The Research Of Intrusion Detection In Wireless Sensor Networks Based On Hidden Markov Model

With the continuous development ofInternet of Things, Wireless Sensor Networks(WSN), which is the basis of Internet of Things,whose security has become an important problemwhich is needed to besolved. In recent years, many scholars have done researches on the problem, and have givenmany solutions. Most of them focused on WSN’s protocols and framework. And they try to change the structure of the communication protocol or use cryptography to enhance the safety level of WSN.This paper presents a more active approach to protect WSN, Which introduces Intrusion Detection System in WSN. This paper designs an Intrusion Detection System which is based on Hidden Markov Model(HMM). First, this paper describes the characteristics of WSN and HMM and analyzes the advantages of using HMM. Then, this paper classifiesthe attacks of WSN and analyzes the characteristics of the attacks in order to determine the way of Intrusion Detection. Then, according to the HMM’s Forward-Backward algorithm, this paperpresents methodsandprocesses in detecting attacks based on HMM in WSN. Then, according to the characteristics ofZigBee protocol and ZigBee’s framework, this paper designs a method ofextracting frame sequence and determines the methodof building HMMwhich uses HMM’s Baum-Welch algorithm.Then, this paper designs a process of buildingHMM anddetecting intrusions. Finally, this paper does experiments to verify the value of the system.