Font Size: a A A

The Research Of Malicious Node Detection In WSN

Posted on:2017-12-30Degree:MasterType:Thesis
Country:ChinaCandidate:F RaoFull Text:PDF
GTID:2428330485474504Subject:Information security
Abstract/Summary:PDF Full Text Request
Because wireless sensor network(WSN)is deployed in complicated network environment,which is easily being attacked by different kinds of network intrusion.The attackers analyze the feature information and network communication protocol stack by intercepting the data packets between the communication of nodes,which has done great harm on the network.In recent years,many scholars in the world have made some progress on the research of security in WSN,but these researches didn't form a complete set of network protection mechanism.So,this paper forms a complete set of network protection mechanism: energy consumption saving,authentication of nodes and detection of intrusion nodes,apply these theoretical research results into the smart power.The contributions of this paper are:First,this paper put forward a algorithm of target location selective movement,this algorithm firstly using the principle of k-means to cluster.And then select the center of each cluster node as the Sink,which can be placed in the target location,and then use the ideas of continuous Hopfie ld neural network algorithm to optimal the path of a preset advancing Sink node,the routing algorithm not only can effectively suppress the phenomenon of empty energy but also can solve the energy problem;Secondly,on the basis of analysing the network intrusion,this paper put forward a new communication model.This mode can effectively solve the mutual authentication among nodes and nodes,which can prevent transmission disguise malicious node information intercepted;Then,in the analysis and comparison of several advantages and disadvantages of the malicious node detection algorithm,we propose a method a Adaboosting algorithm,this algorithm first use Adaboosting algorithm to do the training,then to do test,to malicious nodes marked with labels,enabling isolation of malicious nodes;Finally,in order to better integrate theory with practice,this paper theoretical results applied to the message type photovoltaic power plants,w ith running,which proved the theoretical results have real economic value.
Keywords/Search Tags:wireless sensor networks, detect malicious nodes, energy consumption, node authentication, smart power plant
PDF Full Text Request
Related items