Parameter decomposition filter of BDS-3 combined orbit determination using inter-satellite link observations

Abstract Starting in March 2015, 5 new-generation BDS satellites, including 2 satellites in IGSO and 3 satellites in MEO, have been launched with the Inter Satellite Links (ISLs) on-board and the measurements have been tested. With the ISLs, BDS can achieve Autonomous Navigation (AutoNav) without support of ground facilities. In AutoNav, for lack of the external datum, the residual constellation rotation cannot be observed. This problem will be solved with the support of ground facilities. When ground facilities are available, Combined Orbit Determination (COD) can be achieved by using measurements from the ISLs and ground facilities, which can dramatically improve the precision of orbit determination. Normally, there are two kinds of filters adopted in AutoNav, namely the Parameter Integration Filter (PIF) and the Parameter Decomposition Filter (PDF). The PIF, which is also commonly used in COD due to the one-way property of the L-band satellite-ground measurements, had been prove less efficient than the PDF in AutoNav. As an attempt of improving the solving efficiency for COD, a new PDF is presented and implemented in COD of BDS in this paper. Based on the implement of the COD with simulated observations, the performance of PDF in COD is evaluated by comparing to the PIF in regards to precision and solving efficiency. The results indicate that, given enough observations, the PDF is superior to the PIF in terms of CERR and URE, while inferior to PIF in terms of RERR and PERR. Moreover, the solving efficiency of the PDF is much higher than that of the PIF.