基于非差非组合PPP-RTK的大气改正模型及其性能验证

高精度的大气改正是加快PPP-RTK收敛的重要前提。本文以区域跟踪网台站数据为基础，基于非差非组合PPP提取斜路径电离层和天顶对流层延迟，作为PPP-RTK大气建模的数据源。电离层延迟采用基于斜路径星间单差的改正模型，对流层采用非差天顶对流层模型，设计了相关的服务端和用户端软件系统。在系统设计上，通过服务端提取数据构建大气模型并播发，用户端接收参数并用于实时PPP-RTK定位。对上海区域进行服务端和用户端的试验，服务端计算的参数表明：GPS、GALILEO、BDS系统的电离层、对流层模型内符合精度为6~7 mm。用户端的646组PPP-RTK伪动态试验表明：水平方向30 s内收敛的占比为89.16%、1 min内收敛的占比为91.80%、2 min内收敛的占比为95.98%；三维方向收敛结果中，上述收敛时间尺度分别占总数的86.22%、88.70%和93.34%。附加大气约束后，模糊度固定率为95.59%，收敛后水平方向和三维方向定位RMSE分别为2.35和4.63 cm。实时动态试验表明，PPP首次固定时间为36 s，水平和三维定位精度分别达到了1.13和3.21 cm。

Modeling and assessment of regional atmospheric corrections based on undifferenced and uncombined PPP-RTK

High-precision atmospheric modification is an important prerequisite for accelerating the convergence of PPP-RTK. Based on regional tracking network stations, we first extract the slant ionospheric and zenith tropospheric delays based on the undifferential and uncombined PPP model. Then the estimated ionospheric and zenith tropospheric delays are used as the source observations for regional PPP-RTK atmospheric modeling. The slant ionosphere correction model is based on the single difference among satellites, while troposphere correction is modeled as undifferential zenith model. Based on the proposed method, prototype PPP-RTK server and client software systems are developed, where regional server generates and broadcasts atmospheric models and the user client realizes PPP-RTK with fast convergence by receiving the broadcast parameters from the server. Experiments using CORS network in Shanghai, the results on the server end show that both ionospheric and tropospheric model of GPS, GALILEO, and BDS systems achieve an accuracy of few mm. The 646 sets of PPP-RTK experiments on the client side show that: ①For horizontal coordinates, 89.16% of all solutions converged in 30 s, 91.80% in 1 minute, and 95.98% in 2 minutes. ②For the three-dimensional positions, 86.22% converged in 30 s, 88.70% in 1 minute and 93.34% in 2 minutes. With the addition of atmospheric constraints, the fixed rate of ambiguity is 95.59%, and the localized RMSE in the horizontal and three-dimensional directions after convergence is 2.35 and 4.63 cm, respectively. The real-time PPP-RTK test experiment shows similar results, where time to first fix (TTFF) is 36 seconds,accuracy of horizontal and three-dimensional coordinates reach 1.13 and 3.21 cm, respectively.