Air-sea cross-domain clustering network cluster head selection method

Abstract

The invention relates to the technical field of communication networking, and discloses an air-sea cross-domain clustering network cluster head selection method. Aiming at the problem that residual energy of some nodes is exhausted too early due to node energy isomerism in a three-dimensional air-sea cross-domain network, significant differences between overwater and underwater network environments and differences between different types of nodes such as an air unmanned aerial vehicle, a water surface buoy and an underwater sensor are considered; in addition to factors such as difference between radio channels and underwater acoustic channels among nodes, a basic clustering algorithm is improved, and a cluster head selection algorithm based on an improved hawk algorithm for an air-sea cross-domain network is designed. The proposed cluster head selection algorithm comprehensively considers the residual energy level of nodes in the network, the communication coverage range of the nodes and other factors, and node type differences and heterogeneous characteristics of underwater and overwater networks are integrated in a clustering strategy. The method aims at achieving effective balance of energy consumption, prolonging the overall life cycle of the network and showing wide application potential.

Description

Technical Field

[0001] The present invention relates to the technical field of communication networking, and in particular to a cluster head selection method for an air-sea cross-domain clustering network. Background Art

[0002] In order to meet the growing demand for cross-domain air-sea communication, cross-domain air-sea communication network technology has attracted widespread attention. At present, cross-domain air-sea communication network technology has become a highly valued cutting-edge technology, and the demand for cross-domain air-sea heterogeneous networking with aerial drones, surface buoys, and underwater sensors in air-sea environments is increasing. This network has the characteristics of cross-domain communication, node clustering, node type heterogeneity, and network type heterogeneity. Therefore, it is necessary to design a cross-domain heterogeneous network topology suitable for air-sea environments, so that heterogeneous nodes in air-sea environments can effectively complete cross-domain network communication. Among them, clustering algorithm is an efficient topological routing algorithm in the network. However, the current research on clustering algorithm at home and abroad is mainly reflected in the network environment of a single medium. There are few topological designs for cross-domain heterogeneous networks composed of large-scale drones, surface buoys, and underwater sensor nodes in air-sea environments.

[0003] Therefore, it is necessary to provide a cluster head selection method for air-sea cross-domain clustering network to realize the topology of air-sea cross-domain network under the conditions of heterogeneous nodes and links. Summary of the invention

[0004] The objects of the present invention are:

[0005] Aiming at the problem that the residual energy of some nodes is exhausted prematurely due to the heterogeneity of node energy in three-dimensional air-sea cross-domain networks, the basic clustering algorithm is improved based on the differences between the two network environments of water networks and underwater networks, as well as the differences in the types of aerial drones, surface buoys, and underwater sensor nodes, and the differences between radio channels and underwater acoustic channels between nodes. A cluster head selection method for air-sea cross-domain clustering networks is designed. This method comprehensively considers the elected cluster heads from the aspects of the residual energy of nodes in the network, the communication distance of nodes, etc., and considers the node type, the heterogeneous characteristics of underwater and water networks, etc. in the clustering process, which can effectively balance energy consumption and improve the network life cycle, and has broad application prospects.

[0006] The technical solution of the present invention is:

[0007] The present invention provides a cluster head selection method for an air-sea cross-domain clustering network, the method comprising the optimization of a Sky Eagle algorithm and the application of an improved Sky Eagle algorithm in a clustering algorithm.

[0008] The cluster head selection method of the air-sea cross-domain clustering network specifically comprises the following steps:

[0009] 1. Network Model

[0010] The research object of this paper is a three-dimensional air-sea cross-domain network. The network model mainly includes drones, buoys, and underwater sensor nodes. This paper makes the following assumptions about the network:

[0011] (1) Assume that the number of nodes in a fixed monitoring area is large enough and the distance between nodes is always smaller than the communication radius;

[0012] (2) All underwater nodes use acoustic communication to transmit messages, and all above-water nodes use radio communication to transmit messages;

[0013] (3) All common nodes have communication processing capabilities and location awareness capabilities;

[0014] (4) The underwater nodes periodically collect data and send it to the surface buoy through the clustering network. The surface buoy then sends the collected data to the drone. After the drone collects the data, one round of transmission is considered to be completed;

[0015] (5) The drone is regarded as a base station, and the node energy is infinite.

