A wireless sensor network coverage optimization method and system

Abstract

The application discloses a wireless sensor network coverage optimization method and system. The method comprises the following steps: randomly deploying static nodes and dynamic nodes in a monitoring area, and determining each coverage hole in a wireless sensor network composed of the static nodes as an initial sparrow population; dividing individuals in the sparrow population into different non-dominated levels, and calculating the corresponding crowding distance entropy of each individual; sorting the individuals according to the non-dominated level and the crowding distance entropy of the individuals, and determining explorers, followers and sentinels according to the sorting result; updating the positions of the explorers, the followers and the sentinels, wherein the position updating of the followers adopts a double-sample learning strategy; when a set iteration condition is reached, outputting a global optimal deployment strategy, an optimal coverage rate and a node moving distance; the scheme can effectively balance the target area coverage rate and the node moving distance, reduce the cost in the network deployment stage, and improve the overall network performance.

Description

Technical Field

[0001] This invention relates to the field of wireless sensor network technology, and more specifically to a method and system for optimizing wireless sensor network coverage. Background Technology

[0002] Wireless Sensor Networks (WSNs) are now widely used in target area monitoring, industrial control, and many other fields, playing a crucial role in the development and application of Internet of Things (IoT) technology. However, when deploying sensor nodes in highly unique environments such as forests, deserts, and oceans—environments inaccessible to humans—random scattering is often the only option. This leads to coverage gaps and redundancy, resulting in decreased network performance and wasted resources. Coverage is one of the most fundamental issues in WSN research, determining the quality of service (QoS). A reasonable and effective node deployment strategy can not only reduce network costs but also significantly improve network efficiency. Deploying mobile nodes to improve network coverage has become a mainstream coverage optimization method, but this method significantly increases network deployment costs. Therefore, how to control node mobility costs while improving network coverage is of great significance for the future development of WSN applications.

[0003] Currently, scholars both domestically and internationally have conducted extensive research on coverage optimization of wireless sensor networks in planar areas, achieving some progress. The main research methods fall into two categories: geometric-based coverage optimization and intelligent algorithm-based coverage optimization. Mahboubi et al. proposed a node deployment algorithm based on Voronoi diagrams. During the optimization deployment process, virtual forces applied by the vertices and boundaries of the polygon are considered to determine the new location of the nodes. Simulation results demonstrate that this algorithm can improve coverage. Liu et al. optimized node deployment using an improved virtual force algorithm, considering the interaction forces between nodes on top of the traditional virtual force algorithm. These interaction forces can drive sensor nodes to cover the entire monitoring area, but ultimately, achieving a high coverage rate remains difficult. Wang et al. combined the improved virtual force algorithm with an improved gray wolf algorithm, effectively making the node distribution more uniform and reducing the average movement distance of the nodes. However, this algorithm has high complexity and requires a long running time.

[0004] In recent years, with the development and widespread application of intelligent algorithms, researchers have found that intelligent algorithms can effectively solve the WSN coverage optimization problem, such as particle swarm optimization, whale optimization, firefly optimization, and gray wolf optimization. Intelligent algorithms have advantages such as simple computation and strong search capabilities. Mahnaz et al. proposed an improved whale optimization algorithm for WSN coverage optimization, using methods such as exploration, spiral attack, and bubble net attack to solve complex point coverage problems. Experimental results show that the algorithm can extend the network lifetime, but it is prone to getting trapped in local optima. Kavita et al. proposed a node deployment method based on the gray wolf algorithm, which maximizes the coverage range while ensuring network connectivity by improving the fitness function and population position update method, but this algorithm does not consider the movement cost of nodes. Chao et al. proposed a wireless sensor network coverage optimization method based on an improved artificial bee colony algorithm and a teaching strategy. It combines the global search capability of the artificial bee colony algorithm with the fast convergence of the teaching strategy, and introduces dynamic search balancing to avoid the algorithm getting trapped in local optima. Simulation results show that the algorithm achieves a better balance between global search and local search, but the algorithm has high complexity. Aparajita et al. combined the firefly algorithm with Voronoi diagrams and the K-means algorithm, using the Voronoi cell structure to calculate the optimal sensing radius, and then using the firefly algorithm to find the optimal deployment location of nodes. Simulation results show that this algorithm can improve the coverage and lifespan of wireless sensor networks. Wang et al. proposed an adaptive, discrete-space-oriented wolfpack optimization algorithm for mobile wireless sensor networks. First, they designed an adaptive expansion strategy based on a minimum overlap full coverage model to achieve minimum overlap and gapless coverage of the monitoring area. Second, they improved the target node probability matrix and adaptive step size. Simulation results show that the algorithm has better convergence speed and global optimization capability. Zhu et al. proposed a coverage optimization method based on an improved hybrid strategy weed algorithm. By improving different stages of the weed algorithm, they improved the search capability and convergence speed of the algorithm, thereby improving the network coverage. Teng Zhijun et al. proposed a fireworks virtual force hybrid coverage strategy based on μ-law explosion operator in WSN. The algorithm improves the optimization ability and accuracy by using the μ-law idea in non-uniform quantization. At the same time, the virtual force algorithm is introduced to accelerate the evolution of fireworks population and improve the node deployment efficiency. However, the algorithm does not optimize the movement distance of nodes during the network coverage optimization process.

