Topology repair method for underwater wireless sensor networks suitable for dynamic water flow scenarios

Abstract

The present invention provides an underwater wireless sensor network topology repair method suitable for dynamic water flow scenarios; the repair method comprises two stages: repair path establishment and repair node arrangement; first: a repair path generation method based on path clipping, according to the location distribution characteristics of all underwater sub-networks that are not interconnected, constructs a repair path connecting all sub-networks and a surface aggregation node; second: an adaptive underwater repair node arrangement method based on greedy mobility, so that the number of deployed repair nodes is reduced on the premise that the links between nodes meet the transmission success rate requirements; after the above two steps, the topology of the underwater wireless sensor network can be repaired; the method can perform topological connectivity repair for the underwater wireless sensor network according to the underwater communication channel and the water flow movement characteristics, and can improve the transmission reliability while reducing the number of deployed repair nodes.

Description

Technical field:

[0001] The present invention mainly relates to the technical field of underwater wireless sensor networks, and in particular to a topology repair method for underwater wireless sensor networks suitable for dynamic water flow scenarios. Background technology:

[0002] Underwater wireless sensor networks are one of the important ways to obtain ocean data. They can provide the required ocean data for various marine application scenarios in real time and conveniently, such as tsunami warning, fish monitoring, seabed resource exploration, assisted navigation, and underwater intrusion detection. Underwater wireless sensor networks usually include four links: underwater data perception, underwater data transmission, surface data aggregation, and data processing and application. In a typical underwater data generation and underwater transmission process, underwater sensor nodes first perceive valuable underwater information data, and then transmit these data to the surface aggregation node through multi-hop transmission or underwater mobile node assisted collection. Since underwater mobile nodes usually move slowly, multi-hop transmission is more suitable for the transmission of delay-sensitive data.

[0003] The application scenario of the present invention is an underwater wireless sensor network with a multi-hop transmission mode. Many underwater wireless sensor networks are usually deployed in a complex and changeable marine environment without on-site supervision. The deliberate destruction by marine environmental factors or hostile forces can easily cause a large number of underwater sensor nodes to fail, thereby causing the entire underwater wireless sensor network to be divided into several unconnected sub-networks, which will cause many valuable and delay-sensitive underwater data to no longer be sent to the surface aggregation node in a timely manner. How to repair the unconnected sub-networks that have been damaged so that they can reliably send data to the surface aggregation node is a very challenging task. There are two main reasons. First, low-quality underwater wireless channels are prone to data retransmission problems during the forwarding process, which seriously affects data efficiency and success rate. Secondly, the movement of underwater nodes caused by ocean currents can easily interfere with the topology repair process, destroy the network topology connectivity, and reduce the reliability of data transmission in the underwater wireless sensor network.

[0004] One of the common methods for underwater wireless sensor network topology repair is to deploy additional underwater repair nodes to connect all disconnected sub-networks and surface aggregation nodes, so that underwater data can continue to be reliably sent to the surface aggregation nodes. Through relevant research, the existing underwater wireless sensor network topology repair strategy mainly considers how to deploy the minimum number of underwater repair nodes to connect all disconnected underwater sub-networks, lacks consideration of the poor communication quality caused by underwater channel characteristics and the impact of relay node position offset caused by dynamic water flow on transmission reliability. For example, the journal paper "ATCFS: Effective Connectivity Restoration Scheme for Underwater Acoustic Sensor Networks" proposes a connectivity repair strategy for underwater wireless sensor networks with the purpose of minimizing deployment costs (deploying the minimum number of underwater repair nodes). This paper designs a heuristic strategy based on minimum spanning tree and Fermat point selection. During the operation of this strategy, the surface repair node sinks to a specified depth to repair the network topology connectivity. However, the network model of this strategy is simple and does not consider the impact of underwater channel quality and water flow movement on deployment cost and deployment effect, which can easily lead to problems such as poor transmission reliability of the repaired network and high deployment cost of repair nodes. In view of the problems existing in the topology repair strategy of existing underwater wireless sensor networks, the present invention proposes an underwater wireless sensor network topology repair method suitable for dynamic water flow scenarios; it can effectively ensure network transmission reliability and reduce deployment costs. Summary of the invention:

[0005] In order to solve the problems of the prior art, the present invention provides an underwater wireless sensor network topology repair method suitable for dynamic water flow scenarios. It can deploy as few repair nodes as possible by analyzing the location distribution characteristics of underwater sub-networks and the movement characteristics of ocean currents, so that the network can be restored to connectivity while having reliable data transmission performance.

