| Time synchronization is a key technology for wireless sensor networks that provides a global timescale for the local sensor nodes,which is the basic requirement of common time notion for various applications in the networks,such as data fusion,target tracking and coordination.Consensus-based time synchronization is an important category of time synchronization mechanism,providing information interaction and compensation rules for nodes,so that time notion of each node can approach consensus and achieve global time synchronization.On this basis,this kind of synchronization mechanism maintains the advantages of scalability and robustness,becoming a promising research field of time synchronization in recent years.In wireless sensor networks,the energy of nodes is limited,but frequent information interaction causes large communication overhead when applying consensus-based time synchronization,which shortens the life of wireless sensor networks.To solve the above problem,event-triggered consensus-based time synchronization has been proposed recently,making nodes transmit to it neighboring nodes at the requirement of certain triggering condition,so as to significantly reduce the communication overhead in the synchronization process.However,research of the above-mentioned synchronization method is still in the preliminary stage.On the one hand,current research could not counteract the influence of communications delays and time-varying clock skew in static network,and on the other hand,does not consider dynamic network scenario where the nodes are moving randomly.To address the aforementioned problems,considering static wireless sensor networks and random mobile sensor networks where communication delays exist,this thesis studies the consensus-based time synchronization based on event-triggered mechanism.The main contributions of the thesis are given below:1.The thesis investigates the current research of consensus-based time synchronization in wireless sensor networks,and gives the summary and analysis of relevant technical foundation,including consensus theory,models of network and clock,and evaluation index of time synchronization under wireless sensor networks.2.For the problem of large communication overhead caused by consensus-based time synchronization,we consider the scenario of communication delays and time-varying clocks,and propose an event-triggered consensus-based time synchronization.Firstly,a node interaction strategy is designed,triggering function of which maintains an appropriate convergence speed in the above scenario,and its threshold determines the communication overhead and synchronization accuracy.Under the communication scheme,sequential least square method is applied to estimate the relative skew between nodes to counteract the influence of communications delays with low storage overhead.At the same time,the proportional-integral estimator is used to update the logic skew compensation,so as to force the compensation value approach consensus under the influence of time-varying skew,achieving clock synchronization of global network.Then,theoretically prove that the proposed method achieves input-to-state stability under the influence of delays and timevarying skews and the error of logical clock is bounded,3.Since current research of event-triggered consensus-based time synchronization do not consider the scenario of random mobile sensor networks,this thesis aims at the network with delays,proposing an event-triggered maximum-value-based consensus time synchronization.An exponential trigger scheme as the node interaction rule is adopted for the above scenario.Assuming the communication frequency between nodes follow Poisson distribution,provide a maximum-value-based update rule for logical skew compensation.To counteract the influence of communications delays,decision variables are introduced and sequential least square method is applied to eliminate relative skew.Under the assumption communication delays follow normal distribution,theoretically prove that the proposed could reach convergence with probability 1.4.The two proposed time synchronization methods are verified by simulation.Simulating static wireless sensor networks and random mobile networks,proposed methods are compared with similar consensus-based synchronization methods from the perspective of synchronization accuracy and communication overhead.And trade-off parameters of triggering equation are studied by simulation.The scalability and robustness of the proposed method for static network are verified.Simulation results prove the effectiveness of the two proposed synchronization methods.Considering static wireless sensor networks and random mobile sensor networks where communication delays exist,this thesis studies consensus-based time synchronization method based on event-triggered node interaction mechanism.The proposed method could achieve network time synchronization in certain networks,with less communication overhead under the condition of ensuring synchronization accuracy,which is significant for the development of consensus-based time synchronization technology. |
PDF Full Text Request