Clustering method for underwater wireless sensor networks based on improved AP clustering

Abstract

This invention discloses a clustering method for underwater wireless sensor networks based on improved AP clustering, which relates to the field of underwater wireless sensor network topology optimization. By acquiring node status information in the underwater wireless sensor network and initializing node feature data, the method improves its adaptability to underwater environments by comprehensively considering inter-node distance, received signal strength, and node residual energy. It also employs an energy and position-aware mechanism to iteratively update attraction and affiliation information to form a preliminary cluster structure, effectively addressing node energy consumption and position changes in complex underwater environments. Based on the node communication range, preference values ​​are updated to adjust the cluster structure to ensure that no nodes leave the cluster. A comprehensive weight function for cluster head updates is designed to avoid bias due to a single factor. This invention can better adapt to complex underwater environments, effectively optimize network topology, balance sensor node energy consumption, and improve the stability and energy efficiency of underwater wireless sensor networks.

Description

Technical Field

[0001] The present invention relates to the technical field of underwater wireless sensor network topology optimization, in particular to an underwater wireless sensor network clustering method based on improved AP clustering. Background Art

[0002] With the rapid advancement of marine technology and the increasing demand for marine resource development, underwater wireless sensor networks are becoming increasingly widely used in the marine field. However, due to the complexity of the marine environment and the limited energy of nodes, designing a clustering method suitable for underwater wireless sensor networks is a huge challenge.

[0003] In recent years, with the advancement of machine learning, clustering algorithms have been applied to the clustering process of underwater wireless sensor networks. Clustering algorithms are typically applied to the initial cluster formation process. However, most existing algorithms ignore the unique characteristics of the underwater environment, such as the nonlinear attenuation of acoustic wave propagation, multipath effects, and signal interference from ocean currents, making these algorithms less applicable in underwater environments. Furthermore, considering the dynamic changes of nodes is crucial. In underwater environments, nodes may change position due to factors such as currents and turbulence. Existing algorithms fail to adjust the network topology in a timely manner to accommodate these changes, resulting in network outages or data loss. Furthermore, improper cluster head elections can lead to uneven energy consumption among some nodes, severely impacting the network's lifespan. Especially in underwater environments, where sensor nodes typically rely on limited battery power, balancing energy consumption among sensor nodes becomes a key factor in determining network energy efficiency and longevity.

[0004] In summary, the current clustering-based clustering methods have problems such as poor adaptability to underwater environments, poor network stability, and low network energy efficiency. These problems increase the difficulty of clustering design in underwater wireless sensor networks. Summary of the Invention

[0005] In order to overcome the above problems existing in the prior art, the present invention proposes an underwater wireless sensor network clustering method based on improved AP clustering.

[0006] The technical solution adopted by the present invention to solve the technical problem is: a clustering method for underwater wireless sensor networks based on improved AP clustering, comprising the following steps:

[0007] Step 1: Obtain the status information of the nodes in the underwater wireless sensor network and initialize the characteristic data of the nodes;

[0008] Step 2: Calculate the similarity between nodes by combining the relative residual energy of nodes, received signal strength, and physical distance between nodes, construct the similarity matrix S of the underwater wireless sensor network, and set the preference value S(k,k);

[0009] Step 3: Clustering and optimizing the nodes using the similarity matrix S obtained in step 2, combining the energy-aware mechanism with the location-aware mechanism, iteratively updating the attraction information and affiliation information to form a preliminary cluster structure;

[0010] Step 4: According to the communication range between nodes, adjust the cluster structure by updating the preference value S(k,k) to ensure that there are no out-cluster nodes;

[0011] Step 5: Update the cluster head within the cluster according to the comprehensive weight of the nodes.

[0012] In the above-mentioned underwater wireless sensor network clustering method based on improved AP clustering, step 2 is specifically as follows:

[0013] ;

[0014] ;

[0015] ;

[0016] ;

[0017] ;

[0018] ;

[0019] in, Represents the similarity between node i and node k; when i=k, is the preference value, which determines the number of clusters; and are the coordinates of node i and node k respectively; N is the total number of sensor network nodes; is the three-dimensional Euclidean distance between node i and node k; and is the spatial coordinate of the node; is the received signal strength, which indicates the signal power received by node i from node k; is the transmit power of the sensor node, is the underwater acoustic channel propagation loss; is the normalization coefficient; d is the propagation distance between nodes; k is the expansion factor; is the absorption coefficient; is the relative remaining energy of the node, and its value reflects the current energy status of the node; is the current remaining energy of sensor node k, is the initial energy of node k.

