

以双星定位系统的天基测控技术为应用背景,提出了一种能够自适应估计模型误差的轨道确定方法。详细推导了观测模型中的系统误差形态,建立了能表征实际特征的部分线性轨道改进模型,并利用二阶段法和核函数估计法对混合误差进行补偿,在此基础上对补偿模型进行逐步回归分析,从中提取动力学模型误差,从而抑制了动力学模型误差的影响,提高了轨道改进的精度。在本文的仿真环境下,部分线性轨道改进法能够有效抑制混合误差对定轨精度的影响,提高天基测控的轨道确定精度。

Take the BeiDou navigation system as the application background, a new orbit determination method with model error adaptive estimation was proposed. Because of the adverse spaee environments, the space bases and the LEO satellite dynamic model is imprecise, while the error of dynamic model and error of the space bases will transit to observation model to form the new mixed error with observation error itself. The partial linear model of orbit determination was proposed as well as algorithm of kernel function for this model. It was been proved that the partial linear model of orbit determination can describe the true observation of space based SIT＆C, therefore, the precision of orbit determination will be. increased on the condition of simulation environments in this paper.