WO2017035853A1 - 一种能量节约型无线传感器网络的构造和维护方法 - Google Patents

一种能量节约型无线传感器网络的构造和维护方法 Download PDF

Info

Publication number
WO2017035853A1
WO2017035853A1 PCT/CN2015/089234 CN2015089234W WO2017035853A1 WO 2017035853 A1 WO2017035853 A1 WO 2017035853A1 CN 2015089234 W CN2015089234 W CN 2015089234W WO 2017035853 A1 WO2017035853 A1 WO 2017035853A1
Authority
WO
WIPO (PCT)
Prior art keywords
node
nodes
energy
hypergraph
wireless sensor
Prior art date
Application number
PCT/CN2015/089234
Other languages
English (en)
French (fr)
Inventor
周怀北
邵珩
孔若杉
胡继承
Original Assignee
武汉大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 武汉大学 filed Critical 武汉大学
Publication of WO2017035853A1 publication Critical patent/WO2017035853A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
    • H04W16/18—Network planning tools
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W52/00—Power management, e.g. TPC [Transmission Power Control], power saving or power classes
    • H04W52/02—Power saving arrangements
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W52/00—Power management, e.g. TPC [Transmission Power Control], power saving or power classes
    • H04W52/02—Power saving arrangements
    • H04W52/0209—Power saving arrangements in terminal devices
    • H04W52/0212—Power saving arrangements in terminal devices managed by the network, e.g. network or access point is master and terminal is slave
    • H04W52/0216—Power saving arrangements in terminal devices managed by the network, e.g. network or access point is master and terminal is slave using a pre-established activity schedule, e.g. traffic indication frame
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W84/00—Network topologies
    • H04W84/18—Self-organising networks, e.g. ad-hoc networks or sensor networks
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Definitions

