Study On Coding-aware Capacity Optimization Methods For Dynamic Space Networks

Space information networks have a multi-layer structure with long time delay and a highly dynamic topology, but the transmission capacity is highly restricted by the limited space resources such as frequency spectrum, orbital position and satellite power. Different from the traditional store-and-forward transmitting method, the network coding theory can reduce transmission costs and increase network throughput by combining the packets received at relay nodes. Therefore, communication links between network nodes with potential coding opportunities should be given priority for routing strategy and resource allocation, which is called a coding-aware network design method. Exploiting the coding-aware mechanism for data transmission and processing in space information networks, the transmission capacity can be improved qualitatively with an acceptable level of coding complexity by utilizing coding-aware network optimization methods, while the number of network nodes, link bandwidth and repeater power remain the same. Considering the motion laws of space nodes and the characteristic of space links, optimization methods to improve the transmission capacity for space information networks are studied and explored under the framework of network coding technology in this thesis, where an analytical model of transmission capacity and a coding-aware routing method for dynamic space networks are emphatically studied.Firstly, the maximum-flow problem in the graph theory is proposed to establish a mathematical model to describe the capacity of space networks. Then a hybrid MEO/GEO constellation is designed as an analytical model of the issue and a solution to the maximum-flow problem for the single-source multi-sink directed graph is studied based on the Ford-Fulkerson method. Furthermore, factors affecting the transmitting ability of space information networks are discussed in a specific scenario and the relations between network capacity and the capacity of user data links, minimum ground elevation constraint and satellite onboard processing capability are simulated and analyzed specifically.Secondly, the necessity and feasibility of applying the network coding theory in satellite networks are illustrated from the aspect of capacity optimization. Simulation results show that the network transmission capacity can be effectively enhanced when coding and decoding methods of random network coding and Gaussian elimination are employed in space nodes, which lays a theoretical foundation for the following parts.Finally, considering the features of space information transmission, a coding-aware routing scheme for dynamic space networks NC-MCF is proposed on the basis of the minimum-cost maximum flow problem in the graph theory and the coding-aware transmission mechanism. Moreover, a self-adaption bandwidth allocation scheme is applied to effectively raise resource utilization for the further improvement of the network transmission capacity. Additionally, the changes of key indicators of network capacity with time and orbital altitude are studied. By comparing the shortest path method, the maximum flow method and NC-MCF, simulation results show that NC-MCF excels in data transmission efficiency and bandwidth usage efficiency.