Research On SD-WSN Intrusion Detection Based On Cluster Analysis

Wireless sensor networks(WSNs)has many advantages such as self-organization,low cost,low power consumption,large-scale deployment and so on.It has great potential in many applications,such as military target tracking,natural disaster rescue,dangerous environment detection and so on.However,the inherent problems such as special design,difficult to change the behavior strategy after network deployment,and difficult to manage the network function make the research,design and practical application of wireless sensor networks face great challenges.Software defined wireless sensor network(SDWSN)is a new wireless sensor network architecture,which greatly improves the performance of traditional wireless sensor networks.However,due to the centralization of the control function,the new network simplifies security management,resulting in less resource constraints on the node itself and makes the network vulnerable to network attack.The security of software defined wireless sensor networks is still in its infancy,and not enough attention has been paid to it.Therefore,it is of great significance to study the security of software defined wireless sensor networks.Firstly,this thesis studies the security of the software-defined wireless sensor network.And found that the software-defined wireless sensor network is vulnerable to the security attack due to the lack of the main security components such as middle ware and transport-layer security(TLS).The controller as a single point of failure is also the most attacked component in the SDWSN.In order to solve the security problem of software defined wireless sensor networks,an improved artificial colony optimization K-means clustering algorithm is designed to reduce the influence of initial clustering center selection on clustering results.The KDD CUP99 data set is used to compare the algorithm with the traditional artificial bee colony optimization K-means algorithm and K-means algorithm.It is proved that the improved K-means algorithm is more effective than the traditional K-means algorithm and the traditional K-means algorithm.Finally,in order to achieve the purpose of intrusion detection.the thesis designs and implements the intrusion detection system,and describes in detail the implementation process of the system in the login interface,configuration file introduction,data collection module,data preprocessing module,data training module and intrusion detection module.