Research On Wireless Communication Link Intrusion Detection And Repair Method Based On Machine Learning Method

With the continuous development of the Internet and the Internet of Things,the ecological network derived from the interconnection of everything has been widely used in our daily life and work production.In the current large-scale complex network,there are malicious people who are wirelessly sensing the underlying Internet of Things.The network's attack behavior and the phenomenon that the damaged wireless communication link needs to be repaired.It is important to judge the intrusion behavior based on the current network traffic data and select the optimal link to be repaired according to the network node status.Due to the limited energy of sensor nodes and insufficient computing power in wireless sensor networks,when the network traffic data dimension becomes higher and higher,network anomaly intrusion detection will become more and more difficult.There are many factors that affect the stability of communication links in the IoT application environment,and each impact factor has a different proportion.It is difficult to select the best one by using a single impact factor to measure the quality of a link to be repaired.Based on the above problems,this paper uses the restricted Boltzmann machine algorithm to extract and reduce the dimensionality of network traffic data features in the wireless network,and uses the multi-layer support vector machine algorithm to classify abnormal data to achieve intrusion behavior detection.The fuzzy control algorithm brings the network node state information into the fuzzy control rules to calculate the link state value,and chooses the best link to be repaired.The main work of this article is as follows:(1)In view of the fact that wireless sensor networks are vulnerable to various network attacks,and the network nodes have limited energy and insufficient computing power,it is difficult to analyze abnormal network behavior under high-dimensional traffic data transmission.Detect wireless sensor network attacks.First,the RBM model is used to extract and reduce the dimensions of high-dimensional traffic data features,and the WSN network cluster structure is used to establish a multi-layer SVM classification to detect network anomalies.This paper uses the NSL_KDD public intrusion detection data set to verify the proposed intrusion detection model.The test results show that the model has an accuracy rate of 99.06% for network traffic detection.(2)In view of the problem that the wireless communication network composed of wireless sensor nodes has communication link damage,it is necessary to select the optimallink to be repaired among multiple links to be repaired,this paper proposes a link stability algorithm based on fuzzy control decision Repair strategy,extract the influence factors that affect the stability of the wireless communication link in the network nodes,quantify the influence factors into the fuzzy control decision model,establish the membership function of the input influence factor and the output decision value,establish the appropriate rule base,through fuzzy inference and Inverse fuzzification obtains the decision value of the node in the link,and repairs the link with the best stability and longer life cycle in the link.This paper uses matlab simulation software to simulate the wireless communication link,and uses simulink to establish a fuzzy controller to prove the feasibility of this method.