CN103607747A

CN103607747A - Inter-cluster virtual backbone route protocol method based on power control

Info

Publication number: CN103607747A
Application number: CN201310666860.4A
Authority: CN
Inventors: 李霞; 方世良; 王丽玮; 王永倩; 安良; 王晓燕
Original assignee: Southeast University
Current assignee: Southeast University
Priority date: 2013-12-10
Filing date: 2013-12-10
Publication date: 2014-02-26
Anticipated expiration: 2033-12-10
Also published as: CN103607747B

Abstract

A power control-based method for inter-cluster virtual backbone routing protocol, which is applicable to the field of underwater acoustic communication. This method is improved on the basis of the CLUSTERPOW protocol, combined with the power control technology to select the appropriate number of power levels, each backbone node uses the handshake protocol to establish a corresponding number of routing tables for each power level, and the backbone node forwards data according to the destination node. Adaptively select the optimal path for power consumption, and only perform a route lookup at the source node during the sending of a data packet, and encapsulate the searched route into the data packet, effectively avoiding the occurrence of routing loops. And other forwarding nodes do not need to execute complex routing lookup protocols.

Description

A method of inter-cluster virtual backbone routing protocol based on power control

技术领域technical field

本发明涉及水声通信网络网络层的路由协议，尤其涉及一种基于功率控制的簇间虚拟骨干路由协议方法，属于水声信号处理技术领域。The invention relates to a routing protocol at the network layer of an underwater acoustic communication network, in particular to a power control-based inter-cluster virtual backbone routing protocol method, which belongs to the technical field of underwater acoustic signal processing.

技术背景technical background

很多分簇协议在簇头选举和簇构建完成后，数据的传输仅仅依靠簇头节点通过单跳通信模式将数据直接发送到基站，如果基站远离被监测区域，这将会消耗较多的能量，另外，这也要求簇头节点有足够大的通信功率，能够直接与基站通信，而这对于水下传感器节点来说很不现实，因此许多研究机构和个人针对簇间多跳通信模式即基于簇结构的多跳路由协议展开了研究。In many clustering protocols, after the cluster head election and cluster construction are completed, the data transmission only depends on the cluster head node to send the data directly to the base station through a single-hop communication mode. If the base station is far away from the monitored area, it will consume more energy. In addition, this also requires that the cluster head node has enough communication power to be able to communicate directly with the base station, which is very unrealistic for underwater sensor nodes. Structured multi-hop routing protocols are studied.

在基于簇结构的网络中，簇间通信通常结合虚拟骨干网技术和功率控制技术，骨干路由的查找要求骨干节点尽可能的以最小功率（最短路径）方法来连接另一个骨干节点。VikasKawadia和P.R.Kumar提出的CLUSTERPOW协议使用协商路由表的方法来寻找骨干网络里的最佳路由。In a cluster-based network, inter-cluster communication usually combines virtual backbone network technology and power control technology. The search for backbone routes requires backbone nodes to connect to another backbone node with the least power (shortest path) method possible. The CLUSTERPOW protocol proposed by VikasKawadia and P.R.Kumar uses the method of negotiating the routing table to find the best route in the backbone network.

CLUSTERPOW协议是一种基于功率控制技术的针对不同目的节点自适应地调整到合适发射功率的路由协议。该协议通过握手协议来建立本地功率等级路由表，每个骨干节点建立完自己本地的各个功率等级的路由表后，将自己的路由表信息通过包在全网内广播，这样网络中的骨干节点就知道了其他骨干节点的本地路由表信息。为了寻找功率最优化的路由，每一个骨干节点在转发数据包时,还必须建立一个核心路由表，核心路由表的功能是提供了到达目的节点的最短路由。核心路由表是在协商了本地的路由表之后，选择最小功率路由并且把将其复制到核心路由表内。核心路由表比普通路由表多出多出一个“发射功率”域，该域指出了到达下一跳节点所需的发射功率等级。骨干节点根据核心路由表将发射功率调整到相应等级，然后将数据转发给下一跳节点。类似地，下一跳节点在转发数据时建立自己的核心路由表并转发数据。每个中继节点建立自己的核心路由表同时转发数据包，直到将数据包发送到目的节点。另外，为避免路由环路的出现，即防止节点之间的信息传递陷入死循环，CLUSTERPOW规定在单个数据包的传输过程中节点所使用的发射功率必须是单调非递增的。但是CLUSTERPOW协议为避免产生路由回路，在数据包的一次发送过程中使用非递增的发射功率，这在一定程度上造成了发射功率的浪费。The CLUSTERPOW protocol is a routing protocol based on power control technology to adaptively adjust to the appropriate transmission power for different destination nodes. This protocol establishes local power level routing tables through the handshake protocol. After each backbone node establishes its own local routing table of each power level, it broadcasts its own routing table information in the entire network through packets, so that the backbone nodes in the network The local routing table information of other backbone nodes is known. In order to find a power-optimized route, each backbone node must also establish a core routing table when forwarding data packets. The function of the core routing table is to provide the shortest route to the destination node. The core routing table selects the least power route and copies it into the core routing table after negotiating the local routing table. The core routing table has one more "transmission power" field than the ordinary routing table, which indicates the transmission power level required to reach the next hop node. The backbone node adjusts the transmission power to the corresponding level according to the core routing table, and then forwards the data to the next hop node. Similarly, the next-hop node builds its own core routing table and forwards data when forwarding data. Each relay node establishes its own core routing table and forwards the data packet at the same time until the data packet is sent to the destination node. In addition, in order to avoid the occurrence of routing loops, that is, to prevent the information transmission between nodes from falling into an infinite loop, CLUSTERPOW stipulates that the transmission power used by nodes must be monotonically non-increasing during the transmission of a single data packet. However, in order to avoid routing loops, the CLUSTERPOW protocol uses non-incremental transmission power during one transmission of data packets, which causes waste of transmission power to a certain extent.

