Research On Key Technologies Of Monitoring And Analyzing Tactical MANET

Tactical Mobile Ad Hoc Network(MANET)is an instant network composed of a group of autonomous devices or nodes in a battlefield environment.It has no infrastructure,easy deployment,no center,self-organization,high dynamics,etc.features.Monitoring and analysis of nodes and network attributes of tactical MANET play an important supporting role in the field of network security analysis and countermeasures.The traditional monitoring and analysis technology under the condition that the network can be accessed and the information content can be resolved must be based on access and decryption.The application scenarios are greatly restricted,especially the active detection type monitoring and analysis technology,which has large detection traffic and affects network operation performance and other issues are difficult to apply to scenarios such as network load-sensitive and non-cooperative monitoring.This article focuses on the tactical MANET non-cooperative and passive monitoring and analyzing problems under the condition of not accessing the target network and analyzing the information content,focusing on the key technologies of monitoring and analysis from four aspects: node localization,topology inference,backbone network identification and critical node identification.(1)Node localizationIn view of the limited detection range of traditional positioning methods and it is difficult to locate all tactical MANET communication nodes in the area,a solution is proposed to locate tactical MANET nodes in a specific area by spreading sensors as monitoring nodes.Firstly,based on the MDS-MAP(P)algorithm to complete the cooperative positioning between the monitoring nodes,and then use the target node location method based on Fibonacci Branch Search(FBS)proposed in this paper according to the results of the monitoring nodes’ perception of the presence of the target node’s signal.The node positioning method completes the target node positioning through optimization.Experimental results show that the proposed method has higher positioning accuracy than similar methods.(2)Topology inferenceOn the basis that the location information of the target node has been obtained,according to whether the signal information of the target node(signal sending time and signal duration)can be monitored,the topology inference methods under different conditions are studied.(1)For situations where signal information is available,a topology inference method based on the analysis of interframe space feature(IFSF)is proposed.First,the statistics of the labeled data frame interval are extracted as features,and then the K-Nearest Neighbor(KNN)algorithm is used to implement topology inference.Experimental results show that the inference accuracy of this method is significantly higher than similar methods.(2)For situations where signal information cannot be obtained,a topology inference method based on the analysis of critical transmitting range(CTR)is proposed.First,calculate the CTR based on the network connectivity rate,and then implement topology inference based on the CTR.Experimental results show that this method can complete topology inference in near real-time,and can effectively cope with the dynamic changes of topology.(3)Backbone network identificationBased on the results of network topology inference,and aiming at the real-time nature of network topology changes,the backbone network identification methods in different situations are studied.(1)Aiming at the quasi-static scene where the topology changes slowly,a backbone network identification method based on topological potential and unifying connected dominating set(TP-UCDS)is proposed.First,the topological potential of the node is obtained as the static dominance factor,and then the identification of the backbone network is completed according to the relevant rules of the unified connected dominance set.(2)Aiming at the high dynamic scene where the topology changes rapidly,a backbone network identification method based on dynamic dominating factor and unifying connected dominating set(DDF-UCDS)is proposed.First,the dynamic dominance factor is obtained by combining the time window and the attenuation coefficient to fuse the static dominance factors of the nodes at different times,and then combined with the unifying connected dominating set to realize the dynamic identification of the backbone network.Experimental results show that this method can not only effectively identify the relatively stable backbone network structure in the network when the tactical MANET is in a group mobile state,but also the identified backbone network has a longer life cycle and can effectively cope with the dynamic changes of the tactical MANET topology.(4)Critical node identificationBased on the identification of the backbone network,the critical node identification methods from different perspectives are studied for different network failure mechanisms.(1)From the perspective of network fragmentation,a critical node identification method based on Graph Signal Processing(GSP)is proposed.First,perform graph signal analysis on the candidate node set,and then use the smoothness of graph signal to complete the grouping of nodes,and finally obtain the minimum node cut set of the link interruption between different groups of nodes as the recognition result.Experimental results show that when the same number of nodes are removed,the critical nodes identified by this method are significantly more damaged than other methods after they are removed.(2)From the perspective of cascading failure,a critical node identification method based on Differential Load Balancing(DLB)is proposed.First,build a differentiated load-capacity model,then use the cascade potential as a measure to complete the preliminary screening,and finally combine the harmony search algorithm to complete the optimization of the identification results.Experimental results show that the recommended number of nodes removed by this method is less than other methods under the constraints of the established damage effect.