Objectives BeiDou-3 satellite navigation system (BDS-3) has been officially completed and provides services, network real time kinematic(RTK) is the main means to improve positioning accuracy, we present a long-distance undifference network RTK method based on BDS multi frequency and multi system observation data.
Methods First, the ambiguity resolution model considering atmospheric error parameters is used to determine the integer ambiguity of multi-frequency carrier phase of long-distance BeiDou satellite navigation system (BDS) reference station.Then, the double difference integer ambiguity between reference stations is converted into non-difference integer ambiguity of each reference station through linear change, and calculate the undifference observation error of each reference station.Considering the weakening of spatial correlation of observation errors between long-distance stations.According to the difference of error characteristics, obtain the correction number of classified undifference error of reference station network.Finally, by inverse distance weighted interpolation method, we calculate the error correction of mobile station, high precision positioning for users.
Results Select the measured data of long-distance continuous operation reference station network for experiment, the results show that: Compared with BeiDou-2 satellite navigation system (BDS-2), BDS-3 increases by 15.7%, 28% and 11.9% in E, N and U directions respectively. Compared with GPS, the positioning performance is equivalent or even better. The ambiguity fixing success rate and convergence time of multi-frequency data are better than GPS dual-frequency data. Dual system positioning results are better than single system, both of which can achieve centimeter level positioning of mobile stations.
Conclusions This method can avoid the correlation effect of double difference observations, high efficiency of spatial error calculation, mobile station positioning is more flexible, shows the advantages of BDS multi frequency observation data.