Temporal correlation length of network based rtk techniques

Abstract

Temporal correlation in network based Real-Time Kinematic (NRTK) data exists due to the unmodeled multipath, tropospheric (especially for wet delay) and ionospheric variations, and slowly changing spatial geometry of satellites. Ignoring the temporal correlation can result in too much confidence being placed in the estimated uncertainty of the coordinates. Temporally correlated errors occur when the magnitude of an error is constant or similar over time. In this study, temporal correlation lengths of commonly used Virtual Reference Station (VRS), Flächen Korrektur Parameter (FKP) and Master Auxiliary Concept (MAC) NRTK techniques were computed for the different length of baselines between the rover and the master station (closest CORS station) in the Turkish GNSS active network (TUSAGA-Aktif). Thus, appropriate time separation between the measurements for revisits of the points was determined in order to avoid overly optimistic coordinate uncertainty and unreliable coordinate checking. The results show that there is no clear relationship between the correlation length and the baseline distance. Correlation lengths of FKP and VRS are similar, whereas MAC produced a much longer correlation time for the horizontal component. When the correlation length of each technique is averaged, the correlation lengths are estimated as 17–13–16 min for northing, easting projection coordinates and ellipsoidal height, respectively. According to the measurements using different geodetic-grade receivers, the results show that temporal correlation length varies between the receivers.