[0006] The present invention adopts an underwater mobile device with underwater omnidirectional mobility as an underwater repair node to repair the network topology. The present invention includes two stages: repair path establishment and repair node layout. First, in the repair path establishment stage, the present invention designs a repair path generation method based on path clipping, and constructs a repair path connecting all subnetworks and surface aggregation nodes according to the location distribution characteristics of the underwater subnetwork. This path can not only repair the connectivity from the underwater subnetwork to the surface aggregation node, but also effectively reduce the number of transmission hops from the subnetwork to the surface aggregation node. Then in the repair node layout stage, the present invention designs an adaptive underwater repair node layout method based on greedy mobility, so that the number of repair node deployments is reduced under the premise that the inter-node link meets the transmission success rate requirements. After the above two steps, the underwater wireless sensor network topology can be repaired. If the marine environment mutates, the network topology can be repaired again according to the above steps.

[0007] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0008] The method for repairing underwater wireless sensor network topology applicable to dynamic water flow scenarios includes the following steps:

[0009] Step 101: Select the cluster head node of each underwater sub-network and set the connection state of all cluster head nodes to be activated;

[0010] Step 102: Determine whether the transmission from each underwater sensor node to the surface aggregation node is reliable within the time period t. If the transmission is reliable, the underwater sensor node replaces the original cluster head node of the underwater sub-network where it is located to become the new cluster head node of the underwater sub-network, and changes its own connection state to a dormant state; otherwise, proceed to the next step 103;

[0011] Step 103: Set the sink node as the target point, and define an empty set as the execution set of the target point; store all the cluster head nodes in the active state into the target point execution set, and sort the cluster head nodes of the target point execution set from large to small according to their respective Euclidean distances to the target point;

[0012] Step 104: Select the cluster head node that ranks the top and is in an activated state in the target point execution set as the current execution node, and define two empty sets as the candidate follower set and the follower set of the execution node respectively; store the other cluster head nodes in the target point execution set in the candidate follower set of the current execution node, and sort the cluster head nodes of the candidate follower set of the current execution node from small to large according to their respective Euclidean distances to the execution node;

[0013] Step 105: Process of sequentially determining whether each cluster head node in the candidate follower set of the current execution node can be added to the follower set of this execution node:

[0014] Deposit the current execution node as the first element into its follower set;

[0015] Select the cluster head node at the forefront of the candidate follower set as the candidate node, and calculate the median point of the data set composed of all elements in the follower set of the current execution node and the target point;

[0016] Calculate the sum of distances from all elements of this data set to this median point, and denote it as d1, calculate the distance from the candidate node to the sink node, and denote it as d2, calculate the median point of the data set composed of all elements in the follower set of the current execution node EN, the target point G, and the candidate node CA, and calculate the sum of distances from all elements of this data set to this median point, and denote it as d3;

[0017] Compare the relationship among d1, d2, and d3. If d3 < d1 + d2, then add this candidate node to the follower set of the current execution node, and delete this candidate node from the execution set;

[0018] Whether or not it is added to the follower set of the current execution node, this candidate node is deleted from the candidate follower set of the current execution node; repeat this step until the candidate follower set of the current execution node becomes an empty set;

[0019] Step 106: Repeat steps 104 - 105 until the execution set becomes an empty set;

[0020] Step 107: Process of constructing a topology repair path based on the cluster head node information in each follower set;

[0021] Step 108: For any follower set, use its corresponding median point as the new target point, and turn this follower set into the execution set;

[0022] Step 109: Execute the above steps 103 - 108 until the execution set corresponding to this follower set becomes an empty set;

[0023] Step 110: Repeat step 109 until the execution sets corresponding to all follower sets become empty sets;

[0024] Step 111: Repeat the above steps 109 - 110 until the connection status of all CNs is in the sleep state; deposit all the target points generated in the above steps into the empty set of the branch point set.

[0025] Furthermore, the process of constructing a topology repair path according to the cluster head node information in each follower set in step 107 includes the following steps:

[0026] Calculate the median point of the data set consisting of all cluster head nodes and the target point in the follower set with the number of cluster head nodes greater than 1; connect the target point G and the point P of each follower set to construct a repair path;

[0027] Determine whether the arrival point of each cluster head node in each follower set to the median point of its follower set is reliable within the time period t. If the transmission is reliable, delete the cluster head node from the corresponding follower set, change the connection state of the cluster head node to the dormant state, connect the dormant cluster head node and the median point of the corresponding follower set, and build a repair path.

[0028] Calculate the number of cluster head nodes in the current follower set. If it is 1, change the connection state of the cluster head node in the follower set to a dormant state; connect the dormant cluster head node to the median point of the corresponding follower set to construct a repair path;

[0029] Among them: in the process of judging whether the transmission of the midpoint of the arrival point of each cluster head node in each follower set is reliable within the time period t, the midpoint is regarded as an underwater sensor node with unchanged position.

[0030] Furthermore, the network topology repair method also arranges the repair node process by greedy movement, including the following steps:

[0031] Step 201: deploy a branch node at the position coordinates of all elements in the branch point set.

