基于小波多尺度分析的卫星钟差预报方法研究

Study on Prediction of Satellite Clock Error Based on Wavelet Multi-scale Analysis

利用IGS提供的精密钟差作为已知钟差序列,通过小波多尺度分析,将原始钟差序列分解到不同的时频层上,对每一层的高频系数和低频系数进行小波降噪处理,然后根据小波系数的特性,对不同层的小波系数采用不同的数学建模方法来预测,通过小波重构得到钟差预报序列,并将预报结果与二次多项式模型进行对比分析。实验结果表明,该方法具有较好的预报精度,可以为GPS实时精密单点定位提供较高精度的卫星钟差。

By using the wavelet multi-sale analysis,the original clock error which provided by the IGS is decomposed into the different time-frequency layers,and then the high-frequency coefficient and low-frequency coefficient in each layer were dealt with denosing process. The wavelet coefficient in different layers were predicted by different mathematics modeling method according to the characteristic of the wavelet coefficient. The clock error prediction sequence will be calculated through wavelet reconstruction and contrast with the quadratic polynomial model. The experimental result shows that our proposed method has better prediction accuracy,and it can provide clock error for the GPS real-time precise point position.