Research On The Maximum Consensus Time Synchronization For Clustered Wireless Sensor Networks

Time synchronization is a crucial supporting technology for the stable operation of wireless sensor networks.In recent years,the distributed consensus-based time synchronization protocol has become a research hotspot in the field of clock synchronization because of its good robustness and scalability.However,such algorithms have slow convergence speed and large communication overhead.In the deployed wireless sensor networks,clustering architecture is used in many networks.Since Clustering can improve network scalability and promote data fusion.Therefore,in order to overcome the shortcomings of the consensus-based time synchronization algorithm,time synchronization methods that combine the consensus-based time synchronization protocols and the clustered network are proposed.However,this type of algorithm still needs further optimization.For example,each node needs to periodically broadcast synchronization messages to neighboring nodes,causing massive messages exchange in the synchronization process.Driven by this,based on the structure characteristics of clustered wireless sensor network,this thesis propose improved maximum consensus time synchronization methods for clustered wireless sensor networks.The main contributions of this thesis include:Firstly,this thesis investigates the time synchronization technology in wireless sensor network.It introduces and analyzes the current status of time synchronization and the consensus-based time synchronization algorithms.Secondly,to address the problem of the large communication overhead of the consensus-based time synchronization algorithms for clustered wireless sensor networks,this thesis proposes a forwarding-based maximum consensus time synchronization method for clustered wireless sensor networks(FBTS).In this method,the time synchronization process is started by the cluster heads and is divided into three steps,i.e.,the first intra-cluster time synchronization,the forwarding-based inter-cluster time synchronization and the second intra-cluster time synchronization.In the forwarding-based inter-cluster synchronization,the common nodes in the cluster do not need to communicate with neighbors to be synchronized.Compared with the existing algorithms that nodes broadcast or reply messages in each cycle,the proposed method can effectively reduce the communication overhead between nodes in the synchronization process and improve the convergence speed.Theoretical derivation and comparative analysis demonstrate that the proposed FBTS is superior to the maximum time synchronization(MTS)and the cluster-based maximum consensus synchronization(CMTS)in convergence speed and communication overhead.Thirdly,to further reduce the network traffic during the time synchronization process,this thesis proposes a threshold-based logical clock filtering maximum consensus time synchronization method(TBTS).This method divides the synchronization process into three stages: threshold-based intra-cluster time synchronization,inter-cluster time synchronization and one-way intra-cluster time synchronization.In the threshold-based intra-cluster time synchronization,a logical clock filtering method is proposed to reduce the number of interactions between nodes.Theoretical derivation and comparison show that TBTS and FBTS have the same convergence speed,and the communication overhead in TBTS is smaller than that of FBTS.Finally,through simulation,this thesis analyzes the convergence speed,scalability and the messages exchanges of the proposed FBTS and TBTS.Besides,we demonstrate the performance of the proposed methods through comparison with the current mainstream algorithms.Aiming at the consensus-based time synchronization for the clustered wireless sensor networks,this thesis presents improved consensus-based time synchronization methods,which can effectively improve the convergence speed and reduce the communication overhead.Moreover,the proposed methods can compensate the clock skew and offset at the same time,which prolong the period of re-synchronization.Thus,the research has certain reference value for the application and development of the consensus-based time synchronization algorithm.