Project_01
Beam Acquisition and Tracking for LEO Communication Systems: A Learning based Kalman Filter Approach
A method is proposed that combines the Kalman filter and long short-term memory model to estimate the position of a LEO satellite using two-line element and the global navigation satellite system information, aiming to improve satellite tracking accuracy. This method is applied to beam acquisition and tracking, effectively reducing beam search time and maximizing the signal-to-noise ratio. Experimental results demonstrate that the proposed approach outperforms traditional methods in satellite position and beam direction prediction accuracy while significantly reducing search time.