提出一种新的钟差算法——开窗分类因子抗差自适应序贯平差,即首先对一维钟差数据进行开窗处理,在窗口内利用抗差等价权削弱粗差影响,在窗口间构造自适应因子抵制钟跳异常,从而达到消除和削弱观测异常和状态异常的目的。针对不同星钟参数不符值描述不同的扰动特性,提出构造分类自适应因子来抵制钟差时间序列中的扰动异常。计算结果表明,新算法一方面引入抗差估计,控制了粗差影响,拟合精度和预报精度与没有进行抗差处理的自适应序贯平差相比,分别提高78.9%和60.4%;另一方面由于新算法构造分类自适应因子,分别处理不同特征的状态异常,钟差拟合精度和预报精度与单因子抗差自适应序贯平差相比,分别提高4.3%和29.2%。
It is proposed that a novel clock error estimation and prediction algorithm—adaptively robust sequential adjustment with classified adaptive factors based on opening windows.The main ideas are firstly one dimensional clock error series are processed to eliminate the gross errors with robust equivalence weight estimation in opening windows,where adaptive factors are used between windows to reduce the anomaly of clock errors.Secondly,as for the different satellite clocks,based on their different error characteristics,the classified adaptive factors are adapted to mitigate the effects form clock outliers.Tests indicate as for the fitting precision and prediction precision,the new proposed algorithm can improve 78.9% and 60.4% respectively by comparing with no robust estimation method.Besides,even for the robust estimation,classified factors method are better than single factor method,tests show about 4.3% and 29.2% improvement in fitting precision and prediction precision respectively.