[0020] The above clustering method for underwater wireless sensor networks based on improved AP clustering, the specific content of step 3 is as follows: Through the energy awareness mechanism and the location awareness mechanism, the attraction information and the membership information are iteratively updated. When the change in the value of r(i, i) + a(i, i) for each node i is less than 10 -6 or the maximum preset number of iterations is reached, the update stops;

[0021] When a(i, i) + r(i, i) > T, node i is selected as the cluster head node; when a(i, i) + r(i, i) < T, node i is assigned to the cluster where the cluster head node with the maximum r(i, k) + a(i, k) is located. Here, T is a set threshold representing the criterion for a node to select a cluster head; a(i, i) is the attraction information of node i, and r(i, i) is the membership information of node i.

[0022] The above clustering method for underwater wireless sensor networks based on improved AP clustering, the role of the energy awareness mechanism and the location awareness mechanism is to comprehensively consider the energy level and location change of nodes when updating the attraction information r(i, k) and the membership information a(i, k), and dynamically adjust the network topology. The specific formula is:

[0023] ;

[0024] ;

[0025] ;

[0026] ;

[0027] ;

[0028] [[ID=**31**]]Among them, w(i) is a regulation factor combining energy awareness and location awareness, which is used to dynamically adjust the message transmission frequency of nodes; is the energy awareness coefficient, which is used to reflect the influence of the remaining energy of nodes; is the current remaining energy of node i; is the maximum energy of the node; is the location awareness coefficient, which is used to represent the influence of the location change of nodes on message transmission; is the amplitude of the location change of node i; is the maximum allowable amplitude of location change; λ is the weight factor of energy awareness and location awareness, 0 ≤ λ ≤ 1.

[0029] The above clustering method for underwater wireless sensor networks based on improved AP clustering, the specific content of step 4 is as follows: Calculate the actual physical distance between each cluster head node h and its cluster member node m , and take Communication radius with the cluster head node Compare, if there is a node that satisfies , then return to step 2 to adjust the updated preference value s(k, k); if there is no node that satisfies , then go to step 5.

[0030] In the above-mentioned underwater wireless sensor network clustering method based on improved AP clustering, the update formula of the preference value s(k,k) is specifically as follows:

[0031] ;

[0032] in, is the adjustment factor, which controls the adjustment amplitude of the preference value.

[0033] In the above-mentioned underwater wireless sensor network clustering method based on improved AP clustering, step 5 is specifically as follows:

[0034] ;

[0035] ;

[0036] ;

[0037] ;

[0038] ;

[0039] ;

[0040] ;

[0041] ;

[0042] in, is the cluster head of each cluster; are all nodes in cluster a; is the comprehensive weight of node k; is the energy weight of the node; is the current remaining energy of node k, is the maximum energy of the node; is the communication quality weight of the node; RSS(k) is the received signal strength of node k, is the propagation delay of node k; is the load weight of the node; is the frequency of data transmission of node k; is the neighbor density weight of the node; is the number of neighbor nodes of node k, is the maximum number of neighbor nodes in the network; is the mobility weight of the node; is the average displacement, and the position change of the node within a certain period of time is used as the mobility index; T is the total length of the time series, t is the time point, is the three-dimensional space coordinate of node k at time point t.

[0043] The beneficial effects of the present invention are as follows: first, the present invention initializes the node feature data by acquiring the node status information in the underwater wireless sensor network. Then, a method for improving the similarity calculation formula is proposed, which comprehensively considers the distance between nodes, the received signal strength, and the remaining energy of the nodes to improve the adaptability of the method to underwater environments. Secondly, an energy and position perception mechanism is adopted to iteratively update the attraction information and affiliation information to form a preliminary cluster structure, effectively addressing the issues of node energy consumption and position changes in complex underwater environments. Then, based on the node communication range, the cluster structure is adjusted by updating the preference value to ensure that there are no out-cluster nodes. Finally, a weighted scoring function is designed to update the cluster head within the cluster by integrating multiple factors such as the communication quality between nodes, load conditions, and neighbor density, avoiding the bias of a single factor and balancing the energy consumption of sensor nodes.

