An event-driven graph-based min-cost delivery algorithm in earth observation DTN networks

The Low Earth Orbit (LEO) satellites for Earth observation confront intermittent connection and high packet-loss rate, which significantly limits the returns of scientific data due to non-continual short-time windows. Disruption-Tolerant Networking (DTN) provides an alternative approach for networking for the Earth observation (EO) constellation, which could exploit adequately each short-time contact of individual LEO member. In this paper, a Minimum Cost Transmission Strategy (MCTS) is proposed to find an energy cost-optimal route with least power consumption and acceptable delivery delay. In the proposed MCTS, we establish an event-driven graph-based minimum cost flow model with temporal constraints, which could effectively feature the time-varying topology of EO constellation. In particular, a well-designed bundles fragmentation mechanism is provided for matching the time-windows capacity according to the resulted parameters of the model. The experimental results demonstrate the significant efficiency of delivery delay, power cost and time constraint satisfaction ratio of the proposed MCTS, compared with current contact graph routing-based delivery scheme.