High-Accuracy Clock Synchronization Algorithm for LEO Satellites Based on Particle-Enhanced Extended Kalman Filtering

Inter-satellite clock synchronization is critical for maintaining consistent timing across low Earth orbit (LEO) constellations. Conventional Kalman-filter-based methods often ignore frequency shift information and may converge to local optima, limiting accuracy. To address this, this paper proposes a particle-enhanced extended Kalman filter (PE-EKF) that integrates two-way timing with carrier frequency observations. An observation model is first constructed from two-way measurements, and initial estimates of clock skew and offset are obtained via an extended Kalman filter (EKF) algorithm. A particle evolution mechanism then refines these estimates by generating a feasible region around the EKF output, followed by sampling and weighted fusion. Simulations show that PE-EKF improves skew estimation accuracy by 20.08% and 40.83% over F-EKF and ETV-KF, respectively, while reducing clock offset errors to 0.63%. These results confirm its effectiveness for high-accuracy synchronization in LEO networks.