[0032] Step 202: The branch node of each level is taken as the initial node, and the branch node of the next level or the leaf node of the same level is taken as the terminal node; that is, the road sections of all repair paths can be regarded as being connected by an initial node and a terminal node;

[0033] Step 203: Each terminating node connected to the initial node determines whether the transmission from itself to the initial node is reliable. If so, it means that the section of the repair path does not need to deploy a repair node; if not, it means that the section of the deployment path needs to deploy a repair node to ensure data transmission reliability;

[0034] Step 204: The process of deploying a repair node on a section that does not meet the transmission reliability requirement so that the initial node can reliably transmit with the terminal node:

[0035] A repair node is placed at a position r away from the initial node along the direction of the end node on the road section, and it is determined whether the repair node can reach the initial node reliably. If so, the repair node is deployed at the position; if not, the repair node is moved toward the initial node by a distance Δd;

[0036] The expression of Δd is as follows:

[0037] Δd=mr (1)

[0038] Where: r is the communication radius of the underwater repair node; m is the moving distance adjustment coefficient, which is usually set to a positive real constant equal to or less than 0.05;

[0039] Repeat the above process until the repair node moves to a position where it can reach the source node with reliable transmission;

[0040] The repair node is converted into an initial node, and the original initial node is no longer used as an initial node; it is determined whether the transmission from the repair node to its terminal node is reliable. If it is reliable, the repair node deployment is completed on the road section; if it is not reliable, a repair node is placed at a position r away from the initial node along the direction of the terminal node on the road section, and the above movement steps are repeated to complete the deployment of the repair node; wherein:

[0041] The condition for stopping adding repair nodes on the road section is that there is an RN whose termination node is reliable in transmission;

[0042] Step 205: Repeat step 204 until the terminal nodes on all paths reach their initial nodes and the transmission is reliable.

[0043] Beneficial Effects

[0044] 1. The present invention designs an underwater wireless sensor network topology repair method suitable for dynamic water flow scenarios by sensing the current underwater environmental state. The topology connectivity repair method can be performed for the underwater wireless sensor network according to the underwater communication channel and the water flow movement characteristics, thereby improving the transmission reliability and reducing the number of repair node deployments.

[0045] 2. The repair path generation method based on path clipping designed in the repair path establishment stage of the present invention can construct a repair path connecting the subnetwork and the surface aggregation node according to the distribution characteristics of the underwater subnetwork in the underwater environment. While ensuring network connectivity, the path can effectively reduce the number of transmission hops from the underwater sensor node to the surface aggregation node, thereby improving data transmission reliability.

[0046] 3. The adaptive underwater repair node deployment method based on greedy movement designed in the repair node deployment stage of the present invention can ensure that data transmission in a dynamic underwater environment meets a predetermined signal-to-noise ratio threshold requirement, improve data transmission reliability, reduce the number of repair node deployments, and reduce network costs. Description of the drawings:

[0047] Figure 1 Mobile model diagram for underwater sensors;

[0048] Figure 2 A flowchart for the execution of this method. Specific implementation method:

[0049] The purpose of the present invention is achieved by the following technical solution, comprising the following steps:

[0050] 1. Repair path establishment phase

[0051] Step 101: Select the cluster head node (CN) of each underwater sub-network (USN), and set all CNs to be connected in an activated state.

[0052] Step 102: Determine whether the transmission from each underwater sensor node to the surface aggregation node (BS) is reliable within the time period t. If the transmission is reliable, the underwater sensor node replaces the original CN of the USN to become the new CN of the USN and changes the connection state to the sleep state.

[0053] Each underwater sensor node is fixed by an anchor chain connected to the bottom of the water, but the water flow will cause the underwater sensor node to shift position. The node shift model is as follows: Figure 1 shown.

[0054] The underwater sensor node vi 到 Whether the BS transmits reliably in time period t means that the BS receives the data from node v i The signal-to-noise ratio (SNR) of the data packet ij Is it always greater than a preset signal-to-noise ratio threshold SNR? set . Where t is the preset transmission reliability maintenance time.

[0055] The underwater subnetwork (USN) is defined as the underwater sensor nodes in any USN can communicate with each other through multi-hop, and two underwater sensor nodes belonging to different USNs cannot communicate with each other through multi-hop.

[0056] The cluster head node (CN) of any one USN refers to the underwater sensor node selected to represent the USN for repairing the network topology connectivity. The CN can be one or more underwater sensor nodes of the USN. For example, it can be set that the underwater sensor node with the most remaining energy in each USN is the CN.

[0057] Step 103: Set the BS as the target point G. Then store all the active CNs into the execution set of the target point G, and sort the CNs in the execution set of the target point G in descending order according to their Euclidean distances to the point G.

[0058] The execution set of each target point G is initialized as an empty set.

[0059] Step 104: Select the CN that ranks first and is active in the execution set as the current execution node (EN). Store the other active CNs in the execution set into the candidate follower set of the EN, and sort the CNs in the candidate follower set of the EN in ascending order according to their Euclidean distances to the EN.

[0060] The candidate follower set of each EN is initialized as an empty set.

