A comparative analysis of navigation signals in BDS-2 and BDS-3 using zero-baseline experiments

The Chinese Beidou navigation satellite system (BDS) has provided regional and global navigation and positioning services for the users via the BDS-2 and BDS-3 satellites. However, BDS-2 and BDS-3 send navigation signals on different frequencies; three of those are identical to assure system compatibility and interoperability. To comparatively analyze these navigation signals, we carried out three zero-baseline experiments with five brands of receivers, including Trimble, Septentrio, NovAtel, ComNav and Unicore. BDS-2 and BDS-3 multi-frequency data were collected during these experiments. The data were processed to evaluate the code and carrier phase measurement noises and investigate the inter-system biases (ISBs) on the three identical frequencies. The effects of the ISB on ambiguity resolution (AR) and position estimation (PE) were also demonstrated using the zero-baseline data with one Trimble receiver and one Septentrio receiver. The results show some new findings: (1) the code and phase noises in BDS-3 are smaller than those in BDS-2 on the same frequencies, and the three new signals B1C, B2b and B2a have comparable noises to the present signals. Besides, millimeter-level code noises are achieved on B3I and B2a signals in NovAtel receivers; (2) the code ISBs on B1I, B2I/B2b and B3I between BDS-2 and BDS-3 were found and distinguished, while no phase ISB could be found. The code ISB seems to be receiver-related and can be as large as 1 m in heterogeneous receivers. It is stable during a whole day but may also vary due to a restart of the receiver. No code and phase ISB on all three frequencies is observed for homogeneous receivers; (3) the code ISB will hamper the reliable AR and precise PE. The HMW (Hatch-Melbourne-Wubbena) combination on B1I and B3I is severely biased by about 0.6 cycles when double differencing between BDS-2 and BDS-3. After ISB calibration, the ratio value of wide-lane (WL) AR has a slight increase, and the accuracy of DD B1I code positioning has improved more than 10%, but few improvements can be seen in the fixed solution. It should be noted that although the inter-system code bias (ISCB) between BDS-2 and BDS-3 is detected in the baseline data, this bias should be considered in the absolute positioning as well.