Multiple Chord Distance Regression Algorithm to Judge Constellations Without Prior Information

With the development of large-scale low earth orbit satellite constellation networks (LS-LEOSCN), it is key to make constellation judgment based on observation data. The problem lies in that the realistic constellation configuration (CC) may be different from declared. To solve this problem, a multiple chord distance regression (MCDR) algorithm is proposed to identify the realistic CC based on the Two-Line Elements (TLE) data. Experiments are done for both the constructed data and real TLE data, including the Satrlink, IridiumNext and OneWeb data. It turns out that these constellations can be well judged and given their maintenance parameters.