Research On Intrusion Detection Method Of Wireless Sensor Network In Advanced Metering Infrastructure Based On Trust Management

With the continuous occurrence of smart grid security incidents,the information security technology of Advanced Measurement System(AMI)has attracted wide attention.Researchers have made a thorough and detailed study of the internal node attacks in AMI system,but hardly detected the external node attacks.They only use signature,authentication and encryption to passive defense.Aiming at the shortcomings of the external node intrusion detection method in AMI system,such as large computation,incomplete detection and environmental interference,this paper proposes a feasible external node intrusion detection scheme by means of signature and authentication combined with trust management mechanism.Firstly,on the basis of defining the security requirements of the wireless communication network(WSN)of AMI,an abnormal message recognition method is designed.The advantages and disadvantages of the existing encryption algorithms are analyzed,and the Elliptic Curve Cryptography(ECC)with high security and practicability is selected.The Elliptic Curve Digital Signature Algorithm(ECDSA)algorithm is improved by using the Keyed-Hashing for Message Authentication(HMAC);an abnormal message recognition method is designed by checking the integrity of the communication message.Analyzed the performance of the method,the key management scheme is designed to realize the test of message recognition method.Then,according to the statistical results of communication messages by abnormal message recognition method,a trust value calculation method for WSN network nodes based on Beta distribution theory is designed.Based on the analysis of abnormal message recognition method and Beta distribution theory,a node trust value calculation method is designed according to the characteristics of WSN network communication.The smoothness and robustness of node trust value updating are improved.The regular anomalous behavior in the communication process is analyzed,so that the calculation method can quickly identify the regular anomalous behavior.Build the network simulation model and analyses the change of trust and the interference of environment factors in normal-abnormal message communication.Finally,according to the trust value of nodes,an intrusion detection method for external nodes of WSN communication network in AMI is designed.In this method,an early warning factor is introduced,which is based on fast update of trust value of nodes.analyzed the influence of environmental factors,and the regional anomaly detection based on improved D-S algorithm is introduced in the early warning stateto distinguish environmental factors from intrusion behavior.The confidence level threshold is obtained by the normal communication process data.A joint simulation model and a simple WSN communication network are built.It is used to operate and analyze the results.The simulation results show that the proposed scheme has good anti-jamming ability and superior detection performance.