[0061] Step 105: For the current execution node EN and its candidate follower set, first store the current execution node EN as the first element into its follower set. Then select the CN that ranks first in the candidate follower set as the candidate node CA. Calculate the median point of the data set composed of all elements in the follower set of the EN and the target point G, and then calculate the sum of the distances from all elements of this data set to this median point, and denote it as d1. Calculate the distance from the candidate node CA to the BS, and denote it as d2. Calculate the median point of the data set composed of all elements in the follower set of the EN, the target point G, and the candidate node CA, and then calculate the sum of the distances from all elements of this data set to this median point, and denote it as d3. If d3 < d1 + d2, then add the candidate node CA to the follower set of the EN, and delete the candidate node CA from the execution set. Finally, whether or not it is added to the follower set of the EN, the candidate node CA is deleted from the candidate follower set of the EN. Repeat this step until the candidate follower set of the EN becomes an empty set.

[0062] The follower set of each EN is initialized as an empty set.

[0063] Step 106: Repeat steps 104 - 105 until the execution set becomes an empty set.

[0064] Step 107: Calculate the median point P of the data set consisting of all CNs and the target point G in each follower set whose number of CNs is greater than 1. Connect the target point G and the point P of each follower set to construct a repair path. Then determine whether the transmission of each CN arrival point corresponding to the point P in each follower set is reliable within the time period t. If the transmission is reliable, delete the CN from the corresponding follower set and change its connection state to a dormant state, i.e., g=0. Connect the dormant CN and the point P of the corresponding follower set to construct a repair path. Then, calculate the number of CNs in the current follower set. If it is 1, change the connection state of the CN in the follower set to a dormant state, i.e., g=0. Connect the dormant CN and the point P of the corresponding follower set to construct a repair path.

[0065] In the process of judging whether the transmission of each CN arrival point corresponding to point P in each follower set is reliable within the time period t, point P is regarded as an underwater sensor node with unchanged position. That is, only the displacement of CN caused by water flow is considered. Therefore, the judgment process can be regarded as judging whether the transmission from a SN to BS is reliable.

[0066] Step 108: For any follower set, take its corresponding point P as the new target point G, and then turn this follower set into an execution set. Execute the above steps 103-107 until the execution set corresponding to the follower set becomes an empty set.

[0067] Step 109: Repeat step 108 until the execution sets corresponding to all follower sets become empty sets.

[0068] Step 110: Repeat the above steps 108-109 until the connection status of all CNs is in the dormant state. All target points G generated in the above steps (excluding the water surface convergence node) are stored in an empty set called a branch point set.

[0069] Through the above steps, we can connect the CN of each sub-network to the water surface aggregation node BS directly or through the elements in the branch point set. Next, we need to deploy repair nodes on the repair path so that the data packets sent by the CN of each sub-network can be reliably transmitted to the BS.

[0070] 2. Repair Node Arrangement Phase

[0071] Step 201: deploy an underwater repair node at the position coordinates of all elements in the branch point set, and these underwater repair nodes are called branch nodes KN. The surface convergence node BS is defined as the level 0 branch node KN_0. The branch node KN connected to KN_0 is defined as the level 1 branch node KN_1. In addition to KN_0, the branch node connected to the level 1 branch node KN_1 is called the level 2 branch node KN_2. And so on, each branch node can obtain its own level. Then, we set the SN connected to a certain level of branch nodes as the leaf node of that level. For example, the SN connected to a level 2 branch node KN_2 is a level 2 leaf node LN_2.

[0072] Step 202: The branch node of each level is taken as the initial node, and the branch node of the next level or the leaf node of the same level is taken as the terminal node. That is, all the sections of the repair path can be regarded as being connected by an initial node and a terminal node.

[0073] Step 203: Each terminating node connected to the initial node determines whether the transmission from itself to the initial node is reliable. If so, it means that the deployment path does not need to deploy a repair node (RN). If not, it means that the deployment path needs to deploy a RN to ensure data transmission reliability.

[0074] It can be known from steps 201 and 202 that when the BS is used as the initial node, the terminal node connected to the initial node is a level 1 branch node.

[0075] Any RN or SN v i to RN or BSv j Whether the transmission is reliable means: within time t, node v j Received from node v i The signal-to-noise ratio (SNR) of the data packet ij Is it always greater than the pre-set signal-to-noise ratio threshold SNR? set . Where t is the preset transmission reliability maintenance time.

[0076] Step 204: For the case where the transmission from the termination node to the initial node of a certain section is unreliable, place an underwater repair node RN at a position r away from the initial node along the direction of the termination node on this section. Determine whether the transmission from this RN to the initial node is reliable. If it is, then this RN is deployed at this position; if not, then this RN moves a distance of Δd in the direction of the initial node. Then, determine again whether the transmission from this RN to the initial node is reliable. If it is, then this RN is deployed at this position; if not, then this RN moves a distance of Δd in the direction of the initial node again. Repeat the above process until this RN moves to a position where the transmission from it to the source node is reliable. Then this RN is transformed into the initial node, and the original initial node is no longer the initial node. Then determine whether the transmission from this RN to its termination node is reliable. If it is reliable, then the repair node deployment for this section is completed; if it is not reliable, then continue to place an RN at a position r away from the initial node along the direction of the termination node on this section, and repeat the above moving process to complete the deployment of this RN; the condition for stopping adding RNs on this path is that there is an RN whose transmission to its termination node is reliable.

