Five new-generation BeiDou-3 experimental satellites, called BeiDou-3e, have been launched into inclined geosynchronous orbit (IGSO) and medium orbit (MEO) since March 2015. In addition to newly designed signals and intersatellite links, different satellite buses, updated rubidium atomic frequency standards (RAFSs), and new passive hydrogen masers (PHMs) have been used. Using 15 stations, mainly in the Asia–Pacific region, we determined orbits and clock for both the BeiDou-3e and the regional BeiDou-2 satellites using the Extend CODE (Center for Orbit Determination in Europe) Orbit Model (ECOM). The orbit consistency, indicated by 3D orbit boundary discontinuity, is 50–70 and 40–60 cm for BeiDou-3e IGSO and MEO satellites, respectively, and better than 15 cm in radial component. Satellite laser ranging (SLR) validation gives about 17 and 10 cm for BeiDou-3e IGSO and MEO satellites. The BeiDou-3e satellites orbits show slightly better performance than the BeiDou-2 satellites as indicated by SLR. However, errors depending on the sun elongation angle were identified in SLR residuals for the BeiDou-3e IGSO C32 satellite, while such errors did not exist for BeiDou-2 IGSO/MEO and BeiDou-3e MEO satellites. No orbit accuracy degeneration for BeiDou-3e IGSO and MEO satellites was observed when the elevation angle (β angle) of the sun above the orbital plane was between − 4° and + 4°. In that case, the BeiDou-2 IGSO and MEO satellites are in orbit normal (ON) mode. An analysis of the yaw attitude identified that BeiDou-3e satellites did not use the ON mode, but experienced midnight- and noon-point maneuvers when the β angle is approximately between − 3° and + 3°. Compared with BeiDou-2 satellites, the onboard clocks of the BeiDou-3e IGSO satellites showed dramatic improved performance. The stability of BeiDou-3e IGSO satellites can be compared to the latest type of RAFSs employed onboard the GPS IIF satellites as well as the PHMs used onboard the Galileo satellites.

By end‐2018, the Chinese BeiDou satellite system (BDS) has entered its third phase of global scale coverage (BDS‐3). This study presents and validates a modified modeling aspect concerning an intersatellite link (ISL) hardware delay estimation. The precise orbit determination (POD) results of BDS‐3 demonstrate this aspect along with calibrated phase center offsets (PCOs) of the ISL terminal to help improve the orbit accuracy. After onboarding the ISL data, two additional sets of POD solutions using regional and global ground stations are conducted. About a 40% improvement can be achieved for regional cases and 20% for global cases after ISL data were used for POD. The best POD performance solution is obtained with observations from both the global ground stations and ISL when the averaged 3D root mean squares of the orbit overlapping differences can reach around 20 cm and the signal‐in‐space range errors (SISREs) obtained by 6‐hour predicted clock and orbit solutions show an average value of 25 cm.

A key limitation for precise orbit determination of BeiDou satellites, particularly for satellites in geostationary orbit (GEO), is the relative weak geometry of ground stations. Fortunately, data from a low earth orbiting satellite with an onboard GNSS receiver can improve the geometry of GNSS orbit determination compared to using only ground data. The Chinese FengYun-3C (FY3C) satellite carries the GNSS Occultation Sounder equipment with both dual-frequency GPS (L1 and L2) and BeiDou (B1 and B2) tracking capacity. The satellite-induced variations in pseudoranges have been estimated from multipath observables using an elevation-dependent piece-wise linear model, in which the constant biases, i.e., ambiguities and hardware delays, have been removed. For IGSO and MEO satellites, these variations can be seen in onboard B1 and B2 code measurements with elevation above 40°. For GEO satellites, a different behavior has been observed for these signals. The GEO B2 pseudoranges variations are similar to those of IGSO satellites, but no elevation-dependent variations have been identified for GEO B1. A possible cause is contamination of the larger noise in GEO B1 signals. Two sets of precise orbits were determined for FY3C in March 2015 using onboard GPS-only data and onboard BeiDou-only data, respectively. The 3D RMS (Root Mean Square) of overlapping orbit differences (OODs) is 2.3 cm for GPS-only solution. The 3D RMS of orbit differences between BeiDou-only and GPS-only solutions is 15.8 cm. Also, precise orbits and clocks for BeiDou satellites were determined based on 97 global (termed GN) or 15 regional (termed RN) ground stations. Furthermore, also using FY3C onboard BeiDou data, two additional sets of BeiDou orbit and clock products are determined with the data from global (termed GW) or regional (termed RW) stations. In general, the OODs decrease for BeiDou satellites, particularly for GEO satellites, when the FY3C onboard BeiDou data are added. The 3D OODs reductions are 10.0 and 291.2 cm for GW and RW GEO solution with respect to GN and RN solution, respectively. Since the OODs in the along-track direction dominate the OODs reduction, no improvement has been observed by satellite laser ranging, which mainly validates the accuracy of the radial orbital component. With the GW BeiDou orbit and clock products, the FY3C orbits determined with onboard BeiDou-only data also show improvement in comparison with those determined with BeiDou GN products.