Velocity‐aware handover prediction in LEO satellite communication networks

Summary The previous satellite handover prediction and ground cellular network were mainly studied in the static steady satellite links. Rare work has been conducted to resolve the dynamic satellite handover prediction. The complexity of this problem comes from that the multiple link states of the satellite change with the users' movements. To reckon with these challenges, the time‐expanded graph, in which the covering satellite and the handover strategy are treated as its node and arc weight respectively, is constructed in this paper. And the handover prediction is viewed as finding the shortest path of the time‐expanded graph dynamically based on the previous state with the perception of velocity. Compared with the previous approaches in the satellite handover prediction, the proposed method can effectively avoid the handover prediction failure and reduce the redundant computation triggered by static update algorithms. Simulation results are presented to show the improvements.