Research On Time Synchronization Technology Of Satellite Network

In recent years,as a new and efficient communication method,satellite network communication has received widespread attention and has become increasingly mature.Satellite communication is not restricted by geographical factors,and is also applicable in sparsely populated areas;it is not affected by natural disasters,and local facilities can be quickly restored by intervention when they are damaged due to natural disasters.At the same time,satellite communication has a large ground coverage area,which can provide global users with full coverage services,and has flexible networking and high mobility.When a small number of satellites fail,backup satellites can be used for rapid replacement.Based on these advantages,the satellite network has more and more extensive application scenarios,and its role in communication is becoming more and more significant.The normal operation of the satellite network is inseparable from the time synchronization of the satellite network.In application scenarios such as satellite joint observation and satellite positioning,highprecision time synchronization within the satellite network is an indispensable prerequisite;in addition,the coordinated control,mutual positioning,and information exchange between satellites are also inseparable from the inter-satellite time Synchronize.Therefore,this article takes the time synchronization of the satellite network as the subject,and makes the following related research.(1)The time synchronization process of the satellite network is carried out layer by layer according to the hierarchical structure,and the influence of different time reference nodes on the time synchronization performance is analyzed.The traditional best time reference node selection usually takes a breadth-first traversal of all satellite nodes in the network,constructs a spanning tree,and compares the maximum number of layers that complete the synchronization of the entire network to select,which requires a large amount of calculation.In view of this situation,this paper proposes a breadth-first traversal time synchronization scheme based on pruning.In the process of constructing a spanning tree,some redundant nodes in the network are deleted by mining the information contained in the spanning tree,which reduces the amount of calculation.The theoretical basis and detailed steps of this scheme are given in the article.Finally,a simulation comparison is carried out to prove that it can effectively reduce the amount of calculation and realize the optimization of the time synchronization performance of the satellite network.(2)When the network scale is large or the overall topology of the network cannot be known,the processing method of selecting network time reference nodes in a centralized manner is no longer applicable.In view of this situation,this paper proposes a distributed time reference star selection method based on the message table.Each satellite in the network maintains a message table to record the path to other satellites and the number of layers required to complete synchronization when each node is used as a time reference node.Through the message table interaction of adjacent satellites,the update of path information and the best time are completed Selection of reference nodes.In view of the defect of redundant information in the message table,it has been improved.The level information of all nodes is no longer maintained and only the estimated best time reference node is recorded.Also through the message table interaction of adjacent nodes,the completion Based on the selection of the best time reference node,the amount of information that needs to be transmitted in the message table is reduced.(3)Aiming at the shortcomings of the time reference node being overly dependent in the step-by-step synchronization mode based on the time reference node,a collegiate and consistent time synchronization scheme based on the non-master-slave mode is proposed.In this non-master-slave synchronization scheme,all nodes have the same status.Each node in the network sends its current time information to neighboring satellites.The neighboring satellites add the received information to its own time and calculate the average value,and use this average value to update itself The current clock ticks.After several rounds of iteration,all nodes in the network finally reach the same time scale value,thus realizing the non-master-slave distributed time synchronization of the network.