  • the invention relates to a construction and maintenance method of an energy-saving wireless sensor network, in particular to a construction method of an energy-saving wireless sensor network using a hypergraph theory.
  • This scheme includes graph theory, algorithm design, statistics and mathematical modeling, and proposes a new technology for the construction and optimization of wireless sensor networks.
  • WSN Wireless Sensor Network
  • WSN is a distributed sensor network with a sensor that senses and examines the outside world.
  • the sensors in the WSN communicate wirelessly, so the network settings are flexible, the device location can be changed quickly, and the Internet can be connected in a wired or wireless manner.
  • Hypergraph is the generalization of ordinary graphs, and one side (super edge) can contain no fewer than two nodes.
  • hypergraph type k-means hypergraph which is characterized in that the size of each super-edge in the hypergraph (ie, the number of nodes included) is k.
  • the purpose of hypergraph partitioning is to divide the nodes of the hypergraph into k roughly equal parts, and the case where nodes of the same hypergraph connecting multiple parts are minimized.
  • Module degree is used to measure the structure of the network. The higher the degree of modularity of a network, the closer the nodes within the network are. Correspondingly, these nodes are sparsely connected to nodes of other modules. Modularity is often used to find optimal strategies for constructing networks.
  • the invention mainly solves the deficiencies of the existing wireless sensor network constituent schemes; and provides an energy-saving wireless sensor network construction method based on the hypergraph as a theoretical basis, and plans the global topological structure of the wireless sensor network globally, for each
  • the sensor node performs scheduling to realize a wireless sensor network topology structure with reasonable structure and energy saving.
  • the hypergraph theory it makes up for the shortcomings of traditional simple graphs in information display and partitioning algorithm. It divides the wireless sensor network into multiple clusters, which makes the whole network topology hierarchical. The resulting wireless sensor network has high internals. Poly and low coupling characteristics.
  • the topology of the wireless sensor network is transformed into a hypergraph, and the hypergraph is divided and clustered to complete the high cohesion and low coupling structure;
  • each sensor node The location information of each sensor node is collected by the positioning system. Under the action of the positioning system, each node will receive a time point T, which means that they will send their own position information to the positioning system at time point T. Thereby achieving substantial synchronization of these location information;
  • the clustering is optimized.
  • the clusters are divided into high cohesion and low coupling characteristics. Therefore, a random walk is performed, and the range of motion is likely to be limited to one cluster.
  • the possibility of clusters is relatively small;
  • Control the entire network by minimizing the backbone network consisting of cluster head nodes and base stations. They are responsible for maintaining network connectivity, network coverage, and reducing energy consumption of the entire wireless sensor network.
  • the energy consumption of the network is reduced by scheduling the sensor nodes in the wireless sensor network. Nodes that are asleep can avoid unnecessary energy consumption due to your idle listening, and they will then be woken up by the cluster head node to which they belong, thus entering a working state.
  • the purpose of introducing the idle state is to buffer the process of transitioning from the sleep state to the working state, that is, when a node in a working state is to be converted into a sleep state, all the sleeping nodes in the sleep state will enter an idle state, and then The cluster head nodes to which they belong determine which candidate nodes are finally in the working state, and those nodes that are not selected will re-enter the sleep state;
  • RNG-based power control is implemented to reduce the communication power of all sensor nodes to be able to communicate with its farthest neighbor nodes, thereby minimizing the energy consumption caused by communication while ensuring network connectivity.
  • Step 1 Implementation of an energy-saving wireless sensor network construction and maintenance method relies on being able to acquire location information of all sensor nodes, so a positioning system is needed to collect all node locations and store this information in the base station.
  • the number of base stations may be more than one, but the construction method can only be run on one of the base stations, while the other base stations only store the same information as the primary base station as a copy or candidate of the primary base station;
  • Step 2 The primary base station sends a time point T to each node, which means that all sensor nodes will send their own location information to the positioning system at time point T, thereby achieving substantial synchronization of the location information;
  • Step 3 Adjust the communication power of each node through RNG-based power control, reduce the communication power to be able to communicate with its farthest neighbor node, and minimize the network connectivity and coverage. Energy consumption caused by communication;
  • Step 4 Map the sensor node to the vertices in the hypergraph, and then form a super edge according to the type of data collected by the sensor node, so that all the nodes in a super edge collect the same data type, and each node has at least one neighbor node. ;
  • Step 5 assign a weight to each super edge according to the number of vertices in the super edge
  • Step 6 Divide and cluster the hypergraph of the wireless sensor network based on the hypergraph theory
  • Step 7 Select a cluster head in each cluster through Wc.
  • the expression of Wc is specifically:
  • Er is the residual energy of the sensor node
  • Pc is the power of the node when the node is selected as the cluster head node
  • Pa is the sum of the powers of all nodes when the node is selected as the cluster head node
  • Ct and Cp are factors of Er/Pc and Pa, respectively.
  • Step 8 Dispatched the sensor nodes between sleep, idle, and active states to avoid unnecessary energy consumption.
  • the nodes are scheduled by the cluster head node to which they belong, and the frequency of network coverage redundancy and the data collected by the sensor nodes are taken into account. Therefore, if a node does not have to work all the time or is redundant for the coverage of the current network, the node will be scheduled to sleep and periodically wake up to collect data;
  • Step 9 When a node is about to transition from the working state to the sleep state, the cluster head node to which the node belongs will select the appropriate replacement node to switch the state of these nodes from the sleep state to the idle state. A node is finally selected as a replacement node and needs to be judged according to the weight W.
  • the weight W is calculated by the node itself, and the expression is specifically:
  • Er is the remaining energy of the sensor node
  • P is the communication power of the node
  • Step 1 The hypergraph is iteratively divided into two points, each time two points are performed according to a multi-level hypergraph partitioning algorithm
  • Step 2 The multi-level hypergraph partitioning algorithm is mainly composed of three parts, namely a coarsening stage, an initial dividing stage and a refinement stage;
  • Step 3 In the roughening phase, the initial hypergraph is performed based on the MHEC algorithm, which is applied to the whole process of the roughening phase, so the hypergraph is clustered at each level of the roughening phase;
  • Step 4 In the initial division phase, the initial division is performed on the hypergraph with the highest degree of coarsening.