骨干节点的定义是：一个无线传感器网络划分为多个独立自治的小网络，每个小网络就是一个簇，根据簇头选取协议选取每个簇的簇头结点，簇头节点担任簇间虚拟骨干网的骨干节点（虚拟骨干网通常由簇头和一些普通节点共同构成，也可以仅由功能较强的簇头形成虚拟骨干网络；骨干节点可以由通过分簇协议选举出的簇头节点来担任）进行数据的转发工作；结合功率控制技术选取特定的功率等级数目（传输功率与节点间距离有关，增加功率等级数目可以更精确地调节发射功率，由此推测当功率等级数超过某个值时继续增加将不再明显节约能量，但小的功率等级数则更易于实现，功率等级数目的选取需要根据不同的网络应用来设定），并利用握手协议建立相应数目的本地功率等级路由表和全局功率等级路由表（骨干节点以每一个离散的功率等级p_i发送一个probe包，去探测在接收范围之内的其他骨干节点，其他骨干节点一旦收到probe包，就会回复一个prack包，骨干节点将给自己发送确认包的节点编号存入自己的p_i等级的路由表中。这样每个骨干节点就为每个功率等级维护一个路由列表，最后每个骨干节点形成本地功率等级路由表，结合网络中其他骨干节点的本地功率等级路由表形成全局功率等级路由表）；当骨干节点有数据包需要发送时，此骨干节点称为源节点。The definition of the backbone node is: a wireless sensor network is divided into multiple independent and autonomous small networks, each small network is a cluster, and the cluster head node of each cluster is selected according to the cluster head selection protocol, and the cluster head node acts as the inter-cluster virtual node. The backbone nodes of the backbone network (the virtual backbone network is usually composed of cluster heads and some common nodes, or only cluster heads with strong functions can form a virtual backbone network; the backbone nodes can be formed by the cluster head nodes elected by the clustering protocol In charge of forwarding data; combined with power control technology to select a specific number of power levels (transmission power is related to the distance between nodes, increasing the number of power levels can adjust the transmission power more accurately, so it is speculated that when the number of power levels exceeds a certain value It will no longer obviously save energy if the number of power levels continues to increase, but a small number of power levels is easier to implement. The selection of the number of power levels needs to be set according to different network applications), and use the handshake protocol to establish a corresponding number of local power level routing tables and the global power level routing table (the backbone node sends a probe packet at each discrete power level p _i to detect other backbone nodes within the receiving range, once other backbone nodes receive the probe packet, they will reply with a prack packet , the backbone node stores the node number sending the acknowledgment packet to itself in the routing table of its _pi level. In this way, each backbone node maintains a routing list for each power level, and finally each backbone node forms a local power level routing table, combined with the local power level routing tables of other backbone nodes in the network to form a global power level routing table); when a backbone node has data packets to send, this backbone node is called a source node.

技术内容：Technical content:

解决的技术问题：针对CLUSTERPOW协议为避免产生路由回路，在数据包的一次发送过程中使用非递增的发射功率而造成发射功率浪费的现象，对CLUSTERPOW协议进行改进，改进的CLUSTERPOW协议有效避免了路由环路的出现，同时降低了发射功率消耗。Technical problem to be solved: In order to avoid routing loops in the CLUSTERPOW protocol, the use of non-incremental transmission power during one-time transmission of data packets causes waste of transmission power, and the CLUSTERPOW protocol is improved. The improved CLUSTERPOW protocol effectively avoids routing. The appearance of the loop reduces the transmission power consumption at the same time.