[0016] 2. Algorithm Description

[0017] In order to prevent nodes with insufficient residual energy or far away from the base station from being elected as cluster head nodes, this paper divides the air-sea cross-domain network into several clusters of different sizes, and uses the improved Sky Eagle algorithm to elect cluster heads. Each cluster consists of a cluster head node and ordinary nodes. The nodes in the cluster communicate in a single-hop manner. The cluster head fuses the data and forwards it between clusters. The selection of cluster head nodes is considered from the residual energy of nodes in the network and the communication distance of nodes.

[0018] First, consider the energy factor. Suppose there are N nodes in the network, there are K cluster head nodes in the current round, and the energy of node i is E(n i ), the remaining energy of cluster head node k is E(CH k ), the number of nodes in the current round cluster is C k , the cluster head node is CH k The calculation formula of the fitness function is:

[0019]

[0020] Among them, the value range of ω1 and ω2 is [0,1], and the sum of their values ​​is 1.

[0021]

[0022] Assume that the network contains N nodes, there are K cluster head nodes in the current round, and the number of candidate cluster heads is M. In general, M>>K, then the possible clustering methods are: In order to determine the optimal clustering method, it can be regarded as an optimization problem. This paper uses the improved Tianying optimization algorithm to solve this problem.

[0023] First, an initial population is generated. In the initialization stage, the population is randomly initialized within the search range. Then, the Tent chaotic map is introduced to generate a good initialization population, improve population diversity, prevent the problem of uneven population distribution that may occur during the optimization process, speed up the convergence speed, and improve the convergence accuracy.

[0024] The optimization process of the improved Skyhawk algorithm is represented by four methods: selecting the search space through high-altitude flight with vertical dives, exploring in the divergent search space through contour flight with short gliding attacks, exploiting in the convergent search space through low-altitude flight with slow descent attacks, and swooping and grabbing prey on foot. If t≤2 / 3T, the Skyhawk algorithm can use different behaviors to transfer from the exploration step to the exploitation step; otherwise, the exploitation step will be executed.

[0025] 1 Expand search (X1)

[0026] In the first method (X1), the Skyhawk identifies the prey area and selects the best hunting area by flying high with a vertical bend. At this time, AO lets the explorer from the high altitude fly around to determine the search space area where the prey is located. The mathematical model of this behavior is:

[0027]

[0028] where X1(t+1) is the solution of the t+1th iteration generated by the first search method X1; X best (t) is the best solution obtained before the tth iteration, which reflects the approximate position of the prey; Used to control the expansion search by the number of iterations; rand is a random value between 0 and 1, t and T represent the current iteration and the maximum number of iterations, respectively. Dim is the size of the dimension, and N is the number of candidate solutions (population size).

[0029] 2Narrow your search (X2)

[0030] In the second method (X2), when the prey area is discovered from high altitude, the Skyhawk hovers above the target prey, prepares to land, and then attacks. This method is called contour flight with short glide attack. Here, the AO narrowly explores the selected area of ​​the target prey to prepare for the attack. The mathematical model of this behavior is:

[0031]

[0032] where X2(t+1) is the solution of the t+1th iteration generated by the second search method X2; X R(t) is the random solution obtained in the range [1, N] at the t-th iteration. s = 0.01, μ and ν are respectively subject to N (0, σ 2 ) and N(0,1) Gaussian distributed random numbers, σ is calculated as:

[0033]

[0034] Where β = 1.5, in the calculation formula of X2(t+1), y and x are used to represent the spiral shape in the search, and the calculation is as follows:

[0035] y=r×cos(θ)

[0036] x=r×sin(θ)

[0037] Among them, r = r1 + U × D1, θ = -ω × D1 + θ1, r1 takes a value between 1 and 20, which is used to fix the number of search cycles; U is a value fixed to 0.00565; D1 is an integer from 1 to the dimension of the search space (Dim); ω is a value fixed to 0.005.