  • the specific process is that the BFS algorithm randomly searches for a node from a node v until the ratio of the searched node to the entire hypergraph is ⁇ , and the value of ⁇ is adjusted between 0 and 100 according to the actual situation, and the default value is 50;
  • Step 5 Since the node v is randomly selected in the initial division stage, the result is not necessarily optimal, and as the hypergraph is refined, it will also affect the initial hypergraph partition, so The initial division stage holds ten best-performing initial divisions, which are screened according to their respective effects in the refinement phase;
  • Step 6 The single node is moved in the refinement phase, and the specific meaning is that the optimal module degree Q is obtained by moving the nodes around the initial partition to obtain an optimal partitioning strategy. Calculate the ⁇ Q values of the nodes near each split line, at which point the status of these nodes is unlocked. Move the node with the largest ⁇ Q value to the other side of the split line, and set the state of the node to be locked. The ⁇ Q values of the nodes near all the dividing lines are then updated, and the previous operation is repeated until the states of the nodes near all the dividing lines are locked or their ⁇ Q values are not greater than zero.
  • the nodes may leave, join the network, or move in the network, so the parameters of the nodes are dynamically variable. Therefore, it is necessary to periodically perform hypergraph partitioning and clustering on the primary base station, so that the optimal structure of the network continues to be effective.
  • FIG. 1 is a schematic diagram of a topology structure of a wireless sensor network.
  • Figure 2 is a simplified schematic diagram of the network partitioning process.
  • Figure 3 is a simplified schematic diagram of the hypergraph binary process.
  • Figure 4 is a schematic diagram of pseudo code in the refinement phase of the hypergraph binary.
  • Figure 5 is a schematic diagram of a single node movement.
  • FIG. 1 is a schematic diagram of a topology of a wireless sensor network, which includes a base station 101 and clusters 102 (1-3).
  • Each cluster 102 (1-3) consists of a cluster head 103 (1-3) and a number of sensor nodes 104 (1-3), 105 (1-4), 106 (1-3).
  • Step 1 Implementation of an energy-saving wireless sensor network construction and maintenance method relies on the ability to acquire all sensor nodes 103 (1-3), 104 (1-3), 105 (1-4), 106 (1) -3) location information, so a positioning system is needed to collect all the nodes 103 (1-3), 104 (1-3), 105 (1-4), 106 (1-3) positions, and these Information is stored in the base station 101.
  • the number of base stations 101 may be more than one but only one of the base stations 101 may be configured to operate, while the other base stations 101 only store the same information as the primary base station as a copy or candidate for the primary base station 101;
  • Step 2 The primary base station 101 sends a time point T to each node, which means that all sensor nodes 103 (1-3), 104 (1-3), 105 (1-4), 106 (1-3)
  • the location information will be sent to the positioning system at the time point T, so that the location information is substantially synchronized;
  • Step 3 Adjust the communication power of each node 103 (1-3), 104 (1-3), 105 (1-4), 106 (1-3) by RNG-based power control, and reduce the communication power to Being able to communicate with its farthest neighbor nodes, thereby minimizing the energy consumption caused by communication while ensuring network connectivity and coverage;
  • Step 4 Map sensor nodes 103 (1-3), 104 (1-3), 105 (1-4), 106 (1-3) to the vertices in the hypergraph, and then according to node 103 (1-3) , 104 (1-3), 105 (1-4), 106 (1-3)
  • the type of data collected forms a super edge, so that all nodes in a super edge collect the same data type, and each node has at least a neighbor node;
  • Step 5 assign a weight to each super edge according to the number of vertices in the super edge
  • Step 6 Dividing and clustering the hypergraph of the wireless sensor network based on the hypergraph theory, thereby forming a plurality of clusters 102 (1-3) in the network;
  • Step 7 Select a cluster head 103 (1-3) in each cluster 102 (1-3) by Wc, and the expression of Wc is specifically:
  • Er is the residual energy of the sensor node
  • Pc is the power of the node when the node is selected as the cluster head node
  • Pa is the sum of the powers of all nodes when the node is selected as the cluster head node
  • Ct and Cp are factors of Er/Pc and Pa, respectively.
  • FIG. 2 is a simplified schematic diagram of a network partitioning process that is presented in the form of a binary tree 201 in which the original hypergraph 202 is iteratively divided by a multi-level hypergraph partitioning algorithm.
  • the original hypergraph 202 is divided into two sections 203 (1-2), and the two sections 203 (1-2) are further divided into four sections 204 (1-4). As such, the partitioning is performed iteratively until a certain set condition is reached.
  • Hypergraph partitioning is based on modularity, so the iteration stops when either of the following conditions is met:
  • FIG. 3 is a simplified schematic diagram of a hypergraph binary process showing three stages 301, 302, 303 of hypergraph partitioning.
  • the three stages are: roughening stage 301, initial dividing stage 302, and refinement stage 303, and their specific implementation processes are as follows:
  • Step 1 In the roughening phase 301, the initial hypergraph is performed based on the MHEC algorithm, which is applied to the entire process of the roughening phase, so the hypergraph is clustered at each level of the roughening phase;
  • Step 2 In the initial partitioning phase 302, the initial partition 304(1) is performed on the hypergraph with the highest degree of coarsening.
  • the specific process is that the BFS algorithm randomly searches for a node from a node v until the ratio of the searched node to the entire hypergraph is ⁇ , and the value of ⁇ is adjusted between 0 and 100 according to the actual situation, and the default value is 50;
  • Step 3 Since the node v is randomly selected in the initial division stage, the result is not necessarily optimal, and as the hypergraph is refined, it will also affect the initial hypergraph partition, so The initial partitioning stage stores ten best-performing initial partitions, and in the refinement phase 303, screening according to their respective effects;
  • Step 4 In the refinement phase 303, a single node is moved, and the specific meaning is that the optimal module degree Q is obtained by moving the nodes around the initial partition to obtain an optimal partitioning strategy. Calculate the ⁇ Q values of the nodes near each split line, at which point the status of these nodes is unlocked. Move the node with the largest ⁇ Q value to the other side of the split line, and set the state of the node to be locked. The ⁇ Q values of the nodes near all the dividing lines are then updated, and the previous operation is repeated until the states of the nodes near all the dividing lines are locked or their ⁇ Q values are not greater than zero.
  • Figure 4 is a pseudo-code diagram of the refinement phase in the hypergraph binary, where calculating and updating ⁇ Q will play a crucial role in the algorithm.
  • ⁇ Q is done in a local range, so the movement of a single node may have no effect on the value of Q.
  • ⁇ Qv ⁇ T - ⁇ ED
  • ⁇ ED can be calculated from the weights of the super-edges belonging to the same cluster
  • ⁇ T represents the degree of coupling of a cluster.
  • Figure 5 is a schematic diagram of a single node movement.
  • a super edge around a dividing line has two possible states, a steady state and a critical state.
  • the super-edge E1 501 belongs to a steady state
  • the super-edge E2 502 belongs to the connected state. Therefore, the value of ⁇ Q can be obtained by calculating the super side in the critical state, and the value of ⁇ Q can be updated by calculating the vertices in the same super side.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