本发明采用的技术方案：一种基于功率控制的簇间虚拟骨干路由协议方法，依据CLUSTERPOW协议方法，使用协商路由表的方法来寻找骨干网络里的最佳路由，结合功率控制技术选取功率等级数目并利用握手协议建立相应数目的本地功率等级路由表和全局功率等级路由表，针对不同数目的节点自适应地调整到合适发射功率实现路由，为了寻找功率最优化的路由，每一个骨干节点在转发数据包时须建立一个核心路由表，核心路由表提供了到达目的节点的最短路由，核心路由表在协商了本地的路由表之后，选择最小功率路由并且把将其复制到核心路由表内，骨干节点根据核心路由表将发射功率调整到相应等级，然后将数据转发给下一跳中继节点，下一跳中继节点在转发数据时建立自己的核心路由表并转发数据，每个中继节点建立自己的核心路由表同时转发数据包，直到将数据包发送到目的节点，发送数据包的骨干节点为源节点，源节点到目的节点最短路径上经历的节点为中继节点，其特征在于：改进CLUSTERPOW协议方法，在一个数据包的发送过程中仅在源节点处根据功率等级路由表执行一次Dijkstra算法查找到达目的节点的功耗最小路径并将该路径上的节点ID号保存到核心路由表，以避免路由环路的出现，同时降低发射功率消耗，包括以下步骤：The technical solution adopted by the present invention: a method of inter-cluster virtual backbone routing protocol based on power control, according to the CLUSTERPOW protocol method, using the method of negotiating the routing table to find the best route in the backbone network, and combining the power control technology to select the number of power levels And use the handshake protocol to establish a corresponding number of local power level routing tables and global power level routing tables, and adaptively adjust to the appropriate transmission power for different numbers of nodes to achieve routing. In order to find a power-optimized route, each backbone node forwards A core routing table must be established for data packets. The core routing table provides the shortest route to the destination node. After negotiating the local routing table, the core routing table selects the minimum power route and copies it to the core routing table. The backbone The node adjusts the transmission power to the corresponding level according to the core routing table, and then forwards the data to the next-hop relay node. When the next-hop relay node forwards the data, it establishes its own core routing table and forwards the data. Each relay node Build your own core routing table and forward the data packet at the same time until the data packet is sent to the destination node. The backbone node sending the data packet is the source node, and the node experienced on the shortest path from the source node to the destination node is a relay node. It is characterized in that: Improve the CLUSTERPOW protocol method, and only execute the Dijkstra algorithm once at the source node according to the power level routing table in the process of sending a data packet to find the path with the least power consumption to the destination node and save the node ID number on the path to the core routing table , to avoid the occurrence of routing loops, while reducing transmission power consumption, including the following steps:

1）基于CLUSTERPOW协议，结合功率控制技术选取的功率等级数目，每个区域中的骨干节点利用握手协议建立相应数目的本地功率等级路由表和全局功率等级路由表；1) Based on the CLUSTERPOW protocol, combined with the number of power levels selected by the power control technology, the backbone nodes in each area use the handshake protocol to establish a corresponding number of local power level routing tables and global power level routing tables;

2）当骨干节点有数据包需要发送时，该骨干节点根据先前保存的各功率等级路由表执行Dijkstra算法查找到达目的节点的功耗最小路径并将该路径上的节点ID号保存到核心路由表；2) When the backbone node has a data packet to send, the backbone node executes the Dijkstra algorithm according to the previously saved routing table of each power level to find the path with the lowest power consumption to the destination node and saves the node ID number on the path to the core routing table ;

3）依据DSR协议，数据包的源节点将核心路由表封装到数据包的头部，并将发射功率调整到相应功率等级，然后将数据包发送到下一跳中继节点；3) According to the DSR protocol, the source node of the data packet encapsulates the core routing table into the head of the data packet, adjusts the transmission power to the corresponding power level, and then sends the data packet to the next-hop relay node;

4）下一跳中继节点在收到数据包时首先向上一跳中继节点发送回复确认包，然后对收到的数据包进行解封装，判断自己是否是目的节点，若是目的节点，则转到步骤6），若不是目的节点，则转到步骤5）；4) When the next-hop relay node receives the data packet, it first sends a reply confirmation packet to the previous-hop relay node, and then decapsulates the received data packet to determine whether it is the destination node. Go to step 6), if it is not the destination node, then go to step 5);

5）取出数据包核心路由表中第一个中继节点ID号码作为自己的下一跳中继节点ID号，并将核心路由表中的剩余路由信息存储到自己的核心路由表，同样地将自己的核心路由表封装到数据包的头部，最后根据自己的功率等级路由表将发射功率调节到相应功率等级，并将该数据包转发给自己的下一跳中继节点；5) Take out the first relay node ID number in the core routing table of the data packet as its own next-hop relay node ID number, and store the remaining routing information in the core routing table in its own core routing table. Its own core routing table is encapsulated into the head of the data packet, and finally the transmission power is adjusted to the corresponding power level according to its own power level routing table, and the data packet is forwarded to its own next-hop relay node;

6）每个中继节点建立自己的核心路由表同时转发数据包，中继节点如果收到回复包，则说明数据转发成功，如果没有收到回复确认包，表明先前的路由失效，该中继节点需要根据自己的各功率等级路由表执行Dijkstra算法查找到达目的节点的功耗最小路径并更新自己的核心路由表，将更新后的路由表封装到数据包头部并发送给下一跳中继节点；6) Each relay node establishes its own core routing table and forwards the data packet at the same time. If the relay node receives the reply packet, it means that the data forwarding is successful. If it does not receive the reply confirmation packet, it indicates that the previous route is invalid. The node needs to execute the Dijkstra algorithm according to its own routing table of each power level to find the path with the least power consumption to the destination node and update its own core routing table, encapsulate the updated routing table into the header of the data packet and send it to the next hop relay node ;

7）下一跳节点重复执行步骤4），直到数据包到达目的结点；7) The next hop node repeats step 4) until the data packet reaches the destination node;

8）数据包发送结束。8) The packet sending ends.

本发明方法的优点和显着效果：Advantage and remarkable effect of the inventive method:

1）在路由查找过程中，结合Dijkstra算法根据目的节点自适应地选取功率消耗最小的路径。1) In the process of route search, combined with Dijkstra algorithm, the path with the least power consumption is adaptively selected according to the destination node.