[0077] Where r is the communication radius of the underwater repair node.

[0078] The expression of Δd is as follows:

[0079] Δd = mr. (1)

[0080] In the formula, m is the moving distance adjustment coefficient, and it is usually set as a positive real constant equal to or less than 0.05.

[0081] Step 205: Repeat Step 3 until the transmission from the termination node to the initial node on all paths is reliable.

[0082] Through the above steps, each underwater sensor can reliably send data to the CN node of the sub - network where it is located, and the CN node of each sub - network can send underwater data to the BS through the repair node or directly.

[0083] To describe the implementation manner more clearly, assume that in a three - dimensional underwater wireless sensor network, there is 1 surface aggregation node fixed on the water surface (abbreviated as BS). N underwater sensor nodes (SN) are distributed in a fixed area in the underwater environment (fixed by anchor chains), forming M non - connected sub - networks (USN), where M < N. The underwater relay node is abbreviated as RN, and the communication radius of the underwater sensor node is R s , and the communication radius of RN is R r . The following will detail the specific manner, structure, features and functions of the topology repair method of the underwater wireless sensor network designed according to the present invention with reference to the accompanying drawings.

[0084] 1. Repair path establishment phase

[0085] S11: Select the cluster head node (CN) of each underwater sub-network (USN), and set the connection state of all CNs to the active state, i.e. g = 1. Then, determine whether the transmission from each underwater sensor node (SN) to the surface sink node (BS) is reliable within the time period t. If the transmission is reliable, the SN replaces the original CN of the USN as the new CN, and changes the connection state to the dormant state, i.e. g = 0.

[0086] Each SN is fixed by an anchor chain connected to the bottom of the water, but the water flow will cause the underwater sensor node to shift position. The node shift model is as follows: Figure 1 shown.

[0087] The underwater subnetwork (USN) is defined as the underwater sensor nodes in any USN can communicate with each other through multi-hop, and two underwater sensor nodes belonging to different USNs cannot communicate with each other through multi-hop.

[0088] The cluster head node (CN) of any USN refers to the underwater sensor node selected to represent the USN in order to repair the network topology connectivity. The CN can be one or more underwater sensor nodes of the USN. For example, the underwater sensor node with the most remaining energy in each USN can be set as the CN.

[0089] Any SNv i to BSv j Whether the transmission is reliable means: within time t, node v j Received from node v i The signal-to-noise ratio (SNR) of the data packet ij Is it always greater than a preset signal-to-noise ratio threshold SNR? set . Where t is the pre-set transmission reliability maintenance time. That is, every time t, SN will actively detect its maximum offset distance. The signal-to-noise ratio is related to the success rate of data packet transmission. The larger the signal-to-noise ratio, the higher the success rate of data packet transmission. Setting the signal-to-noise ratio threshold is often used in the field of underwater wireless sensor networks to maintain the reliability of data transmission. Where v j Received from v i The signal-to-noise ratio (SNR) of the data packet ij The expression is as follows:

[0090]

[0091] Where W i is node v i The transmission power, N(f)Δf is the frequency-dependent underwater noise parameter, l ij is node vi To node v i The Euclidean distance, A(l ij ,f) is the channel attenuation parameter, expressed as follows:

[0092]

[0093] Where A0 is a positive constant value, and a(f) is a positive value related to the frequency. In summary, the present application can find that when the node's transmission power W i When the frequency band f is fixed, SNR ij is a distance l ij Inversely related values.

[0094] Situation analysis, in the scenario without water flow, v j Received from v i The signal-to-noise ratio of the transmitted data packet is SNR ij , when the distance l ij ≤R s When ij ≥SNR set Under the above definition, SN in the marine environment is prone to displacement due to water flow. Before the displacement occurs, v j Receive the v i The signal-to-noise ratio of the transmitted data packet is greater than SNR set , it is likely that after the position shift occurs, the signal-to-noise ratio becomes less than SNR set Therefore, in order to ensure the reliability of data transmission, we analyze the data transmission signal-to-noise ratio in the water flow scene, assuming that SNv i The static coordinates are (x i0 ,y i0 ,z i0 ) The coordinate of the maximum displacement caused by the water flow within time t is (x i ,y i ,z i ),BSv j The coordinates of (x j ,y j ,z j ). Then node v i and v j The maximum Euclidean distance can be estimated as

[0095]

[0096] That is, when v i is a SN, if there is always Then v i to v jData transmission is reliable.

[0097] S12: Set BS as the target point G. Then store all CNs with g = 1 into the execution set of the target point G, and sort the CNs in the execution set of the target point G in descending order according to their Euclidean distances to the point G.