本发明公开了一种能量节约型无线传感器网络的构造和维护方法，尤其是涉及一种利用超图理论的能量节约型无线传感器网络的构造方法。本方案包括图论、算法设计、统计学以及数学建模，为无线传感器网络的构造和优化提出了一种新技术。

Description

一种能量节约型无线传感器网络的构造和维护方法 技术领域
本发明涉及一种能量节约型无线传感器网络的构造和维护方法，尤其是涉及一种利用超图理论的能量节约型无线传感器网络的构造方法。本方案包括图论、算法设计、统计学以及数学建模，为无线传感器网络的构造和优化提出了一种新技术。
背景技术
无线传感器网络(WSN)：WSN是一种分布式传感器网络，它的末梢是可以感知和检查外部世界的传感器。WSN中的传感器通过无线方式通信，因此网络设置灵活，设备位置可以快速更改，还可以跟互联网进行有线方式或者无线方式的连接。
超图理论：超图是普通图的泛化，它的一条边(超边)可以包含的结点数不少于2个。超图的一般定义为H＝(V，E)，其中V为所有的节点的集合，E为所有超边的集合，且E的元素均为V的非空子集；有权重的超图的定义为H＝(V，E，w)，新添加的参数w表示超图H中的每一条超边均被分配了权重w。此外，还有一种超图类型为k均值超图，其特征为超图中的每一条超边的大小(即所包含的结点数)均为k。超图划分的目的在于，将超图的节点划分为k个大致相等的部分，且出现同一个超图连接多个部分的节点的情况被最小化。
模块度：模块度被用来衡量网络的结构，一个网络的模块度越高说明该网络内部的节点联系越紧密。相应地，这些节点与其他模块的节点联系比较稀疏。模块度经常被用来寻找构造网络的最优化策略。
发明内容
本发明主要是解决现有无线传感器网络构成方案所存在的不足；提供了一种以超图作为理论基础的能量节约型无线传感器网络构造方法，从全局统筹规划无线传感器网络的拓扑结构，对各个传感器节点进行调度，从而实现一种结构合理，能耗节约型的无线传感器网络拓扑结构。通过引用超图理论，弥补了传统的简单图在信息展示以及划分算法方面的缺陷，它将无线传感器网络划分为多个簇，使整个网络拓扑结构具有层次性，所得的无线传感器网络具有高内聚和低耦合的特性。
本发明的技术解决方案是：
一种能量节约型无线传感器网络的构造和维护方法，
通过将超图理论运用于全局网络的拓扑结构规划，将无线传感器网络的拓扑结构转化为超图，并对其进行超图划分和聚类，完成高内聚和低耦合的构造；
通过定位系统来收集每个传感器节点的位置信息，在定位系统的作用下，每个节点会接受到一个时间点T，这意味着它们将在时间点T一齐向定位系统发送自己的位置信息，从而实现这些位置信息大致的同步；
通过寻找最优模块度实现对分簇的优化，被划分出来的这些簇具有高内聚和低耦合的特性，因此执行一次随机游走，其活动范围很有可能局限一个簇内，而出现跨簇的可能性比较小；
通过最小化由簇头节点和基站所组成的主干网络来控制整个网络，它们负责通信的同时，还负责维护网络的连通性、网络覆盖率以及减小整个无线传感器网络的能量消耗；
通过对无线传感器网络中传感器节点的调度来减小网络的能量消耗。处于睡眠状态的节点能够避免因为你空闲监听而造成的不必要的能量消耗，它们之后将被其所属的簇头节点唤醒，从而进入工作状态。引入空闲状态的目的在于，使其作为睡眠状态向工作状态转换的过程的缓冲，即当一个处于工作状态的节点将要转换为睡眠状态时，所有的处于睡眠状态的候补节点将进入空闲状态，然后由它们所属的簇头节点来决定最终哪些候补节点进入工作状态，而那些没有被选中的节点将重新进入睡眠状态；
通过基于RNG的功率控制来实现将所有的传感器节点的通信功率缩小到能够与它的最远邻居节点通信，从而在保证网络连通性的前提下，尽可能减小通信所造成的能量消耗。
利用超图理论实现的一种能量节约型无线传感器网络的构造和维护方法，具体构造步骤为：
步骤1：一种能量节约型无线传感器网络的构造和维护方法的实现依赖于能够获取所有的传感器节点的位置信息，所以需要一个定位系统来来收集所有的节点位置，并将这些信息存储在基站之中。基站的数量可能不止一个，但是只能在其中一个基站上运行构造方法，而其他的基站只是存储和主基站相同的信息，作为主基站的拷贝或者候补；
步骤2：主基站向每个节点发送一个时间点T，这意味着所有的传感器节点将在时间点T一齐向定位系统发送自己的位置信息，从而实现这些位置信息大致的同步；
步骤3：通过基于RNG的功率控制来调节每一个节点的通信功率，将通信功率缩小到能够与它的最远邻居节点通信，从而在保证网络连通性和覆盖率的前提下，尽可能减小通信所造成的能量消耗；
步骤4：将传感器节点映射到超图中的顶点，然后根据传感器节点采集数据的类型形成超边，从而在一条超边中所有的节点采集的数据类型相同，且每个节点至少有一个邻居节点；
步骤5：根据超边内顶点的数量为每一条超边分配权值；
步骤6：基于超图理论对无线传感器网络的超图进行划分和成簇；
步骤7：通过Wc在每个簇中选择一个簇头，Wc的表达式具体为：
Figure PCTCN2015089234-appb-000001
其中，Er为传感器节点的剩余能量；Pc为当该节点被选为簇头节点时，该节点的运行的功率；Pa为当该节点被选为簇头节点时，所有节点的运行的功率总和；Ct和Cp分别是Er/Pc和Pa的因子。
步骤8：对传感器节点在睡眠、空闲和工作三个状态之间进行调度，从而避免不必要的能量消耗。初始化时，所有节点处于工作状态，节点由所属的簇头节点进行调度，并将网络覆盖冗余以及传感器节点的采集数据的频率考虑其中。所以，如果一个节点没必要一直工作或者对于目前的网络的覆盖率来说是冗余的，则这个节点将被调度至睡眠状态，并周期性的唤醒来采集数据；
步骤9：当一个节点即将从工作状态转换为睡眠状态时，该节点所属的簇头节点将会选择合适的替换节点，将这些节点的状态从睡眠状态切换至空闲状态。一个节点最终被选择为替换节点需要根据权值W来判断，权值W由节点自己来计算，其表达式具体为：
Figure PCTCN2015089234-appb-000002
其中，Er为传感器节点的剩余能量；P为该节点的通信功率。
利用超图理论实现的分割和成簇，具体构造步骤为：
步骤1：超图被迭代地二分，每次二分都根据多层次的超图划分算法进行；
步骤2：多层次的超图划分算法主要由三个部分组成，分别为粗化阶段、初始划分阶段和细化阶段；
步骤3：在粗化阶段，初始超图基于MHEC算法进行，该算法被应用与粗化阶段的整个过程，所以超图在粗化阶段的各个层级上都会被进行分簇；
步骤4：在初始划分阶段，初始划分是在粗化程度最高的超图上进行的。其具体过程为，根据BFS算法随机从一个节点v开始搜索节点，直至搜索到的节点占整个超图的比率为ρ，而ρ的值则根据实际情况在0到100之间调整，其默认值为50；
