Research On Intrusion Detection Algorithms For Wireless Sensor Network Based On Traffic Prediction

Wireless sensor network can form a multi-hop ad hoc network for its flexible configuration.Generally wireless sensor network is deployed in the complex and unattended environment,and its openness and broadcast characteristic raise the potential risks to the network security.With the expanding of application scope in wireless sensor network,the network security situation is more rigorous than before.Considering the security problem,the intrusion detection technology in wireless sensor network is deeply researched,and a double-layer analysis intrusion detection model based on traffic prediction is proposed.To make up for the disadvantages in the detection of attack with slow traffic affection using ARMA algorithm,a first layer detection algorithm based on the traffic space-time characteristics is proposed by combining with the time and space prediction technology.The trigger judgment is made through the first layer analysis result,and then,the second layer local analysis to the suspicious region is activated.Through such double layer analysis structure,the power consumption after importing security system is reduced,and the detection rate is increased.With the NS2 simulation software,we set up the simulation environment,and carry on simulation test.Test results show that this model can meet high detection rate and low false positive.Finally,based on the analysis model,we design an intrusion detection system in wireless sensor network,do the test in the actual network environment with actual sensor nodes,and verify the feasibility and effectiveness of the system.