[0098] Wherein: The execution set of each target point G is initialized as an empty set.

[0099] S13: Select the CN with the top rank and g = 1 in this execution set as the current execution node (EN). Store the other CNs with g = 1 in the execution set into the candidate follower set of this EN, and sort the CNs in the candidate follower set of this EN in ascending order according to their Euclidean distances to this EN.

[0100] Wherein: The candidate follower set of each EN is initialized as an empty set.

[0101] S14: For the current execution node EN and its candidate follower set, first store the current execution node EN as the first element into its follower set. Then select the CN with the top rank in the candidate follower set as the candidate node CA. Calculate the median point of the data set composed of all elements in the follower set of EN and the target point G, and then calculate the sum of the distances from all elements in this data set to this median point, and denote it as d1. Calculate the distance from the candidate node CA to BS, and denote it as d2. Calculate the median point of the data set composed of all elements in the follower set of EN, the target point G and the candidate node CA, and then calculate the sum of the distances from all elements in this data set to this median point, and denote it as d3. If d3 < d1 + d2, then add the candidate node CA to the follower set of EN, and delete the candidate node CA from the execution set. Finally, whether it is added to the follower set of EN or not, the candidate node CA is deleted from the candidate follower set of EN. Repeat this step until the candidate follower set of EN becomes an empty set.

[0102] Wherein: The follower set of each EN is initialized as an empty set.

[0103] Wherein: For a set of points, the median point P means that the sum of the distances from all points to the point P is the smallest. Specifically, for a data set containing H nodes, assuming the points are {n1, n2, …, n H}, the coordinates of its median point P can be calculated by Where x, y, z respectively represent three-dimensional coordinates, and (x p , y p , z p ) that can satisfy the above formula is the coordinates of the median point P.

[0104] S15: Repeat steps 3-4 until the execution set becomes an empty set.

[0105] Among them: Through step 5 we will obtain multiple follower sets.

[0106] S16: Calculate the median point P of the data set consisting of all CNs and the target point G in each follower set whose number of cluster head nodes is greater than 1. Connect the target point G and the point P of each follower set to construct a repair path. Then determine whether the transmission of each CN arrival point corresponding to the point P in each follower set is reliable within the time period t. If the transmission is reliable, delete the CN from the corresponding follower set, and change the connection state of the CN to the dormant state, that is, g=0. Connect the dormant CN and the point P of the corresponding follower set to construct a repair path. Calculate the number of CNs in the current follower set. If it is 1, change the connection state of the CN in the follower set to the dormant state, that is, g=0. Connect the dormant CN and the point P of the corresponding follower set to construct a repair path.

[0107] Among them: In the process of judging whether the transmission of each CN arrival point corresponding to point P in each follower set is reliable within the time period t, point P is regarded as an underwater sensor node with unchanged position. That is, only the displacement of CN caused by water flow is considered. Therefore, the judgment process can be regarded as judging whether the transmission from a SN to BS is reliable.

[0108] Step S17: For any follower set, take its corresponding point P as the new target point G, and then change this follower set into an execution set. Execute the above steps S12-S16 until the execution set corresponding to the follower set becomes an empty set.

[0109] Step S18: Repeat step S17 until the execution sets corresponding to all follower sets become empty sets.

[0110] Step S19: Repeat the above steps S17-S18 until the connection status of all CNs is in the dormant state, that is, g = 0. All target points G generated in the above steps are stored in an empty set called a branch point set.

[0111] Through the above steps, we can connect the CN of each sub-network to the water surface aggregation node BS directly or through the elements in the branch point set. Next, we need to deploy repair nodes on the repair path so that the data packets sent by the CN of each sub-network can be reliably transmitted to the BS.

[0112] 2. Repair node layout phase

[0113] S21: Deploy an underwater repair node (RN) at the position coordinates of all elements in the branch point set, and these RNs are called branch nodes KN. The surface convergence node BS is defined as the level 0 branch node KN_0. The branch node KN connected to KN_0 is defined as the level 1 branch node KN_1. In addition to KN_0, the branch node connected to the level 1 branch node KN_1 is called the level 2 branch node KN_2. And so on, each branch node can obtain its own level. Then, we set the SN connected to a certain level of branch nodes as the leaf node of that level. For example, the SN connected to a level 2 branch node KN_2 is a level 2 leaf node LN_2.

[0114] S22: Take the branch node of each level as the initial node, and the branch node of the next level or the leaf node of the same level as the terminal node. That is, all sections of the repair path can be regarded as connected by an initial node and a terminal node. Then, each terminal node connected to the initial node determines whether the transmission to the initial node is reliable. If so, it means that RN does not need to be deployed on this deployment path. If not, it means that RN needs to be deployed on this deployment path to ensure data transmission reliability.

[0115] According to steps S21 and S22, when the BS is used as the initial node, the terminal node connected to the initial node is a level 1 branch node.