步骤5：由于初始划分阶段中节点v是随机选取的，所以得出的结果不一定是最优的，且随着超图的不断细化，也会对最初的超图划分有影响，所以在初始划分阶段保存十个效果最好的初始划分，在细化阶段中根据它们各自的效果进行筛选；
步骤6：在细化阶段将引用单个节点移动，其具体含义为通过移动初始划分周围的节点来获取最优模块度Q，从而获得最优的划分策略。计算每个分割线附近的节点的ΔQ值，此时这些节点的状态为未锁住的。移动ΔQ值最大的节点到分割线的另外一边，并设置该节点的状态为锁住的。然后更新所有分割线附近的节点的ΔQ值，重复执行上一个操作，直至所有分割线附近的节点的状态都为锁住的或者它们的ΔQ值都不大于0。
在构建完能量节约型无线传感器网络之后，节点可能离开、加入网络，也有可能在网络中移动，因此节点的参数是动态可变的。所以需要周期性的在主基站上运行超图划分和成簇，使得网络的最优结构持续有效。
附图说明
图1是无线传感器网络的拓扑结构示意图。
图2是简化的网络划分流程示意图。
图3是简化的超图二分流程示意图。
图4是超图二分中细化阶段的伪代码示意图。
图5是单个节点移动的示意图。
具体实施方式
图1是无线传感器网络的拓扑结构示意图，它包含一个基站101以及一些簇102(1-3)。每一个簇102(1-3)由一个簇头103(1-3)和一定数量的传感器节点104(1-3)，105(1-4)，106(1-3)组成。
其具体构造步骤为：
步骤1：一种能量节约型无线传感器网络的构造和维护方法的实现依赖于能够获取所有的传感器节点103(1-3)，104(1-3)，105(1-4)，106(1-3)的位置信息，所以需要一个定位系统来来收集所有的节点103(1-3)，104(1-3)，105(1-4)，106(1-3)位置，并将这些信息存储在基站101之中。基站101的数量可能不止一个但是只能在其中一个基站101上运行构造方法，而其他的基站101只是存储和主基站相同的信息，作为主基站101的拷贝或者候补；
步骤2：主基站101向每个节点发送一个时间点T，这意味着所有的传感器节点103(1-3)，104(1-3)，105(1-4)，106(1-3)将在时间点T一齐向定位系统发送自己的位置信息，从而实现这些位置信息大致的同步；
步骤3：通过基于RNG的功率控制来调节每一个节点103(1-3)，104(1-3)，105(1-4)，106(1-3)的通信功率，将通信功率缩小到能够与它的最远邻居节点通信，从而在保证网络连通性和覆盖率的前提下，尽可能减小通信所造成的能量消耗；
步骤4：将传感器节点103(1-3)，104(1-3)，105(1-4)，106(1-3)映射到超图中的顶点，然后根据节点103(1-3)，104(1-3)，105(1-4)，106(1-3)采集数据的类型形成超边，从而在一条超边中所有的节点采集的数据类型相同，且每个节点至少有一个邻居节点；
步骤5：根据超边内顶点的数量为每一条超边分配权值；
步骤6：基于超图理论对无线传感器网络的超图进行划分和成簇，从而在网络中形成多个簇102(1-3)；
步骤7：通过Wc在每个簇102(1-3)中选择一个簇头103(1-3)，Wc的表达式具体为：
Figure PCTCN2015089234-appb-000003
其中，Er为传感器节点的剩余能量；Pc为当该节点被选为簇头节点时，该节点的运行的功率；Pa为当该节点被选为簇头节点时，所有节点的运行的功率总和；Ct和Cp分别是Er/Pc和Pa的因子。
图2是简化的网络划分流程示意图，它被以二叉树201的形式展示，其中原始超图202被多层次的超图划分算法迭代地二分。原始超图202被二分为两个部分203(1-2)，而这两个部分203(1-2)进一步的被分为四个部分204(1-4)。如此，划分被迭代地执行，直至到达某个设定的条件。
超图划分是基于模块度来实施的，所以迭代在符合以下任意一个条件时停止：
条件1：最优的顶点被找到；
条件2：所有的顶点最终都属于一个簇。
图3是简化的超图二分流程示意图，它展示了超图划分的三个阶段301,302,303。这三个阶段分别为：粗化阶段301、初始划分阶段302以及细化阶段303，它们的具体执行过程如下：
步骤1：在粗化阶段301，初始超图基于MHEC算法进行，该算法被应用与粗化阶段的整个过程，所以超图在粗化阶段的各个层级上都会被进行分簇；
步骤2：在初始划分阶段302，初始划分304(1)是在粗化程度最高的超图上进行的。其具体过程为，根据BFS算法随机从一个节点v开始搜索节点，直至搜索到的节点占整个超图的比率为ρ，而ρ的值则根据实际情况在0到100之间调整，其默认值为50；
步骤3：由于初始划分阶段中节点v是随机选取的，所以得出的结果不一定是最优的，且随着超图的不断细化，也会对最初的超图划分有影响，所以在初始划分阶段保存十个效果最好的初始划分，在细化阶段303中根据它们各自的效果进行筛选；
步骤4：在细化阶段303将引用单个节点移动，其具体含义为通过移动初始划分周围的节点来获取最优模块度Q，从而获得最优的划分策略。计算每个分割线附近的节点的ΔQ值，此时这些节点的状态为未锁住的。移动ΔQ值最大的节点到分割线的另外一边，并设置该节点的状态为锁住的。然后更新所有分割线附近的节点的ΔQ值，重复执行上一个操作，直至所有分割线附近的节点的状态都为锁住的或者它们的ΔQ值都不大于0。
图4是超图二分中细化阶段的伪代码示意图，其中计算和更新ΔQ将在算法中起着至关重要的作用。实际上，ΔQ是在局部范围内完成的，所以对单个节点的移动可能对Q的值并没有什么影响。给定一个节点v，则有表达式ΔQv＝ΔT-ΔED，其中ΔED可以通过属于同一个簇的超边的权值计算得到，ΔT表示一个簇的耦合度。
图5是单个节点移动的示意图。
给定一个划分，它将超图分割为超边E1 501和超边E2 502，顶点B 503属于第一个部分，而其他顶点504,505,506属于第二个部分。如果讲顶点A 504和顶点B 505移动到第一部分，超边E1 501将不再被 分割，而超边E2 502将被分割。相应地，Q值将会发生变化，但是如果只移动一个节点A 504，Q值将不会改变。
一个在分割线周围的超边有两个可能的状态，分别为稳定状态和临界状态。在顶点A 504被移动之前，超边E1 501属于稳定状态，而在移动了顶点A 504和顶点C 505之后，超边E2 502属于连接状态。因此，可以通过计算处于临界状态的超边来获取ΔQ的值，通过对同一个超边内的顶点进行计算来更新ΔQ的值。

