Performance analysis of geostationary orbit satellite autonomous orbit determination algorithm based on few Global Navigation Satellite System satellites

Abstract This study investigates autonomous navigation solutions for geostationary orbit satellites under the condition of insufficient Global Navigation Satellite System (GNSS) visible satellites, such that conventional geometric positioning becomes unavailable. An autonomous orbit determination algorithm based on the extended Kalman filter is proposed, and the usability and positioning accuracy of the GNSS pseudorange filtering algorithm under the condition of few navigation satellites are analyzed. The influence of orbital dynamics model accuracy, thrust error, inner system bias (ISB) deviation, and carrier-phase smoothing pseudorange on the performance of the filtering algorithm is studied. A simulation scenario of GNSS signals received by a high-orbit satellite is constructed for verification. The effectiveness of the algorithm is evaluated through a semi-physical simulation validation platform. The simulation results demonstrate that with a precise orbital dynamics model incorporating both J2 perturbation and third-body effects, the algorithm maintains approximately 10 m positioning accuracy when continuously tracking three satellites, and meanwhile supports temporary operation with fewer than three satellites. During orbital maneuvers with 10% constant thrust bias and 6% random error, the solution sustains 60 m accuracy. In multi-system fusion positioning, the algorithm is still adaptable to the ISB deviation of ±15 ns. Through carrier-phase smoothing pseudorange technology, the improvement of positioning accuracy is limited. Further analysis will be conducted after the telemetry data is transmitted back.