2）将所查到的最优路径封装到数据包中，有效避免了路由环路的出现。2) Encapsulate the found optimal path into a data packet, effectively avoiding the occurrence of routing loops.

3）在一个数据包的发送过程中，协议仅在源节点处被执行，其他转发节点无需执行复杂的路由查找协议，这一定程度上降低了节点的数据处理等开销。3) During the sending process of a data packet, the protocol is only executed at the source node, and other forwarding nodes do not need to execute complex routing lookup protocols, which reduces the data processing and other overhead of nodes to a certain extent.

附图说明:Description of the drawings:

图1是骨干节点X₁到X_N的最优路径；Fig. 1 is the optimal path from backbone node _X1 to _XN ;

图2是CLUSTERPOW协议的数据包格式；Figure 2 is the packet format of the CLUSTERPOW protocol;

图3是改进的CLUSTERPOW协议的数据包格式；Fig. 3 is the packet format of improved CLUSTERPOW agreement;

图4是两协议单个数据包能量消耗差值随功率等级数和路径上节点数目的变化曲线；Fig. 4 is the change curve of the energy consumption difference of a single packet of the two protocols with the number of power levels and the number of nodes on the path;

图5是网络仿真采用的拓扑结构图；Fig. 5 is the topological structure diagram that network emulation adopts;

图6是两协议在不同功率等级数目下单个数据包平均能量消耗的对比；Figure 6 is a comparison of the average energy consumption of a single packet of the two protocols at different power levels;

图7是两协议在高数据速率下单个数据包能量消耗随负载变化的对比；Figure 7 is a comparison of the energy consumption of a single data packet changing with the load of the two protocols at a high data rate;

图8是两协议在低数据速率下单个数据包能量消耗随负载变化的对比；Figure 8 is a comparison of the energy consumption of a single data packet with the load of the two protocols at a low data rate;

图9是两协议在数据包信息位长度取不同值时单个数据包平均能量消耗随负载变化的对比；Figure 9 is a comparison of the average energy consumption of a single data packet with the change of load when the data packet information bit length of the two protocols takes different values;

图10是水池实验监控界面；Figure 10 is the pool experiment monitoring interface;

图11是水池实验中两协议单数据包的平均能量消耗比较。Figure 11 is a comparison of the average energy consumption of a single data packet of the two protocols in the water pool experiment.

具体实施案例：Specific implementation cases:

为更进一步阐述本发明为达成预定发明目的所采取的方法,以下结合实施例，对本发明方法进行详细说明。In order to further illustrate the method adopted by the present invention to achieve the intended purpose of the invention, the method of the present invention will be described in detail below in conjunction with the examples.

实施例1：Example 1:

1、为了准确估计改进的CLUSTERPOW协议的能耗性能，下面从理论上推算一下改进的CLUSTERPOW协议和CLUSTERPOW协议的能量消耗情况。1. In order to accurately estimate the energy consumption performance of the improved CLUSTERPOW protocol, the following theoretically calculates the energy consumption of the improved CLUSTERPOW protocol and the CLUSTERPOW protocol.

如图1所示，假设X₁要向X_N发送数据包，假设用Dijkstra算法查找到功耗最小路径为X₁-X₂-X₃-X₄……X_N-1-X_N。考虑到两协议的发射功耗相差最小情况下CLUSTERPOW协议和改进的CLUSTERPOW协议仅在第一跳即X₁-X₂时存在发射功率等级差异，，假设路径上的其他发射功率等级如图1中所示，其中P₁和P₂为相邻的两个功率等级，且有P₁≤P₂，那么在X₁-X₂这条路径上CLUSTERPOW协议的发射功率为P₂，改进的CLUSTERPOW协议的发射功率为P₁。CLUSTERPOW协议和改进的CLUSTERPOW协议数据包格式如图2和图3所示，那么两协议发送一个数据包的总能量消耗E₁和E₂如式（1）和式（2）所示（此处忽略确认包）。As shown in Figure 1, suppose that X ₁ wants to send data packets to X _N , and assume that the path with the minimum power consumption found by Dijkstra algorithm is X ₁ -X ₂ -X ₃ -X ₄ ... X _N-1 -X _N . Considering that the transmission power consumption difference of the two protocols is the smallest, the CLUSTERPOW protocol and the improved CLUSTERPOW protocol only have a difference in transmission power level at the first hop, that is, X ₁ -X ₂ , assuming that other transmission power levels on the path are shown in Figure 1 As shown, where P ₁ and P ₂ are two adjacent power levels, and P ₁ ≤ P ₂ , then the transmission power of the CLUSTERPOW protocol on the path X ₁ -X ₂ is P ₂ , the improved CLUSTERPOW protocol The transmit power is P ₁ . The CLUSTERPOW protocol and the improved CLUSTERPOW protocol packet format are shown in Figure 2 and Figure 3, then the total energy consumption E ₁ and E ₂ of sending a data packet by the two protocols are shown in formula (1) and formula (2) (here Ignore the acknowledgment packet).