Claims (14)

  1. 一种能量节约型无线传感器网络的构造和维护方法，其特征在于：
    通过构建一个定位系统，监测每一个传感器节点的位置信息；
    通过超图理论对无线传感器网络进行划分和成簇；
    通过调整每个传感器节点的传输功率来达到功率控制的目的；
    通过节点所属的簇头，对传感器节点的状态进行调度，从而达到节约能量的目的。
  2. 根据权利要求1所述的能量节约型无线传感器网络的构造和维护方法，其特征在于：通过构建一个定位系统，监测每一个传感器节点的位置信息，
    该定位系统为每一个传感器节点提供位置信息查询机制，所以基站、簇头节点和其他普通节点都可以随时获取网络中各个节点的位置信息。
  3. 根据权利要求1所述的能量节约型无线传感器网络的构造和维护方法，其特征在于：
    每个传感器节点均由四个部分组成，分别为传感模块、处理模块、能量模块以及通信模块；
    传感器节点之间在处理速度和存储容量方面差异巨大，所以那些处理速度快，存储容量大的节点更有可能成为簇头节点。
  4. 根据权利要求1所述的能量节约型无线传感器网络的构造和维护方法，其特征在于：
    该无线传感器网络是异构的，节点之间在硬件方面以及工作环境方面都有可能差异巨大；
    无线传感器网络在划分和成簇之后，其将由一个或者多个基站，一定数量的簇头节点和其他普通传感器节点组成。
  5. 根据权利要求1所述的能量节约型无线传感器网络的构造和维护方法，其特征在于：通过调整每个传感器节点的传输功率来达到功率控制的目的，
    缩小各个传感器节点的通信功率，叵至其能够与离自己最远的邻居节点通信，因此在减小传感器节点通信功率的同时，也保证了网络的连通性和覆盖范围。
  6. 根据权利要求1所述的能量节约型无线传感器网络的构造和维护方法，其特征在于：通过节点所属的簇头，对传感器节点的状态进行调度，从而达到节约能量的目的，
    传感器节点的状态调度能够使节点状态在睡眠、工作和空闲之间周期性切换；
    睡眠状态使节点能够避免因为空闲监听对于能量的浪费；
    空闲状态作为一个缓冲，尤其是当多个节点同时作为一个将要切换为睡眠状态的节点的候补时，如此可以避免因多个节点重复工作所造成的能量损失。
  7. 根据权利要求4所述的能量节约型无线传感器网络的构造和维护方法，其特征在于：基站和簇头节点，
    它们都在网络架构中有着至关重要的作用，其中，基站保存所有传感器节点的位置信息，而簇头节点是从每个基于超图理论而被划分出的簇中选择出来的；
    当传感器网络中出现多个基站时，只有一个基站充当生成网络架构的角色，而其他节点虽然也保存网络中节点的位置信息，但是它们只是作为该基站的候补。
  8. 根据权利要求7所述的能量节约型无线传感器网络的构造和维护方法，其特征在于：基于超图理论的划分和成簇，
    超图分割可以分为三个阶段，分别为粗化阶段，初始划分阶段和细化阶段；
    超图分割被迭代执行，从而不断的对网络进行二分，叵至达到模块度最优；
    成簇在超图分割的粗化阶段中被执行，其通过MHEC算法将各个传感器节点融合到簇里面。
  9. 根据权利要求8所述的能量节约型无线传感器网络的构造和维护方法，其特征在于：
    模块度的大小表示了一个簇的内聚性；
    通过寻找最优模块度，就可以自动的寻找到基于超图理论的最优划分数k，从而避免了人工计算最优划分数的工作。
  10. 根据权利要求8所述的能量节约型无线传感器网络的构造和维护方法，其特征在于：基于超图理论的划分和成簇，
    簇头的选择是基于传感器节点的权值Wc得到的，Wc的表达式具体为：
    Figure PCTCN2015089234-appb-100001
    其中，Er为传感器节点的剩余能量；Pc为当该节点被选为簇头节点时，该节点的运行的功率；Pa为当该节点被选为簇头节点时，所有节点的运行的功率总和；Ct和Cp分别是Er/Pc和Pa的因子。
  11. 根据权利要求8所述的能量节约型无线传感器网络的构造和维护方法，其特征在于：超图划分的粗化阶段，
    在粗化阶段会通过MHEC算法迭代地生成一个超图序列，在该序列中超图的粗化程度不断呈递增趋势。
  12. 根据权利要求8所述的能量节约型无线传感器网络的构造和维护方法，其特征在于：超图划分的初始划分阶段，
    初始划分阶段分为两个步骤完成，其一，按BFS算法随机从一个节点v开始搜索节点，叵至搜索到的节点占整个超图的比率为p；
    其二，保留十个划分效果最好的划分策略，以便在细化阶段选出最终的划分策略。
  13. 根据权利要求8所述的能量节约型无线传感器网络的构造和维护方法，其特征在于：超图划分，
    超图划分的细化阶段包含单一节点移动，其目的在于通过在分割线附近移动节点来增加模块度Q的值，从而达到获取最优模块度的目的。
  14. 根据权利要求12所述的能量节约型无线传感器网络的构造和维护方法，其特征在于：超图划分的细化阶段，
    细化阶段由三个部分组成，其一，计算每个分割线附近的节点的ΔQ值，此时这些节点的状态为未锁住的；
    其二，移动ΔQ值最大的节点到分割线的另外一边，并设置该节点的状态为锁住的；
    其三，更新所有分割线附近的节点的ΔQ值，然后重复第二步操作，叵至所有分割线附近的节点的状态都为锁住的或者它们的ΔQ值都不大于0。
PCT/CN2015/089234 2015-09-02 2015-09-09 一种能量节约型无线传感器网络的构造和维护方法 WO2017035853A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510560399.3A CN105101233A (zh) 2015-09-02 2015-09-02 一种能量节约型无线传感器网络的构造和维护方法
CN201510560399.3 2015-09-02

