Strategy of Data Processing for GPS Rover and Reference Receivers Using Different Sampling Rates

In Global Positioning System (GPS) applications, the double-difference (DD) model is often employed, which requires that the reference station and the rover receivers adopt the same sampling rate. On the premise of the need fulfillment, could the existing static receivers of Global Navigation Satellite System networks with a low sampling rate be used as the references for the rover with a high sampling rate? If feasible, how to process such data? In this paper, an integrated processing scheme is presented to deal with such data in the posterior mode, i.e., after initialization, both the single-difference (SD) and the DD models will be employed when synchronous data are available, while only the SD model will be used when only the rover receiver data are obtained. The deviations of the results obtained with the SD model from those with the DD model at the synchronous epoch might be considered as the accumulated systematical errors, which could be linearly interpolated into individual nonsynchronous epochs of that interval. This strategy takes full advantage of all available observations, rather than discarding the nonsynchronous GPS data, and it provides a promising application for the precise kinematic point positioning. The total expense of the surveying mission will be greatly reduced by adopting the low-sampling-rate receivers of the Continuously Operating Reference Station or the other similar GPS network as the reference stations. A data set of an airplane flight experiment was processed to verify the new algorithm, and the results revealed that the positioning accuracy could reach a centimeter level.