Data sink selection using consensus leadership: improving target connectivity for a spacecraft constellation

The identification of optimal data sink nodes, which provide high data flow from all source nodes, is presented for a spacecraft constellation network. Selection relies on consensus dynamics, revealing how effective sinks in a data flow system equate to effective leaders in a networked system seeking a new consensus state. This has advantages over maximum flow analysis, including the assurance that connections to all source nodes affect the optimisation’s objective function. This is demonstrated for a system composed of ground stations, satellites, and ground-targets, which are modelled as a network of weighted connections varying according to simulated contact times. The ground stations are data sinks, to which the spacecraft constellation downlink information after receiving data from overflying ground-targets that are sources. A model of the Spire Global, Inc. constellation of 84 satellites with 250 ground-targets - distributed throughout major shipping channels - is used to test combinations of ground station selections from 77 potential sites. Consensus leadership is compared with maximum flow, revealing that consensus-based selections notably improve the data delivered from the least-connected sources at a relatively small cost in terms of total data downlinked. Numerical simulations confirm the optimality of selections for small combinations of ground stations.