[0116] Where: Any RN or CN v i to RN or BS v j Whether the transmission is reliable refers to: within time t, v j Received from v i The signal-to-noise ratio (SNR) of the data packet ij Is it always greater than the pre-set signal-to-noise ratio threshold SNR? set . Where t is the pre-set transmission reliability maintenance time. Where v j Received from v i The signal-to-noise ratio (SNR) of the data packet ij The expression is as shown in formula (2).

[0117] In order to ensure the reliability of data transmission in the dynamic water flow scenario, the following analyzes the cases where the sending node is RN and CN respectively. First, suppose the sending node v i is a RN, receiving node v j RN or BS. Node v i The deployment coordinates are (x i0 ,y i0 ,z i0 ), the coordinate of the deployment position that is farthest offset by the water flow within time Δt is (x i ,y i ,zi )。The time Δt is the time interval for the RN to autonomously correct its own position, and usually Δt < t. Similarly, the receiving node v j is deployed at the coordinates (x j0 , y j0 , z j0 ). The coordinates of the position that is farthest from the deployment position under the influence of water flow within the time Δt are (x j , y j , z j ). Then, the maximum Euclidean distance between nodes v i and v j can be estimated as

[0118]

[0119] That is to say, when v i is an RN, if there is always within the time t, then the data transmission from v i to v j is reliable. If the node v j is a BS, then we can consider Δt = 0, that is, l j0 = 0.

[0120] Then, assume that the sending node v i is a CN, and the receiving node v j is an RN or a BS. The static coordinates of v i are (x i0 , y i0 , z i0 ). The coordinates of the maximum offset under the influence of water flow within the time t are (x i , y i , z i ). The deployment coordinates of v j are (x j0 , y j0 , z j0 ). The coordinates of the position that is farthest from the deployment position under the influence of water flow within the time Δt are (x j , y j , z j ). Then, the maximum Euclidean distance between nodes v i and v j can be estimated as

[0121]

[0122] That is to say, when v i is a CN, if there is always within the time t, then the data transmission from v i to v jData transmission is reliable. If node v j is BS, then we can assume that Δt = 0, that is, l j0 =0.

[0123] S23: For the case where the transmission from the terminal node of a certain section to its initial node is unreliable, an underwater repair node RN is placed at a position r away from the initial node along the direction of the terminal node on the section, and it is determined whether the transmission of the RN to the initial node is reliable. If so, the RN is deployed at this position; if not, the RN moves a distance Δd toward the initial node. Then it is determined again whether the transmission of the RN to the initial node is reliable. If so, the RN is deployed at this position; if not, the RN moves a distance Δd toward the initial node again. Repeat the above process until the RN moves to a position where it can reach the source node with reliable transmission. Then the RN is converted into an initial node, and the original initial node is no longer used as an initial node. Then it is determined whether the transmission of the RN to its terminal node is reliable. If reliable, the deployment of the repair node is completed on the section; if unreliable, continue to place an RN at a position r away from the initial node along the direction of the terminal node on the section, and repeat the above movement link to complete the deployment of the RN; the condition for stopping adding RN on the path is that there is an RN that reaches its terminal node with reliable transmission.

[0124] Where: r is the communication radius of the underwater repair node.

[0125] Where: The expression of Δd is as follows:

[0126] Δd=mr

[0127] Wherein m is the moving distance adjustment coefficient, which is usually set to a positive real constant equal to or less than 0.05.

[0128] S24: Repeat step S23 until the terminal nodes on all paths reach their initial nodes and the transmission is reliable.

[0129] Through the above steps, each underwater sensor can reliably send data to the CN node of its sub-network, and each CN node of the sub-network can send underwater data to the BS through the repair node or directly.

[0130] The technical solution of the present invention and the specific implementation methods described above are merely illustrative and not restrictive. Without departing from the scope of the present invention and the scope of protection of the claims, a person skilled in the art may make many specific changes under the guidance of the present invention, all of which fall within the protection scope of the present invention.

