Research And Simulated Implementation Of Hierarchical Routing For Space/Air/Earth Information Networks Based On Data Mining

Space/air/earth communication systems are mutually independent of each other at present.Nodes in different kind of networks cannot realize direct information interactions,which results in a waste of resource and a lack of QoS support on various applications of aerospace communication platforms.Currently,few efficient prediction algorithms can work with the routing in the space/air/earth information network,and it has been a key problem to be solved to apply effective prediction algorithm in the space/air/earth information networks.Besides,the space/air/earth information network models also failed to show its link connection condition in time and space clearly.Such routing algorithms of space/air/earth information network will not be able to become an efficient routing,and the topology modeling of the space/air/earth information network is also key point.The space/air/earth information network is a kind of delay tolerant network(DTN).Recently,with possible participation of a large number of wireless nodes,how to maintain efficient and dynamic topology becomes crucial.In DTNs,the topology is lack of continuous connectivity,and has characteristics of network heterogeneity,long dealys,unreliable time-variable connections and highly dynamics,which make the design of routing algorithms face severe challenges.In addition,problem about DTN routing topology design is also influenced by its characteristic,and modeling for the space/air/earth information network is also a challenge.This paper proposes a time-space graph based on hierarchical routing scheme for the space/air/earth information network.To construct a semi-deterministic time-space graph,we introduce a data mining algorithm to predict the contact probability for each pair of nodes.Using the time-space graph model,we put forward a simplified topology and determine a path that ensures the connectivity and reliability over time but also significantly reduces the cost of topology,thus achieving the hierarchical routing.The aim of our topology design is to build a sparse time-space structure such that:1)for any pair of network nodes,there always exists a path with the reliability larger than the required threshold;2)the total cost of the topology is minimized.We first show that this problem is NP-hard,and then propose several heuristics which can significantly reduce the total cost of the topology while maintaining the reliable connectivity over time.Simulation results demonstrate that the proposed scheme could find a simplified topology which can ensure the connectivity and reliability of paths over time but also significantly reduce the cost of topology.