Clock Synchronization For Wireless Sensor Networks Under Malicious Attacks

Wireless sensor networks(WSNs),which integrate wireless communication,information processing and other emerging technologies,are receiving more and more attention due to its wide applications including traffic control,environmental monitoring,smart home and so on.In the meantime,their development is restricted by various factors,one of which is the clock synchronization.It is significant for all nodes in a WSN to work together to accomplish complex sensing tasks,which requires all nodes to keep the same time.However,as a common distributed system,it is difficult for all nodes in a WSN to reach an agreement on the time.Therefore,designing the clock synchronization protocols to realize the clock synchronization among nodes is the precondition for many applications of WSNs.In actually,WSNs are vulnerable to numerous network attacks,and the existence of malicious attacks brings great challenges to the realization of accurate clock synchronization.Thus,this paper studies the clock synchronization of WSNs under malicious attacks,and porposes the corresponding clock synchronization schemes.The main achievements are as follows:In order to improve the error accumulation of clock offset in the traditional consensus based clock synchronization algorithm,an improved clock synchronization scheme based on maximum consensus is proposed.In this paper,the clock offset compensation and the relative skew estimation are separated independently,and the relative offset estimation algorithm is proposed for the first time.On this basis,a novel software clock offset compensation algorithm is proposed,which ensures higher synchronization accuracy for WSNs in the presence of bounded transmission delay.Previous studies have shown that the traditional consensus based clock synchronization schemes can resist the impact of denial of service attacks,but can not resist the false data injection attacks.To this end,an attack detection mechanism is designed in this paper,which can accurately detect abnormal data attacked by false data injection.According to the results of the attack detection,a synchronization algorithm based on the maximum consensus is proposed in both delay-free and random-delay cases.Relevant proof and simulation are provided,which prove the security and reliability of the proposed protocol with the attack detection mechanism.In this paper,the clock synchronization problem under denial of service attacks is studied,and a filter based solution is provided.An estimation algorithm is developed in the case of denial-of-service attacks,which is used to estimate the software clock of the neighbor node using the existing clock information.Furthermore,the fliter based software clock reading compensation algorithm using the estimated information is proposed,which guarantees the clock synchronization in the presence of denial of service attacks.Simulation results verify the security and accuracy of the proposed clock synchronization algorithm for WSNs under malicious attacks.