Predictive Data Transportation over Low-Altitude UAV Networks with Time-Varying Topology: A Dynamic Graph Approach

The rapidly growing human and autonomous robot activities in the low-altitude airspace brings new requirements and challenges to wireless networks. It is needed but challenging to design the aerial transmission strategy to ensure the transmission requirements of aerial nodes and avoid interference with ground terminals in a dynamic network topology. In this work, we study a predictive data transportation problem in aerial networks and formulate a integer programming problem to control the transmission route, handover time, and power. Specifically, a cache-and-pass transmission graph is introduced to convert the integer programming problem to a shortest path problem. Then, a graph-based alternating optimization algorithm with convergence guarantee is developed by using shortest path, alternating optimization algorithms. Simulation verifies that the proposed scheme is able to adjust the transmission strategy according to the data requirements in time-varying topology, reaching almost global optimality, and achieves an order of magnitude improvement compared with the classical static scheme, trivial routing scheme, and space-time routing scheme.