灰色粒子群自适应卫星钟差预报方法

Particle swarm adaptive satellite clock error prediction model based on grey theory

高精度卫星钟差预报是当前接收机实时精密单点定位技术(Real-time precise point positioning,RT-PPP)亟需解决的关键技术难题之一.为找到一种基于小样本钟差序列的快速高精度卫星钟差预报方法,在分析常规GM(1,1)灰色模型(grey model)缺点的基础上对其进行了改进,提出了PGM(1,1)模型(particle swarm optimization-grey model)及其算法.该模型利用最新量测值进行初始化,然后通过引入遗忘因子的最小二乘法对新旧信息进行加权处理;再引入优化因子对模型系数进行调节,以归一化的平均相对误差作为精度检验标准,采用粒子群算法对其自适应寻优.最后选取了5颗钟差变化典型的GPS(global positioning system)卫星原子钟进行1 d内的精密钟差预报实验.结果表明,相对于常规GM(1,1)灰色模型和常规二次项拟合模型,所提出的模型及其算法预报精度有显著提升,其平均预报残差达到了亚纳秒级,且所需训练样本小.因此,该预报模型可以应用于卫星钟差快速精准预报.

The high precision satellite clock error prediction is one of the key technical problems for the receiver real-time precision single point positioning technology. To find a rapid and accurate prediction method for small sample satellite clock error sequences,a optimized algorithm model PGM（1,1） is presented based on the drawback analysis of the conventional GM（1,1） predication model. The predication model using the latest measurement for initialization is established, followed by replacing the old information with the latest one to realize model predication. In addition,the attenuated memory recursive least squares method is adopted for weighted handling of both the old and new information. The normalized mean relative error is used as accuracy test standard for fitting coefficient optimization factors and particle swarm optimization adaptive optimization is adopted. The typical clock errors error of five GPS satellites are predicted among one day using the PGM（1,1） model. The prediction accuracy is greatly improved with small training samples compared with the GM（1,1） model and the second order polynomial model,which indicates that the prediction method can be applied to the accurate and rapid forecasting of satellite clock error.