Research Of WSN Time Synchronization Algorithm Based On Nonlinear Clock Model

Time synchronization of wireless sensor networks is a prerequisite for efficient information interaction and data fusion.Existing time synchronization algorithms usually only consider a single factor among the four design factors,such as time delay,synchronization accuracy,network energy consumption and network topology,and cannot meet the comprehensive requirements of large-scale network time synchronization for these factors.This article will consider the above four design elements to make the performance of the time synchronization algorithm balanced.In this paper,under two common time delay distribution models,the network can not guarantee the synchronization accuracy,the energy of the network node is limited,and the applicability of the network topology is insufficient.By estimating the clock drift parameters,reducing the synchronization message volume and generating the specific network topology,etc.The method is designed to design the time synchronization algorithm,and the network topology suitable for the respective time synchronization algorithm is generated.At the same time,the synchronization precision is improved to some extent and the network energy consumption is saved.The main work of the thesis is as follows:1)In view of the problem that the network can not guarantee the long-term synchronization accuracy,this paper considers that the clock drift has great influence on the long-term synchronization accuracy.Through the mathematical analysis of the clock model,the nonlinear clock model with quadratic term is introduced in detail,and the quadratic coefficient is introduced.It is the clock drift parameter,and the estimation of clock drift under the corresponding time synchronization algorithm is studied.2)For the Gaussian delay distribution model,the nonlinear clock model is applied to the paired broadcast synchronization algorithm.First,the joint maximum likelihood estimation and linear regression methods are used in the single cluster network to synchronize the sender-receiver in the network.(Sender-receiver synchronization,SRS)mechanism and(receiver-only synchronization,ROS)mechanism for joint estimation of clock skew,frequency offset and drift;then group-based pairwise selection in multi-cluster networks The algorithm generates atree topology,and the nodes in the group and the nodes use different time synchronization mechanisms.Finally,through simulation experiments,the improvement of synchronization precision,the effective generation of network topology and the reduction of network synchronization message volume are verified,which is beneficial to the long-term network.Time synchronization.3)For the more complex exponential delay distribution model,based on the nonlinear model clock model,firstly,the clock parameters of the SRS mechanism are estimated in a single cluster network,and the clock phase offset,frequency offset and drift estimation are obtained.The fixed time delay estimation is more suitable for the actual situation;then the non-uniform dynamic clustering method is adopted in the multi-cluster network,and the synchronization is initiated by the timing mode through the query mode.From the node energy and the node spacing,the dynamic adjustment can be obtained.Adapting to the cluster network topology where some nodes die due to energy exhaustion,the inter-cluster synchronization adopts continuous multi-hop synchronization algorithm,which reduces the amount of synchronous messages by nearly half compared with the sensor network timing synchronization protocol algorithm,and the synchronization time is also greatly reduced.Finally,through simulation experiments,the improvement of synchronization accuracy and the effective generation of network topology are verified.4)The experiment verifies the two time synchronization algorithms mentioned in the paper.The hardware uses CC2530 module and the software uses IAR EW.The experiment compares the synchronization error and cumulative energy consumption of the two time synchronization algorithms.The results show that both time synchronization algorithms can effectively improve.Accuracy and low cumulative network energy consumption.Figure [24] table [4] reference [59]...