GPS近实时共视观测资料处理算法研究

On the Algorithm of Processing Observation Datum of GPS Common-view in Near Real-time

GPS共视资料的高精度快速处理可实现近实时共视时间传递,常见的平滑方法不能满足近实时并视的要求。分析GPS共视资料特点,设计一种卡尔曼滤波算法,对共视资料进行近实时处理,以便削弱观测噪声,估计异地钟差。对相距2000多公里的中国科学院国家授时中心(NTSC)与日本通信综合研究所(CRL),和相距1000多公里的CRL与韩国计量科学研究院(KRIS)的共视观测资料处理结果表明:卡尔曼滤波算法所得钟差与根据BIPMT公报所得钟差的均方根误差分别优于2.9ns和2.6ns。为进一步提高比对精度,最后对近实时共视应用于多站点间相互比对的情况,提出在卡尔曼滤波算法基础上使用间接观测平差处理技术,根据共视网络中站点间距离设置观测权值,通过解矛盾方程组得到两站钟差。以NTSC、CRL和KRIS3站比对为例,以BIPMT公报得到的钟差为标准,对间接观测平差处理前后的数据比较表明,近实时比对精度可进一步提高。

GPS common-view time transfer in near real-time may be implemented by processing observation data with high speed and high precision, but normal smoothing and filtering method is not satisfied with near real-time case. Based on the datum from GPS common-view, a Kalman filtering algorithm is presented, in which the clock difference at remote different places can be obtained. The reductions of the common-view data from National Time Service Center(NTSC) and Communications Research Laboratory(CRL) which locate at a distance of about 2000 km, and that from CRL and the Korea Research Institute of Standards and Science(KRIS) which locate at a distance of about 1000km show that the root mean square of time differences obtained by Kalman filter, and from Circular T are better than 2.9ns and 2.6ns respectively. Furthermore, when GPS CV in near real-time is applied in multi-sites, in order to improve precision, a technique of indirect observation adjustment after Kalman filtering algorithm is presented, in which the observation weight is set according to the distance between two sites. With reference to the clock difference from Circular T of BIPM, a comparison between data before and after adjustment shows that common-view after adjustment in near real-time is more accurate.