Research On Connectivity Of Mobile Wireless Ad Hoc Network Based On Virtual Backbone Nodes

The future battlefield network will be more and more heterogenous and intensive,which puts forward higher requirements for availability and security.Nodes constitutes the backbone network that serves the terminal,called “node + terminal” communication mode.This mode can better adapt to communication scenarios of large-scale and high-density deployment.However,due to the dynamic nature of network nodes,network connectivity is highly vulnerable,especially in unfamiliar high-dynamic combat scenarios.When links between backbone nodes are broken down,the tactical network confronts the risk of losing connectivity,which greatly restricts the application of the “node + terminal” communication mode in the battlefield.Considering that terminal nodes can also forward data,the terminal nodes can constitute mobile wireless ad hoc network and be used to recover the links between backbone nodes when backbone nodes are severely destroyed,which could quickly restore network connectivity and improve battlefield adaptability of “node + terminal” communication mode without increasing the cost.But there are some serious challenges in application of mobile wireless ad hoc network.Massive information is forwarded in high-density network,which would dramatically increase network load.Free motion of nodes could cause that links are vulnerable.This paper makes use of virtual backbone nodes to restore links,adopts cascade movement to maintain the network connectivity and implements the goal of restoring the connectivity of mobile wireless ad hoc networks under dynamic conditions.The main work and contributions are summarized as follows:(1)We propose a strategy to restore the backbone network using non-backbone nodes.All nodes in mobile wireless ad hoc networks can forward data,which is research foundation of this thesis.The key target of our strategy is to reduce the number of nodes participating the restoration progress.If there are too many non-backbone nodes forwarding data,“broadcasting storm” would easily happen.Firstly,member nodes calculate its own rank,according to the information of two-hop neighbors.Secondly,we construct a strongly connected dominating and absorbent set(SCDAS),utilizing the rank of nodes as its priority.It can choose virtual backbone nodes and reduce the number of nodes participating the restoration progress.Finally,we could achieve the optimal path with minimum hops among several candidate paths by allocating backoff delay for each chosen node.This strategy does not need global information,which is applied to mobile networks.The simulation result shows that the node set is less and can meet the requirement of restoration progress.(2)We present an algorithm which can dynamically maintain the connectivity of networks.The broken links between backbone nodes can be recovered with the help of non-backbone nodes only if the network is strongly connected.The nodes are mobile,and it is difficult to maintain the connectivity dynamically.In this thesis,we adopt protection and restoration technology to ensure that the network is strongly connected when nodes are moving.Firstly,we establish redundant multi-hop links between backbone nodes.When the direct links interrupt,data can be switched to alternate links.Secondly,we predict the movement of nodes so as to identify connectivity of networks with the adoption of Kalman filter.Finally,cascade movement is applied to maintain or restore the broken links.Cascade movement is the key step and has a great impact on the algorithm.It is the focus of the algorithm to reduce the number of chosen nodes.We propose three hybrid recovering algorithms,namely,cascade movement based on nearest nodes,cascade movement based on connected dominating set and cascade movement based on critical nodes.They all adopt neighbor information to judge whether a node participates in the cascade movement and the moving direction if the node needs to participate.We conduct extensive simulations,and the algorithm of cascade movement based on critical nodes achieves the best results.The number of chosen nodes is least and the impact on the task of networks is minimum.