基于最小二乘和卡尔曼滤波方法进行原子时预报的研究

介绍和分析了最小二乘和卡尔曼滤波方法在时间预报中的应用。通过IGS站提供的钟差数据，分别运用这两种方法对其中六个原子钟进行了时间预报的实验。通过对预报结果进行分析，结论证明为了取得较好的预报效果，不同的预报方法和钟参数的模型对于观测数据的要求有一些差别。从实验数据所成图形来看，当采用一天的观测数据进行模型预报时，最小二乘法的预报精度比卡尔曼滤波法稍高一些。

The Study of Time Prediction Based on the Methodsof Least Square and Kalman Filter

This paper introduces and analyzes the application of the least square method and Kalman filter in time prediction.Through the clock bias data offered by IGS station,six atomic clocks are under the experiment of time prediction using these two methods.The analysis of prediction results shows that there are differences in the requirements of the observation data for different prediction methods and clock deviation models in order to obtain a better prediction result.According to the figures of the experimental data,we can deduce that the prediction precision of least square method is higher than Kalman filter while adopting one day's observation data in the prediction of the clock deviation model.