$\begin{matrix} {E E.}_{11} = = {P P}_{22} \times \times \frac{D D.}{{R R}_{b b}} + + {P P}_{rcv rcv} \times \times \frac{D D.}{{R R}_{b b}} + + {P P}_{22} \times \times \frac{D D.}{{R R}_{b b}} + + {P P}_{rcv rcv} \times \times \frac{D D.}{{R R}_{b b}} + + {P P}_{11} \times \times \frac{D D.}{{R R}_{b b}} \\ + + {P P}_{rcv rcv} \times \times \frac{D D.}{{R R}_{b b}} + + {P P}_{11} \times \times \frac{D D.}{{R R}_{b b}} + + . . . . . . + + {P P}_{rcv rcv} \times \times \frac{D D.}{{R R}_{b b}} + + {P P}_{11} \times \times \frac{D D.}{{R R}_{b b}} + + {P P}_{rcv rcv} \times \times \frac{D D.}{{R R}_{b b}} \\ = = \frac{D D.}{{R R}_{b b}} [[22 {P P}_{22} + + ((N N - - 33)) {P P}_{11} + + ((N N - - 11)) {P P}_{rcv rcv}]] \end{matrix} - - - - - - ((11))$

$\begin{matrix} {E E.}_{22} = = {P P}_{11} \times \times \frac{D D. + + ((N N - - 22)) {N N}_{b b}}{{R R}_{b b}} + + {P P}_{rcv rcv} \times \times \frac{D D. + + ((N N - - 22)) {N N}_{b b}}{{R R}_{b b}} + + {P P}_{22} \times \times \frac{D D. + + ((N N - - 33)) {N N}_{b b}}{{R R}_{b b}} \\ + + {P P}_{rcv rcv} \times \times \frac{D D. + + ((N N - - 33)) {N N}_{b b}}{{R R}_{b b}} + + {P P}_{11} \times \times \frac{D D. + + ((N N - - 44)) {N N}_{b b}}{{R R}_{b b}} + + {P P}_{rcv rcv} \times \times \frac{D D. + + ((N N - - 44)) {N N}_{b b}}{{R R}_{b b}} \\ {P P}_{11} \times \times \frac{D D. + + ((N N - - 55)) {N N}_{b b}}{{R R}_{b b}} + + . . . . . . + + {P P}_{rcv rcv} \times \times \frac{D D. + + {N N}_{b b}}{{R R}_{b b}} + + {P P}_{11} \times \times \frac{D D.}{{R R}_{b b}} + + {P P}_{rcv rcv} \times \times \frac{D D.}{{R R}_{b b}} \\ = = \frac{D D.}{{R R}_{b b}} [[((N N - - 22)) {P P}_{11} + + {P P}_{22} + + ((N N - - 11)) {P P}_{rcv rcv}]] + + \\ \frac{{N N}_{b b}}{{R R}_{b b}} [[\frac{((N N - - 11)) ((N N - - 22))}{22} (({P P}_{11} + + {P P}_{rcv rcv})) + + ((N N - - 33)) (({P P}_{22} - - {P P}_{11}))]] \end{matrix} - - - - - - ((22))$

其中，D为CLUSTERPOW协议的一个数据包大小，R_b为发射数据率，P_rcv为传感器节点接收功率，N_b为数据包中节点ID号所占有的比特数。那么CLUSTERPOW协议和改进的CLUSTERPOW协议发送一个数据包总共所消耗的能量差值E为：Among them, D is the size of a data packet of CLUSTERPOW protocol, R _b is the transmission data rate, P _rcv is the sensor node receiving power, N _b is the number of bits occupied by the node ID number in the data packet. Then the total energy difference E consumed by the CLUSTERPOW protocol and the improved CLUSTERPOW protocol to send a data packet is:

E＝E₁-E₂ (3)E＝E ₁ -E ₂ (3)

将式（1）和式（2）带入式（3）可得Put formula (1) and formula (2) into formula (3) to get

$E E. = = \frac{((D D. - - N N + + 33)) (({P P}_{22} - - {P P}_{11})) - - {N N}_{b b} \frac{((N N - - 22)) ((N N - - 11))}{22} (({P P}_{11} + + {P P}_{rcv rcv}))}{{R R}_{b b}} - - - - - - ((44))$

CLUSTERPOW协议和改进的CLUSTERPOW协议发送一个数据包总共所消耗的能量差值E与D、功率等级P₁和P₂、路径上节点数N、R_b、P_rcv以及N_b有关。The total energy difference E consumed by the CLUSTERPOW protocol and the improved CLUSTERPOW protocol to send a data packet is related to D, power levels P ₁ and P ₂ , the number of nodes on the path N, R _b , P _rcv and N _b .