[0038] ③Expanded development (X3)

[0039] In the third method (X3), when the eagle accurately specifies the prey area and is ready to land and attack, the eagle descends vertically and makes a preliminary attack to detect the prey's reaction. However, the traditional eagle algorithm is difficult to play a key role in the escape of the entire population from local constraints at this stage. Although the addition of new individuals can cause certain disturbances at this stage, learning from the newly generated eagle individuals ignores the information carried by other individuals in the population. In the later stage of iteration, the degree of assimilation of the population gradually increases and it is easy to fall into the local optimum. Therefore, the social free foraging strategy is introduced, and the current individual X that executes the spiral foraging strategy is i t It not only follows the previous individual, but also moves along the spiral path toward the optimal individual position, effectively traversing the search space and increasing the possibility of escaping the local optimal solution. The improved X3 model is:

[0040]

[0041] Among them: r2 and r3 are [0, 1]; X r t is a random eagle individual in the current iterative population; X b t is the optimal eagle individual in the current iteration population, μ is the spiral coefficient, t and T represent the current iteration and the maximum number of iterations respectively.

[0042] ④ Reduced development (X4)

[0043] In the fourth method (X4), when the eagle approaches the prey, it attacks the prey on land according to its random movement. This method is called "walk and catch prey", and the AO attacks the prey in the last position. The mathematical model of this behavior is:

[0044]

[0045] Among them, X4(t+1) is the solution of the t+1th iteration generated by the third search method X4; X(t) is the current solution of the tth iteration, t and T represent the current iteration and the maximum number of iterations respectively. Based on the new balance method, the algorithm performs local search in the middle of the iteration, making the algorithm more dependent on Levy flight, thereby searching the solution space more thoroughly; in the later stage of the iteration, it is less affected by the random large-scale search of Levy flight, so the optimal solution X obtained in the previous step best (t) and historical solutions X(t) will have a greater impact on the search results of this step, improving the local search capability in the later stages of iteration.

[0046] The cluster head selection method of the air-sea cross-domain clustering network specifically comprises the following steps:

[0047] First, the positions of drones, buoys, and underwater sensor nodes are deployed to obtain a three-dimensional coordinate set. Then, the improved Sky Eagle algorithm is used to elect cluster heads for each round of data transmission. In the underwater network, after the cluster head election is completed, other ordinary nodes join the cluster head nearby to form a cluster group, and then send messages to the cluster head. After data collection and fusion, the cluster head sends the message to the nearby surface buoy. The data received by the buoy is collected by the drone, and one round is considered completed. Repeat this process in the next round until the network energy is exhausted.

[0048] Compared with the related art, the cluster head selection method for the air-sea cross-domain clustering network provided by the present invention has the following beneficial effects:

[0049] The technical solution of the present invention realizes the election of cluster heads by comprehensively considering the remaining energy of nodes in the network and the communication distance of nodes, and considers factors such as node type and heterogeneous characteristics of underwater and above-water networks in the clustering process, which can effectively balance network energy consumption and improve network life cycle. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a schematic diagram of the network model in the present invention;

[0051] Figure 2 This is a flow chart of the cluster head election algorithm in the present invention. DETAILED DESCRIPTION

[0052] The present invention will be further described below in conjunction with the accompanying drawings and implementation modes.

[0053] The network model in the present invention is as follows Figure 1 As shown, the process of the cluster head election algorithm in the present invention is as follows Figure 2 shown.

[0054] The specific implementation of the present invention is:

[0055] At the beginning of each cluster formation phase, all nodes send information about their current energy status and location to the drone node. Based on this information, the drone calculates the average energy level of all nodes, builds a fitness function based on energy and distance factors, and then uses the improved Sky Eagle algorithm to elect cluster heads to obtain the optimal cluster head position and number of cluster heads for the round. After the drone determines the best set of cluster heads and their associated cluster members, it sends information containing the cluster head ID of each node back to all nodes in the network. The node that becomes the cluster head acts as a local control center to coordinate data transmission in its cluster.