Publications (1)

Publication Number Publication Date
WO2017035853A1 true WO2017035853A1 (zh) 2017-03-09

Family

ID=54580552

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/089234 WO2017035853A1 (zh) 2015-09-02 2015-09-09 一种能量节约型无线传感器网络的构造和维护方法

Country Status (2)

Country Link
CN (1) CN105101233A (zh)
WO (1) WO2017035853A1 (zh)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110645988A (zh) * 2019-10-11 2020-01-03 广州大学 一种基于最优寿命的无人机路径规划方法
CN111698656A (zh) * 2020-05-21 2020-09-22 江苏海洋大学 一种水下移动无线传感器网络的多目标动态调度方法
CN111800751A (zh) * 2020-05-13 2020-10-20 江苏华东新能源勘探有限公司（江苏省有色金属华东地质勘查局八一三队） 一种地下水水质监控系统
CN112055322A (zh) * 2020-08-04 2020-12-08 江苏海洋大学 一种基于区间多目标的水下无线传感器网络调度优化模型
EP3849218A1 (en) * 2020-01-07 2021-07-14 Rockwell Collins, Inc. Stealth position-location information for high value nodes
CN113472573A (zh) * 2021-06-30 2021-10-01 三峡大学 用于无线传感器网络资源调度的高维多目标协同优化方法
CN116232921A (zh) * 2023-05-08 2023-06-06 中国电信股份有限公司四川分公司 一种基于超图的确定性网络数据集构建装置及其方法
CN116456372A (zh) * 2023-06-12 2023-07-18 合肥工业大学 基于动态缩减策略空间的无线监测网络调度方法和系统
CN116600366A (zh) * 2023-06-02 2023-08-15 中国通信建设第三工程局有限公司 一种基于无线通信网络的数据获取方法及系统

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106209390B (zh) * 2016-06-30 2021-04-13 武汉大学深圳研究院 网络设备管理及节能方法及其资源管理装置、路由节能装置
CN109633532A (zh) * 2018-12-21 2019-04-16 青岛安然物联网科技有限公司 一种无线信标定位系统及其定位方法
CN111836282B (zh) * 2020-06-28 2022-04-19 中南大学 一种三维wsn节点调度方法及存储介质
CN112188518B (zh) * 2020-09-07 2024-02-13 北京燧昀科技有限公司 一种传感器节点通信优化方法、装置及可读存储介质
CN113567910A (zh) * 2021-07-22 2021-10-29 青岛安然物联网科技有限公司 一种无线信标定位系统及其定位方法

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103701647A (zh) * 2013-12-31 2014-04-02 武汉大学 一种无线网络拓扑优化生成方法