[0044] The present invention can enable the network to better adapt to complex underwater environments, effectively optimize the network topology, balance the energy consumption of sensor nodes, and improve the stability and energy efficiency of underwater wireless sensor networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic flow chart of the present invention;

[0046] Figure 2 Schematic diagram of the improved AP clustering algorithm results of the present invention;

[0047] Figure 3 This is a schematic diagram of the out-cluster node determination of the present invention;

[0048] Figure 4 It is a schematic diagram of the clustering structure of the underwater wireless sensor network of the present invention. DETAILED DESCRIPTION

[0049] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0050] This embodiment proposes an underwater wireless sensor network clustering method based on improved AP clustering. The overall flow chart is as follows: Figure 1 As shown, the following steps are included:

[0051] Step 1: Obtain the status information of the nodes in the underwater wireless sensor network, such as the location, remaining energy, and transmission power, and initialize the characteristic data of the nodes.

[0052] Step 2: In the improved AP clustering algorithm, in order to be more adaptable to the underwater environment, the similarity matrix is ​​initialized by combining the node residual energy, signal strength, and physical distance.

[0053] Step 2 is as follows:

[0054] By calculating the similarity between nodes, the similarity matrix S of the underwater wireless sensor network is constructed. The calculation of similarity takes into account the physical distance between nodes, received signal strength (RSS), and the relative residual energy of nodes. , in order to accurately evaluate the communication quality and energy status between nodes, its formula is defined as:

[0055] ;

[0056] in, Represents the similarity between node i and node k. When i=k, it is called is the preference value, which determines the number of clusters. and are the coordinates of node i and node k respectively, and N is the total number of sensor network nodes.

[0057] is the three-dimensional Euclidean distance between node i and node k, calculated as:

[0058] ;

[0059] in and is the spatial coordinate of the node.

[0060] is the received signal strength, which represents the signal power received by node i from node k. The calculation formula is:

[0061] ;

[0062] in, is the transmit power of the sensor node, is the underwater acoustic channel propagation loss, and its calculation formula is:

[0063] ;

[0064] in, is the normalization coefficient, d is the propagation distance between nodes, and k is the expansion factor. The value of k is different in different scenarios. In practical applications, k=1.5 is often used.

[0065] is the absorption coefficient, which will change with the frequency f at any time. There are two calculation formulas according to the frequency:

[0066] ;

[0067] is the relative remaining energy of the node, and its value reflects the current energy status of the node. The calculation formula is:

[0068] ;

[0069] in is the current remaining energy of sensor node k, is the initial energy of node k.

[0070] The resulting similarity matrix S is then used to form the initial cluster structure. During the clustering process, nodes are clustered and optimized using the similarity matrix S. Taking into account underwater environmental factors, nodes with high residual energy, good communication quality, and a relatively close distance to member nodes are prioritized as cluster heads. This strategy is more adaptable to complex underwater environments, ensuring a balanced balance between energy consumption and communication performance for cluster heads, thereby improving overall network stability and data transmission efficiency.

[0071] Step 3: To improve network stability, the energy-aware mechanism and the location-aware mechanism are combined to iteratively update the attraction information and affiliation information to form a preliminary cluster structure.

[0072] Step 3 is as follows:

[0073] In the dynamic environment of underwater wireless sensor networks, in order to better adapt to the complex node energy consumption and location changes, a message update mechanism combining energy awareness and location awareness is introduced. The core of the energy awareness mechanism and location awareness mechanism is to comprehensively consider the node energy level and location changes when updating the attraction information r(i,k) and the affiliation information a(i,k), thereby dynamically adjusting the network topology. The specific formula is as follows:

[0074] ;

[0075] ;

[0076] Where w(i) is a control factor that combines energy awareness and location awareness and is used to dynamically adjust the message transmission frequency of the node. It is defined as follows:

[0077] .

[0078] is the energy perception coefficient, which is used to reflect the impact of the node's remaining energy. The formula is as follows:

[0079] ;

[0080] Among them, is the current remaining energy of node i, is the maximum energy of the node.

[0081] is the position perception coefficient, which is used to represent the impact of the node position change on message transmission. The formula is as follows:

[0082] ;

[0083] Among them, is the amplitude of the position change of node i, is the maximum allowable amplitude of the position change.

[0084] λ is the weight factor of energy perception and position perception, satisfying 0 ≤ λ ≤ 1. By adjusting the value of λ, a balance can be achieved between energy priority and position priority.

