非差非组合与组合模型定位性能对比及并行设计

分别采用非差无电离层组合模型与非差非组合模型进行批量静态定位解算,对比两者的定位精度以及ZPD估计精度。设计并实现两种模型的并行解算方法,提高非差模型的计算效率。大网数据实验结果表明,两种模型定位精度基本一致,非差非组合ENU的3个方向平均偏差为(4.2,2.8,6.6)mm,组合模型的平均偏差为(4.0,2.8,6.3)mm,但非差非组合模型的解算更加耗时,约是组合模型计算时间的1.4倍。多核环境下,两种模型的并行解算效率均比传统串行模型得到了提高。结果表明,双核并行和四核并行的计算效率比单核串行计算均分别提高了40%和60%以上。

Performance Comparison and Parallel Computing Between Ionosphere-free Combined and Uncombined Precise Point Positioning

In order to analyze the difference of traditional ionosphere-free precise point positioning (PPP) model and uncombined PPP model,observations are used to compare the static positioning performance whit respect to positioning accuracy between two PPP models.The parallel computing model of two PPP model is further designed and implemented to effectively improve the computation efficiency of un-differential processing strategy for data of huge GNSS network.The experimental results of large data test show that the positioning accuracy for both combined and uncombined PPP model is not significantly different.Compare with the IGS daily products,the average deviation of daily static positioning accuracy of uncombined PPP is (4.2,2.8,4.2) mm,and the average deviation of combined PPP is (4.0,2.8,4.0) mm.However,the calculating time of uncombined PPP model is about 1.4 times than the combined PPP model.In the multi-core computing environment,the efficiency of two PPP parallel computing models is improved higher than the traditional serial computing models.The results show that compared to the single core serial processing,the computational efficiency of dual-core parallel and quadcore parallel are improved approximately 40% and 62% respectively.