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103701647A (zh) * 2013-12-31 2014-04-02 武汉大学 一种无线网络拓扑优化生成方法

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
LI, LEI: "The Research on Hypergraph-based Key Management Protocols for Heterogeneous Sensor Networks", MASTER'S DISSERTATION OF HARBIN ENGINEERING UNIVERSITY, 15 March 2013 (2013-03-15) *
TIAN, WENFENG: "Application of Hypergraph Theory in Wireless Sensor Network", MASTER'S DISSERTATION OF NANJING UNIVERSITY OF POSTS AND TELECOMMUNICATIONS, 22 December 2010 (2010-12-22) *
YOUNIS, M. ET AL.: "Topology management techniques for tolerating node failures in wireless sensor networks: A survey", COMPUTER NETWORKS, vol. 58, 15 January 2014 (2014-01-15), pages 254 - 283, XP028608004 *

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110645988A (zh) * 2019-10-11 2020-01-03 广州大学 一种基于最优寿命的无人机路径规划方法
CN110645988B (zh) * 2019-10-11 2023-08-29 广州大学 一种基于最优寿命的无人机路径规划方法
EP3849218A1 (en) * 2020-01-07 2021-07-14 Rockwell Collins, Inc. Stealth position-location information for high value nodes
CN111800751B (zh) * 2020-05-13 2023-06-02 江苏华东新能源勘探有限公司（江苏省有色金属华东地质勘查局八一三队） 一种地下水水质监控系统
CN111800751A (zh) * 2020-05-13 2020-10-20 江苏华东新能源勘探有限公司（江苏省有色金属华东地质勘查局八一三队） 一种地下水水质监控系统
CN111698656A (zh) * 2020-05-21 2020-09-22 江苏海洋大学 一种水下移动无线传感器网络的多目标动态调度方法
CN112055322A (zh) * 2020-08-04 2020-12-08 江苏海洋大学 一种基于区间多目标的水下无线传感器网络调度优化模型
CN113472573A (zh) * 2021-06-30 2021-10-01 三峡大学 用于无线传感器网络资源调度的高维多目标协同优化方法
CN116232921A (zh) * 2023-05-08 2023-06-06 中国电信股份有限公司四川分公司 一种基于超图的确定性网络数据集构建装置及其方法
CN116232921B (zh) * 2023-05-08 2023-09-12 中国电信股份有限公司四川分公司 一种基于超图的确定性网络数据集构建装置及其方法
CN116600366A (zh) * 2023-06-02 2023-08-15 中国通信建设第三工程局有限公司 一种基于无线通信网络的数据获取方法及系统
CN116600366B (zh) * 2023-06-02 2023-10-27 中国通信建设第三工程局有限公司 一种基于无线通信网络的数据获取方法及系统
CN116456372A (zh) * 2023-06-12 2023-07-18 合肥工业大学 基于动态缩减策略空间的无线监测网络调度方法和系统
CN116456372B (zh) * 2023-06-12 2023-08-11 合肥工业大学 基于动态缩减策略空间的无线监测网络调度方法和系统

Also Published As

Publication number Publication date
CN105101233A (zh) 2015-11-25

Similar Documents

Publication Publication Date Title
WO2017035853A1 (zh) 一种能量节约型无线传感器网络的构造和维护方法
Wang et al. A PSO based energy efficient coverage control algorithm for wireless sensor networks.
Ma et al. ESCVAD: an energy-saving routing protocol based on voronoi adaptive clustering for wireless sensor networks
EP3014444A1 (en) Computing connected components in large graphs
Khorasani et al. Energy efficient data aggregation in wireless sensor networks using neural networks
CN106789320A (zh) 一种用于优化无线传感器网络拓扑的多种群协同进化方法
John et al. Energy saving cluster head selection in wireless sensor networks for internet of things applications
Li et al. An improved algorithm of LEACH routing protocol in wireless sensor networks
Tham et al. Multi-agent systems on sensor networks: A distributed reinforcement learning approach
CN111225398B (zh) 一种基于协作覆盖的微电网无线传感器网络能耗优化方法
Sun et al. A New Energy-efficient Multi-target Coverage Control Protocol Using Event-driven-mechanism in Wireless Sensor Networks.
CN103596203B (zh) 一种局部自维护的无线传感器网络节能分簇拓扑控制方法
Umar et al. Designing of Dynamic Re-clustering Leach Protocol for Calculating Total Residual Time and Performance.
CN106412935B (zh) 一种基于复杂网络理论的网络拓扑结构建立方法
CN112234599B (zh) 一种多元复杂城市电网超前动态自适应分区方法及其系统
Yang et al. Research on clustering routing algorithm based on K-means++ for WSN
CN108684066A (zh) 基于K-medoids的无线传感器网络分簇方法
Zhu et al. A mobile agent-based clustering data fusion algorithm in wsn
Zhang et al. A WSN clustering routing protocol based on improved Whale algorithm
Mehta et al. Comparative Study of BCDCP protocols in Wireless sensor network
Jain et al. Performance analysis of hierarchical agglomerative clustering in a wireless sensor network using quantitative data
Kumar et al. Enhancing network lifetime and throughput in heterogeneous wireless sensor networks
Liu et al. Correlation-model-based data aggregation in wireless sensor networks
Sun A survey on clustering routing protocols based on PSO in WSN
Song et al. Periodicity-and-linear-based data suppression mechanism for WSN

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15902660

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15902660

Country of ref document: EP

Kind code of ref document: A1