[0085] The schematic diagram of the result of the improved AP clustering algorithm is as shown in Figure 2 The algorithm iteratively updates the attraction information and the membership information through the above formula. When the value change of r(i, i) + a(i, i) of each node i is less than 10 -6 or reaches the maximum preset number of iteration times, the update stops. When a(i, i) + r(i, i) > T, then node i is selected as the cluster head node. When a(i, i) + r(i, i) < T, then node i is assigned to the cluster where the cluster head node with the maximum r(i, k) + a(i, k) is located. Among them, T is a set threshold, which represents the standard for node selection of the cluster head. In practical applications, it is necessary to adjust the threshold T according to environmental factors such as the node energy level, communication quality, and underwater noise of the underwater network to adapt to different working scenarios.

[0086] Step 4: According to the communication range between sensor nodes, adjust the clustering structure by updating the preference value to ensure that there are no off-cluster nodes.

[0087] Step 4 is specifically as follows: [[ID=四十一]]

[0088] First, calculate the actual physical distance between each cluster head node h and its cluster member node m. The calculation formula is as follows:

[0089] ;

[0090] Among them and are the three-dimensional space coordinates of the member node m and the cluster head node h respectively.

[0091] Secondly, the schematic diagram of off-cluster node determination is as shown in Figure 3As shown, in order to ensure that all member nodes m can effectively communicate with the cluster head node h, it is necessary to Communication radius with the cluster head node For comparison, the judgment conditions are:

[0092] ;

[0093] If no node meets this condition, it means that the cluster head node and the member nodes can effectively communicate, and then proceed directly to the next step S5. If a node meets this condition, it means that node i is too far away to effectively communicate with the cluster head node k, and it is necessary to return to step S2 to adjust and update the preference value s(k, k). If the distance between the cluster member node and the cluster head node is greater than the communication range, it means that the current preference value s(k, k) is too large and needs to be lowered. The preference value update formula is as follows:

[0094] ;

[0095] in is the adjustment factor, which controls the adjustment amplitude of the preference value.

[0096] Schematic diagram of underwater wireless sensor network clustering structure Figure 4 As shown in the figure, after clustering through the above steps, a certain communication relationship is established between the cluster head node and its member nodes. The cluster head node is responsible for collecting data from cluster member nodes and transmitting it to the aggregation node or upper-layer network to achieve efficient data aggregation and transmission. It then performs preliminary processing and fusion of the cluster data to reduce data redundancy and transmission overhead. However, due to the complexity of the underwater environment and the dynamic changes in node status, the performance of the cluster head node may be affected by various factors, such as energy depletion and decreased communication quality. Therefore, to ensure long-term stable operation of the network and efficient data transmission, it is necessary to dynamically update the cluster head node based on the comprehensive performance of the nodes.

[0097] Step 5: Update the cluster head within the cluster based on multiple factors such as inter-node communication quality, load conditions, neighbor density, mobility, etc.

[0098] In the cluster head update phase, multiple factors such as node residual energy, communication quality, load, neighbor density, mobility, etc. are comprehensively considered. The introduction of these factors can make the cluster head selection more reasonable, avoid the bias of a single factor, and improve the stability and energy efficiency of the network. The update selection is performed using the following formula. According to the comprehensive weight of the nodes, the node with the largest weight is selected as the cluster head. The calculation formula is:

[0099] ;

[0100] in, are all nodes in cluster a.

[0101] is the comprehensive weight of node k, taking into account factors such as energy, communication quality, neighbor density and mobility, and the calculation formula is:

[0102] .

[0103] is the energy weight of the node, which is calculated as:

[0104] ;

[0105] in is the current remaining energy of node k, is the maximum energy of the node.

[0106] is the communication quality weight of the node, and the calculation formula is:

[0107] ;

[0108] Where RSS(k) is the received signal strength of node k, is the propagation delay of node k.

[0109] is the load weight of the node, which is calculated as:

[0110] ;

[0111] in is the frequency of data transmission of node k.

[0112] is the neighbor density weight of the node, and the calculation formula is:

[0113] ;

[0114] in is the number of neighbor nodes of node k, is the maximum number of neighbor nodes in the network.

[0115] is the mobility weight of the node, which is calculated as:

[0116] ;

[0117] in is the average displacement, and the position change of the node within a certain period of time is used as the mobility index. The calculation formula is as follows:

[0118] ;

[0119] Where T is the total length of the time series, t is the time point, is the three-dimensional space coordinate of node k at time point t.

[0120] The above embodiments are merely exemplary embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art may make various modifications or equivalent substitutions to the present invention within the spirit and scope of protection of the present invention, and such modifications or equivalent substitutions shall also be deemed to fall within the scope of protection of the present invention.