由式（4）易于看出，对于水声通信网络来说，多数应用场合下会有E＞0，即改进的CLUSTERPOW协议的能量消耗会低于CLUSTERPOW协议。这是因为：对于水声通信网络来说，水声信道的特殊性以及水下传感器节点昂贵的硬件成本决定了通常情况下网络的规模不会很大，尤其是在分簇式网络里，簇头的数目不会很多，在簇间通信过程中，数据转发跳数一般较少，所以N的取值一般也不会很大；数据包头中路由控制信息的存在使得D与N的差值较大；考虑到实现的复杂性水声通信网络中的功率等级数目也不会很大且通常发射功率比接收功率大得多，因此P₂与P₁的差距相对来说不会很小。It is easy to see from formula (4) that for underwater acoustic communication networks, E>0 in most applications, that is, the energy consumption of the improved CLUSTERPOW protocol will be lower than that of the CLUSTERPOW protocol. This is because: for the underwater acoustic communication network, the particularity of the underwater acoustic channel and the expensive hardware cost of the underwater sensor nodes determine that the size of the network is usually not very large, especially in the clustered network, the cluster The number of headers will not be very large. In the process of inter-cluster communication, the number of data forwarding hops is generally small, so the value of N is generally not very large; the existence of routing control information in the data packet header makes the difference between D and N relatively small. Large; considering the complexity of the implementation, the number of power levels in the underwater acoustic communication network will not be very large and usually the transmit power is much larger than the receive power, so the gap between P ₂ and P ₁ will not be relatively small.

2、当D=512bit、R_b=320bit/s、P_rcv=0.9W、N_b=8bit、网络中相邻节点间的最远距离为50km、最短距离为18km时，E随功率等级数GradeNum和N的变化情况如图4所示。由图中可以看出，功率等级数目越小，能量差值越大，N越小即路由跳数越少，能量差值亦越大。并且不同的功率等级数目下，当N值较小时总有E＞0。由于分簇式网络里N的取值通常不会很大，因此改进的CLUSTERPOW协议在分簇式网络的簇间虚拟骨干路由协议上会表现出一定的优越性能。2. When D=512bit, R _b =320bit/s, P _rcv =0.9W, N _b =8bit, the furthest distance between adjacent nodes in the network is 50km, and the shortest distance is 18km, E varies with the power grade GradeNum And the change of N is shown in Fig. 4. It can be seen from the figure that the smaller the number of power levels, the greater the energy difference, and the smaller N means the fewer routing hops, the greater the energy difference. And under different numbers of power levels, when the value of N is small, there is always E>0. Because the value of N in the clustered network is usually not very large, the improved CLUSTERPOW protocol will show certain superior performance in the inter-cluster virtual backbone routing protocol of the clustered network.

实施例2：Example 2:

在OPNETModeler14.5平台上对CLUSTERPOW协议和改进的CLUSTERPOW协议进行仿真比较。The CLUSTERPOW protocol and the improved CLUSTERPOW protocol are simulated and compared on the OPNETModeler14.5 platform.

参看图5，为网络仿真采用的拓扑结构图，在100km×100km的仿真场景中随机布放了50个节点，其中10个簇头节点和40个普通节点。仿真中骨干节点仅由所有的簇头节点来担当。仿真中每个节点产生自己的数据包，如果节点是普通节点，先将数据包发送到自己的簇头节点，经簇头节点转发。如果目的节点是普通节点，则先将数据包发送到其簇头节点进行转发。MAC层采用简单的载波检测（侦听）多路访问CSMA（CarrierSenseMultipleAccess）协议。主要仿真参数如下：传播速度为1500m/s；中心频率为10kHz；数据包中节点ID号所占有的比特数为8bit；节点数据速率和数据包中信息位长度均在仿真中按需设定。仿真事务流的包产生时间间隔服从均匀分布，包到达时间间隔均值亦在仿真中按需设定。节点的负载包括输入负载以及中继的数据流。通过仿真分别得到了CLUSTERPOW协议和改进的CLUSTERPOW协议的单个数据包平均能量消耗随功率等级数目、网络平均负载以及数据包中信息位长度的变化对比图6～9。Referring to Figure 5, the topology diagram used for network simulation, 50 nodes are randomly placed in the simulation scene of 100km×100km, including 10 cluster head nodes and 40 ordinary nodes. In the simulation, the backbone nodes are only played by all the cluster head nodes. In the simulation, each node generates its own data packet. If the node is an ordinary node, the data packet is first sent to its own cluster head node and forwarded by the cluster head node. If the destination node is an ordinary node, it will first send the data packet to its cluster head node for forwarding. The MAC layer uses a simple carrier detection (sense) multiple access CSMA (CarrierSenseMultipleAccess) protocol. The main simulation parameters are as follows: the propagation speed is 1500m/s; the center frequency is 10kHz; the number of bits occupied by the node ID number in the data packet is 8bit; the node data rate and the length of information bits in the data packet are set as required in the simulation. The packet generation time interval of the simulated transaction flow is uniformly distributed, and the average value of the packet arrival time interval is also set as required in the simulation. The load of the node includes the input load and the data flow of the relay. Through simulation, the average energy consumption of a single data packet of the CLUSTERPOW protocol and the improved CLUSTERPOW protocol are compared with the number of power levels, the average load of the network, and the length of information bits in the data packet. Figure 6-9.