Claims

1. A topology repair method for underwater wireless sensor networks suitable for dynamic water flow scenarios, characterized in that: The network topology repair method constructs a repair path phase process connecting all sub-networks and the surface aggregation node according to the location distribution characteristics of all the underwater sub-networks that are not connected to each other: the process includes the following steps: Step 101: Select the cluster head node of each underwater sub-network and set the connection state of all cluster head nodes to be activated; Step 102: Determine whether the transmission from each underwater sensor node to the surface aggregation node is reliable within the time period t. If the transmission is reliable, the underwater sensor node replaces the original cluster head node of the underwater sub-network where it is located to become the new cluster head node of the underwater sub-network, and changes its own connection state to a dormant state; otherwise, proceed to the next step 103; Step 103: Set the sink node as the target point, and define an empty set as the execution set of the target point; store all the cluster head nodes in the active state into the target point execution set, and sort the cluster head nodes of the target point execution set from large to small according to their respective Euclidean distances to the target point; Step 104: Select the cluster head node that ranks the top and is in the activated state in the target point execution set as the current execution node, and define two empty sets as the candidate follower set and the follower set of the execution node respectively; The other cluster head nodes in the target point execution set that are in an activated state are stored in the candidate follower set of the current execution node, and the cluster head nodes in the candidate follower set of the current execution node are sorted in ascending order according to their respective Euclidean distances to the execution node; Step 105: sequentially determine whether each cluster head node in the candidate follower set of the current execution node can be added to the follower set of the execution node. Process: Store the current execution node as the first element in its follower set; Select the cluster head node with the highest ranking in the candidate follower set as the candidate node, and calculate the median point of the data set composed of all elements in the follower set of the current execution node and the target point; Calculate the sum of the distances from all elements of the data set to the median point, and record it as d1. Calculate the distance from the candidate node to the sink node, and record it as d2. Calculate the median point of the data set consisting of all elements in the follower set of the current execution node EN, the target point and the candidate node. Calculate the sum of the distances from all elements of the data set to the median point, and record it as d3. Compare the relationship between d1, d2, and d3. If d3 < d1 + d2, add the candidate node to the follower set of the current execution node and delete the candidate node from the execution set. Regardless of whether the candidate node is added to the follower set of the current execution node, the candidate node is deleted from the candidate follower set of the current execution node; repeat this step until the candidate follower set of the current execution node becomes an empty set; Step 106: Repeat steps 104-105 until the execution set becomes an empty set; Step 107: A process of constructing a topology repair path according to the cluster head node information in each follower set; Step 108: For any follower set, take its corresponding median point as the new target point, and change this follower set into an execution set; Step 109: Execute the above steps 103-108 until the execution set corresponding to the follower set becomes an empty set; Step 110: Repeat step 109 until the execution sets corresponding to all follower sets become empty sets; Step 111: repeat the above steps 109-110 until the connection status of all cluster head nodes is dormant; store all target points generated in the above steps into the empty set of the branch point set.

2. The underwater wireless sensor network topology repair method applicable to dynamic water flow scenarios according to claim 1 is characterized in that: The process of constructing a topology repair path according to the cluster head node information in each follower set in step 107 includes the following steps: Calculate the median point of the data set consisting of all cluster head nodes and the target point in the follower set with the number of cluster head nodes greater than 1; connect the target point and the corresponding median point of each follower set to construct a repair path; Determine whether the arrival point of each cluster head node in each follower set to the median point of its follower set is reliable within the time period t. If the transmission is reliable, delete the cluster head node from the corresponding follower set, change the connection state of the cluster head node to the dormant state, connect the dormant cluster head node and the median point of the corresponding follower set, and build a repair path. Calculate the number of cluster head nodes in the current follower set. If it is 1, change the connection state of the cluster head nodes in the follower set to the dormant state. Connect the dormant cluster head node and the median point of the corresponding follower set to construct a repair path; Among them: in the process of judging whether the transmission of the midpoint of the arrival point of each cluster head node in each follower set is reliable within the time period t, the midpoint is regarded as an underwater sensor node with unchanged position.

3. The underwater wireless sensor network topology repair method applicable to dynamic water flow scenarios according to claim 1 is characterized in that: The network topology repair method also arranges the repair node process through greedy movement, including the following steps: Step 201: deploy a branch node at the position coordinates of all elements in the branch point set; Step 202: The branch node of each level is taken as the initial node, and the branch node of the next level or the leaf node of the same level is taken as the terminal node; that is, the road sections of all repair paths can be regarded as being connected by an initial node and a terminal node; Step 203: Each terminating node connected to the initial node determines whether the transmission from itself to the initial node is reliable. If so, it means that the section of the repair path does not need to deploy a repair node; if not, it means that the section of the repair path needs to deploy a repair node to ensure data transmission reliability. Step 204: The process of deploying a repair node on a section that does not meet the transmission reliability requirement so that the initial node can reliably transmit with the terminal node: A repair node is placed at a position r away from the initial node along the direction of the end node on the road section, and it is determined whether the repair node can reach the initial node reliably. If so, the repair node is deployed at the position; if not, the repair node is moved toward the initial node by a distance Δd; The expression of Δd is as follows: Δd=mr. (1) Where: r is the communication radius of the underwater repair node; m is the moving distance adjustment coefficient, which is usually set to a positive real constant equal to or less than 0.05; Repeat the above process until the repair node moves to a position where it can reach the source node with reliable transmission; The repair node is converted into an initial node, and the original initial node is no longer used as an initial node; it is determined whether the transmission from the repair node to its terminal node is reliable. If it is reliable, the repair node deployment is completed on the road section; if it is not reliable, a repair node is placed at a position r away from the initial node along the direction of the terminal node on the road section, and the above movement steps are repeated to complete the deployment of the repair node; wherein: The condition for stopping adding repair nodes on the road section is that there is an RN whose termination node is reliable in transmission; Step 205: Repeat step 204 until the terminal nodes on all paths reach their initial nodes and the transmission is reliable.

Images