[0056] In each cluster, the cluster head node broadcasts the information of the elected cluster head node to other nodes. If the candidate cluster head node receives the information of the elected cluster head node from other nodes in the same layer, it will join this sub-cluster; if it does not receive any broadcast information, it will be elected as the cluster head node of the sub-cluster and broadcast the information of the elected cluster head node. When the ordinary nodes in other clusters receive the election information of the cluster head of the same layer, they will join this sub-cluster; if they have not received the election information of the cluster head node of this layer, they will choose the sub-cluster closest to themselves to join.

[0057] After the underwater network cluster is completed, the ordinary nodes send data to the cluster head node, and the cluster head node aggregates it. Considering that the current direct cross-domain communication technology does not have the conditions for large-scale long-distance use, it is assumed that when the underwater node needs to communicate with the aerial node, it must be forwarded through the surface node. After the underwater sensor cluster head node collects the information, it transmits it to the surface buoy, and then the drone completes the collection of the surface buoy information, which is considered a successful round of data collection. Repeat this process until the network energy is exhausted.

[0058] The above only describes the preferred embodiments of the present invention. It should be understood that the present invention is not limited to the above specific implementation methods. Although the present invention has been disclosed as above in the preferred embodiments, it is not used to limit the present invention. Any technician familiar with the profession can make some changes or modify the technical contents disclosed above into equivalent embodiments with equivalent changes without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent replacement and improvement made to the above embodiments without departing from the content of the technical solution of the present invention, based on the technical essence of the present invention, within the spirit and principles of the present invention, still fall within the protection scope of the technical solution of the present invention.

Claims

1. A cluster head selection method for air-sea cross-domain clustering network, characterized in that: The following steps are involved: S1. Deploy the positions of drones, buoys, and underwater sensor nodes, obtain the three-dimensional coordinate set, and build a three-dimensional air-sea cross-domain network model; S2, based on energy and distance factors, construct a fitness function, use the improved Sky Eagle algorithm to elect the cluster head for each round of data transmission, and determine the cluster head node; S3, ordinary nodes join the nearest cluster head to form a cluster group, and then send messages to the cluster head. After the cluster head node collects and fuses the data, it sends the message to the nearby surface buoy. The data received by the buoy is collected by the drone, completing a round of data transmission; S4. Repeat the above process in the next round until the network energy is exhausted.

2. The cluster head selection method for air-sea cross-domain clustering network as claimed in claim 1, characterized in that: The method further comprises: dividing the air-sea cross-domain network into a plurality of clusters of different sizes to prevent nodes with insufficient residual energy or far from the base station from being elected as cluster head nodes.

3. The cluster head selection method for air-sea cross-domain clustering network as claimed in claim 1, characterized in that: The network model assumptions include: a. The number of nodes in the fixed monitoring area is sufficient, and the distance between nodes is always smaller than the communication radius; b. All underwater nodes use hydroacoustic communication, and all above-water nodes use radio communication; c. All common nodes have communication processing capabilities and location awareness capabilities; d. The underwater nodes periodically collect data and send it to the surface buoy through the clustering network. The buoy then sends the data to the drone. e. The drone is regarded as a base station, and the node energy is infinite.

4. The cluster head selection method for air-sea cross-domain clustering network as claimed in claim 1, characterized in that: The optimization process of the improved Sky Eagle algorithm is represented by the following four methods: Select the search space by high altitude flight with vertical dives; Exploration within the divergent search space through contour flights with short glide attacks; Low-level flight via slow-descent attack is developed within a convergent search space; By swooping down on foot and grabbing prey.

5. The cluster head selection method for air-sea cross-domain clustering network as claimed in claim 1, characterized in that: The calculation formula of the fitness function is: Where N is the number of nodes in the network, K is the number of cluster head nodes in the current round, and E(n i ) is the energy of node i, E(CH k ) is the residual energy of cluster head node k, C k is the number of nodes in the cluster in the current round, CH k is the cluster head node, the value range of ω1 and ω2 is [0,1], and the sum of their values ​​is 1.

Images