参看图6，为两协议在不同功率等级数目下单个数据包平均能量消耗的对比。从图中可以看出CLUSTERPOW协议和改进的CLUSTERPOW协议的单个数据包平均能量消耗均随着功率等级数目的增大而逐渐减小。当功率等级数大于等于2时，改进的CLUSTERPOW协议的单个数据包能量消耗明显低于CLUSTERPOW协议。这是因为采用更多的功率等级可使通信系统更有效、更精确地调节功率的分配，这样节点就可以根据传输需要选用合适的功率发送数据，从而减少总的能量消耗和节点干扰，但是使用太多的功率等级并不实际，这会增加系统设计的复杂度。并且从图中可以看出，当等级数大于3时，两种协议均不能显著地改进系统性能，另外结合图2并通过综合考虑单个数据包平均能量消耗和实现复杂性，功率等级数采用3和4均是比较合适的。Referring to FIG. 6 , it is a comparison of the average energy consumption of a single data packet of the two protocols under different numbers of power levels. It can be seen from the figure that the average energy consumption of a single packet of the CLUSTERPOW protocol and the improved CLUSTERPOW protocol decreases gradually with the increase of the number of power levels. When the number of power levels is greater than or equal to 2, the energy consumption of a single packet of the improved CLUSTERPOW protocol is significantly lower than that of the CLUSTERPOW protocol. This is because the use of more power levels can make the communication system adjust the power allocation more effectively and accurately, so that the nodes can choose the appropriate power to send data according to the transmission needs, thereby reducing the total energy consumption and node interference, but using Too many power levels are impractical and add complexity to the system design. And it can be seen from the figure that when the number of levels is greater than 3, neither of the two protocols can significantly improve system performance. In addition, combined with Figure 2 and by comprehensively considering the average energy consumption and implementation complexity of a single data packet, the number of power levels is 3 and 4 are more appropriate.

参看图7，为两协议在高数据速率下单个数据包能量消耗随负载变化的对比，此时仿真中设定数据速率为320bit/s，仿真事务流的包到达时间间隔均值分别0.2秒，0.5秒，0.9秒，20秒，90秒，150秒，200秒，300秒。每组仿真的时间设为2小时，每组仿真运行5次，取平均得到仿真结果。由图可以看出两种协议的单个数据包平均能量消耗均随着负载的增大逐渐增大，最后趋于平稳。CLUSTERPOW协议为避免产生路由回路，在数据包的一次发送过程中使用非递增的发射功率，这在一定程度上造成了发射功率的浪费即造成了能量的浪费，而改进的CLUSTERPOW协议解决了这一缺陷，使用功耗最优路径，故单个数据包平均能量消耗明显比CLUSTERPOW协议减小了很多。Referring to Figure 7, it is a comparison of the energy consumption of a single data packet with the load of the two protocols at a high data rate. At this time, the data rate is set to 320bit/s in the simulation, and the mean value of the packet arrival time interval of the simulated transaction flow is 0.2 seconds and 0.5 seconds respectively. seconds, 0.9 seconds, 20 seconds, 90 seconds, 150 seconds, 200 seconds, 300 seconds. The time of each group of simulation is set to 2 hours, and each group of simulation runs 5 times, and the simulation results are obtained by taking the average. It can be seen from the figure that the average energy consumption of a single data packet of the two protocols gradually increases with the increase of the load, and finally tends to be stable. In order to avoid routing loops, the CLUSTERPOW protocol uses a non-increasing transmission power during one transmission of a data packet, which to a certain extent causes a waste of transmission power, that is, a waste of energy. The improved CLUSTERPOW protocol solves this problem. The defect is that the optimal path of power consumption is used, so the average energy consumption of a single data packet is significantly lower than that of the CLUSTERPOW protocol.

参看图8，为两协议在低数据速率下单个数据包能量消耗随负载变化的对比，此时仿真中设定数据速率为80bit/s，仿真事务流的包到达时间间隔均值分别6秒，7秒，8秒，9秒，10秒，13秒，15秒，30秒,45秒,60秒,75秒,90秒,105秒,120秒。由图可以看出，低数据率下，两种协议的单个数据包平均能量消耗仍然随着负载的增大均逐渐增大，最后趋于平稳，但改进的CLUSTERPOW协议的单个数据包能量消耗仍然低于CLUSTERPOW协议。与图7相比，低数据率下，两种协议的能量差距加大，这是因为发射时间增大，发射能量消耗增大Referring to Figure 8, it is a comparison of the energy consumption of a single data packet with the load of the two protocols at a low data rate. At this time, the data rate is set to 80bit/s in the simulation, and the mean value of the packet arrival time interval of the simulated transaction flow is 6 seconds, 7 seconds, respectively. Seconds, 8 seconds, 9 seconds, 10 seconds, 13 seconds, 15 seconds, 30 seconds, 45 seconds, 60 seconds, 75 seconds, 90 seconds, 105 seconds, 120 seconds. It can be seen from the figure that at low data rates, the average energy consumption of a single data packet of the two protocols still gradually increases with the increase of the load, and finally tends to be stable, but the energy consumption of a single data packet of the improved CLUSTERPOW protocol is still Below the CLUSTERPOW protocol. Compared with Figure 7, at low data rates, the energy gap between the two protocols increases, because the transmission time increases and the transmission energy consumption increases

参看图9，为两协议在数据包信息位长度取不同值时单个数据包平均能量消耗随负载变化的对比，此时仿真中设定信息位长度分别为256bit,512bit和1024bit。从图（改进前表示CLUTERPOW协议，改进后表示改进的CLUTERPOW协议）中可以看出数据包信息位长度在三种取值下，改进的CLUTERPOW协议的单数据包能量消耗均小于CLUTERPOW协议，并且随着数据信息位长度的增加两者的差值也逐渐增大。Referring to Figure 9, it shows the comparison of the average energy consumption of a single data packet with the load when the two protocols take different values of the data packet information bit length. At this time, the information bit lengths are set to 256bit, 512bit and 1024bit in the simulation. It can be seen from the figure (the CLUTERPOW protocol before the improvement, and the improved CLUTERPOW protocol after the improvement) that the data packet information bit length is under three values, and the energy consumption of a single data packet of the improved CLUTERPOW protocol is less than that of the CLUTERPOW protocol. As the length of the data information bit increases, the difference between the two gradually increases.