Claims

1. The underwater wireless sensor network clustering method based on improved AP clustering is characterized by: It includes the following steps: Step 1: Obtain the status information of nodes in the underwater wireless sensor network and initialize the feature data of the nodes; Step 2: Calculate the similarity between nodes by combining the relative remaining energy, received signal strength, and physical distance between nodes, construct the similarity matrix S of the underwater wireless sensor network, and set the preference value S(k, k); Step 3: Cluster and optimize the nodes through the similarity matrix S obtained in Step 2, combine the energy awareness mechanism and the location awareness mechanism, iteratively update the attraction information and the membership information, and form a preliminary cluster structure; Step 4: Adjust the cluster structure by updating the preference value S(k, k) according to the communication range between nodes to ensure that there are no off-cluster nodes; Step 5: Update the cluster heads within the clusters according to the comprehensive weights of the nodes; The step 3 is specifically as follows: the attraction information and the belonging information are updated iteratively through the energy perception mechanism and the location perception mechanism. When the value of r(i,i)+a(i,i) of each node i changes by less than 10 -6 Or stop updating when the maximum preset number of iterations is reached; When a(i, i)+r(i, i)>T, node i is selected as the cluster head node; when a(i, i)+r(i, i)<T, node i is assigned to the cluster where the cluster head node with the maximum r(i, k)+a(i, k) is located, where T is a set threshold representing the criterion for node selection of the cluster head; a(i, i) is the attraction information of node i, and r(i, i) is the membership information of node i; The role of the energy awareness mechanism and the location awareness mechanism is to comprehensively consider the energy level and location change of nodes when updating the attraction information r(i, k) and the membership information a(i, k), and dynamically adjust the network topology. The specific formula is: ; ; ; ; ; Among them, w(i) is a control factor that combines energy awareness and location awareness, which is used to dynamically adjust the message transmission frequency of the node; is the energy perception coefficient, which is used to reflect the influence of the node's remaining energy; is the current remaining energy of node i; is the maximum energy of the node; is the position awareness coefficient, which is used to represent the impact of node position changes on message transmission; is the amplitude of the position change of node i; is the maximum allowable position change amplitude; λ is the weight factor of energy perception and position perception, 0≤λ≤1.

2. The underwater wireless sensor network clustering method based on improved AP clustering according to claim 1, characterized in that: The specific content of Step 2 is: ; ; ; ; ; ; in, Represents the similarity between node i and node k; when i=k, is the preference value, which determines the number of clusters; and are the coordinates of node i and node k respectively; N is the total number of sensor network nodes; is the three-dimensional Euclidean distance between node i and node k; and is the spatial coordinate of the node; is the received signal strength, which indicates the signal power received by node i from node k; is the transmit power of the sensor node, is the underwater acoustic channel propagation loss; is the normalization coefficient; d is the propagation distance between nodes; k is the expansion factor; is the absorption coefficient; is the relative remaining energy of the node, and its value reflects the current energy status of the node; is the current remaining energy of sensor node k, is the initial energy of node k.

3. The underwater wireless sensor network clustering method based on improved AP clustering according to claim 1, characterized in that: The step 4 is specifically as follows: calculating the actual physical distance between each cluster head node h and its cluster member node m ,Will Communication radius with the cluster head node Compare, if there is a node that satisfies , then return to step 2 to adjust the updated preference value s(k, k); if there is no node that satisfies , then go to step 5.

4. The underwater wireless sensor network clustering method based on improved AP clustering according to claim 3, characterized in that: The update formula of the preference value s(k, k) is specifically: ; in, is the adjustment factor, which controls the adjustment amplitude of the preference value.

5. The underwater wireless sensor network clustering method based on improved AP clustering according to claim 1, characterized in that: The specific content of Step 5 is: ; ; ; ; ; ; ; ; in, is the cluster head of each cluster; are all nodes in cluster a; is the comprehensive weight of node k; is the energy weight of the node; is the current remaining energy of node k, is the maximum energy of the node; is the communication quality weight of the node; RSS(k) is the received signal strength of node k, is the propagation delay of node k; is the load weight of the node; is the frequency of data transmission of node k; is the neighbor density weight of the node; is the number of neighbor nodes of node k, is the maximum number of neighbor nodes in the network; is the mobility weight of the node; is the average displacement, and the position change of the node within a certain period of time is used as the mobility index; T is the total length of the time series, t is the time point, is the three-dimensional space coordinate of node k at time point t.

Images