Research On Monitoring Data Oriented Intrusion Detection Algorithm For Wireless Sensor Networks

As wireless sensor networks play an increasingly important role in life and work,designing intrusion detection technologies for wireless sensor network characteristics is the focus of current network security research.Currently,most intrusion detection techniques are typically deployed on sensor nodes,which is a challenge for energy-limited sensor nodes.In this thesis,two intrusion detection algorithms with the monitoring data received by the base station are proposed,namely,the feature-adaptive wireless sensor networks intrusion detection algorithm and the parameter-adaptive wireless sensor networks intrusion detection algorithm.Aiming at the disadvantages of traditional centralized intrusion detection algorithms that require sensor nodes to transmit network features to base station,and intrusion detection algorithms that use monitoring data as detection features are difficult to identify denial of service attacks,the feature-adaptive wireless sensor network intrusion detection algorithm with monitoring data received by base station is proposed.First of all,the primary task of the algorithm is to design feature processing formula,so that the monitoring data can reflect the dynamic change of the network after being processed by the formula.It enables the algorithm to identify denial of service attacks without network features,reducing the node's energy consumption.At the same time,unsupervised learning is achieved by combining the advantages of different machine learning.Finally,the sliding window module is designed by using the characteristics of sensor monitoring data,which enables the intrusion detection algorithm to have the capability of feature adaptation.The experimental results show that the proposed algorithm has better detection accuracy than support vector machine,naive Bayes and decision tree algorithm.Aiming at the phenomenon that the support vector machine is sensitive to parameters,the parameter-adaptive wireless sensor networks intrusion detection algorithm is proposed.The algorithm uses the improved fruit fly optimization algorithm to achieve adaptive adjustment of parameters.In addition,the parameter-adaptive intrusion detection algorithm adopts the feature processing formula of the feature adaptive intrusion detection algorithm,which enables the algorithm to detect denial of service attacks and data pollution attacks without network features.The experimental results show that detection effect of the proposed algorithm is better than these algorithms that genetic algorithm,particle swarm optimization and standard fruit fly optimization algorithm respectively optimize support vector machine algorithm.