参看图10，为水池实验监控界面，实验中仅采用了四个modem作为簇头节点，而簇成员节点采用虚拟节点来模拟真实modem，每个簇有三个成员节点，图中簇头0和簇头3正在进行通信。Refer to Figure 10, which is the monitoring interface of the pool experiment. In the experiment, only four modems are used as cluster head nodes, while cluster member nodes use virtual nodes to simulate real modems. Each cluster has three member nodes. In the figure, cluster head 0 and cluster Head 3 is communicating.

参见图11，为水池实验中两协议单数据包的平均能量消耗比较。由图可以看出，在某些特定轮上改进的CLUSTERPOW协议的能量消耗低于CLUSTERPOW协议，这是因为CLUSTERPOW协议为避免路由环路使用非递增的功率等级，这一定程度上造成了能源的浪费，而改进的CLUSTERPOW协议选取合适的发射功率等级，节省了能量。See Figure 11, which is the comparison of the average energy consumption of a single data packet of the two protocols in the water pool experiment. It can be seen from the figure that the energy consumption of the improved CLUSTERPOW protocol on some specific rounds is lower than that of the CLUSTERPOW protocol. This is because the CLUSTERPOW protocol uses non-increasing power levels to avoid routing loops, which to some extent causes energy waste. , and the improved CLUSTERPOW protocol selects the appropriate transmit power level to save energy.

综合以上所有的仿真结果分析，可以得出与CLUSTERPOW协议相比，改进的CLUSTERPOW协议减小了网络的能量消耗，并进而延长了网络的生命周期。Based on the analysis of all the above simulation results, it can be concluded that compared with the CLUSTERPOW protocol, the improved CLUSTERPOW protocol reduces the energy consumption of the network, and thus prolongs the life cycle of the network.

Claims

1. A power control-based inter-cluster virtual backbone routing protocol method, based on the CLUSTERPOW protocol method, using the method of negotiating routing tables to find the best route in the backbone network, combining power control technology to select the number of power levels and using the handshake protocol to establish The corresponding number of local power level routing tables and global power level routing tables are adaptively adjusted to the appropriate transmission power for different numbers of nodes to achieve routing. In order to find the power-optimized route, each backbone node must establish A core routing table. The core routing table provides the shortest route to the destination node. After the core routing table negotiates the local routing table, it selects the minimum power route and copies it to the core routing table. Adjust the transmission power to the corresponding level, and then forward the data to the next-hop relay node. When forwarding the data, the next-hop relay node establishes its own core routing table and forwards the data. Each relay node establishes its own core routing table The table forwards the data packet at the same time until the data packet is sent to the destination node, the backbone node sending the data packet is the source node, and the node experienced on the shortest path from the source node to the destination node is a relay node, and it is characterized in that: the improved CLUSTERPOW protocol method, In the process of sending a data packet, the Dijkstra algorithm is only executed once at the source node according to the power level routing table to find the path with the least power consumption to the destination node and save the node ID number on the path to the core routing table to avoid routing loops The emergence of roads, while reducing transmission power consumption, including the following steps:

1) Based on the CLUSTERPOW protocol, combined with the number of power levels selected by the power control technology, the backbone nodes in each area use the handshake protocol to establish a corresponding number of local power level routing tables and global power level routing tables;

2) When the backbone node has a data packet to send, the backbone node executes the Dijkstra algorithm according to the previously saved routing table of each power level to find the path with the lowest power consumption to the destination node and saves the node ID number on the path to the core routing table ;

3) According to the DSR protocol, the source node of the data packet encapsulates the core routing table into the head of the data packet, adjusts the transmission power to the corresponding power level, and then sends the data packet to the next-hop relay node;

4) When the next-hop relay node receives the data packet, it first sends a reply confirmation packet to the previous-hop relay node, and then decapsulates the received data packet to determine whether it is the destination node. Go to step 6), if it is not the destination node, then go to step 5);

5) Take out the first relay node ID number in the core routing table of the data packet as its own next-hop relay node ID number, and store the remaining routing information in the core routing table in its own core routing table. Its own core routing table is encapsulated into the head of the data packet, and finally the transmission power is adjusted to the corresponding power level according to its own power level routing table, and the data packet is forwarded to its own next-hop relay node;

6) Each relay node establishes its own core routing table and forwards the data packet at the same time. If the relay node receives the reply packet, it means that the data forwarding is successful. If it does not receive the reply confirmation packet, it indicates that the previous route is invalid. The node needs to execute the Dijkstra algorithm according to its own routing table of each power level to find the path with the least power consumption to the destination node and update its own core routing table, encapsulate the updated routing table into the header of the data packet and send it to the next hop relay node ;

7) The next hop node repeats step 4) until the data packet reaches the destination node;

8) Data packet sending ends.