[0005] Wang Zhendong et al. used an improved sparrow search algorithm to improve node coverage in wireless sensor networks. The improved algorithm showed improvements in search accuracy and convergence speed, but it neglected the impact of the initial sparrow population's position distribution on algorithm performance. Duan Jin et al. proposed an improved sparrow search algorithm to optimize wireless sensor network coverage. Addressing the uncertainty caused by the random initialization of the population in the standard sparrow search algorithm, they introduced an initial population based on a set of optimal points and proposed a comprehensive average index for the coverage problem. Results showed that initializing the initial position of the population using a set of optimal points effectively improved the algorithm's convergence speed and accuracy; however, this study did not fully consider the impact of the movement distance of mobile nodes on the network coverage optimization cost. Wu Juan et al. proposed a multi-objective coverage optimization method based on an improved sparrow search algorithm. This method not only improved network coverage by improving the sparrow search algorithm but also restored nodes with large movement distances but small improvements in network coverage to their initial positions. Results showed that this algorithm could balance network coverage and node movement costs to a certain extent, but its method for controlling node movement distance was still insufficient, failing to reduce the overall movement distance of all nodes. Summary of the Invention

[0006] The technical problem to be solved by this invention is how to effectively balance the network coverage rate and the dynamic node movement cost in the process of optimizing wireless sensor network node coverage.

[0007] The present invention solves the above-mentioned technical problems through the following technical means:

[0008] A method for optimizing wireless sensor network coverage is proposed, the method comprising:

[0009] Static and dynamic nodes are randomly deployed within the monitoring area, and each coverage gap in the wireless sensor network composed of the static nodes is identified as the initial sparrow population.

[0010] Individuals in the sparrow population are divided into different non-dominant levels, and the crowding distance entropy corresponding to each individual is calculated.

[0011] Individuals are ranked according to their non-dominance level and crowding distance entropy, and explorers, followers, and vigilants are determined based on the ranking results.

[0012] Update the positions of the explorer, the follower, and the vigilant, wherein the position update of the follower employs a two-sample learning strategy;

[0013] When the set iteration conditions are met, the system outputs the globally optimal deployment strategy, optimal coverage, and node movement distance.

[0014] Furthermore, the step of randomly deploying static and dynamic nodes within the monitoring area and identifying coverage gaps in the wireless sensor network composed of the static nodes as the initial sparrow population includes:

[0015] Randomly deploy static and dynamic nodes within the monitoring area and obtain the coordinates of all nodes;

[0016] The Delaunay triangulation strategy is used to find coverage gaps in the wireless sensor network composed of the static nodes as the initial sparrow population.

[0017] Furthermore, the step of employing the Delaunay triangulation strategy to find coverage gaps in the wireless sensor network composed of the static nodes as the initial sparrow population includes:

[0018] The wireless sensor network composed of static nodes is divided using the Delaunay triangulation strategy to obtain a triangular network;

[0019] Based on the relationship between the triangles in the triangular mesh and the uncovered areas, each coverage gap in the triangular mesh is determined as the initial sparrow population. The relationship between the triangles and the uncovered areas is as follows:

[0020] when At that time, there is no uncovered area inside the triangle;

[0021] when When, if and only if the triangle is an acute triangle, there exists an uncovered region inside the triangle that is entirely inside the triangle;

[0022] when If the triangle is acute or right, there is an uncovered area inside the triangle. If the triangle is obtuse and the perpendicular distance from point C to AB is greater than R, there is an uncovered area inside the triangle.

[0023] in, A , B , C These are three vertices of the same triangle in a triangular mesh formed after the statically constructed network is divided using a triangulation strategy. R This represents the sensing radius of the sensor node. d ( A , B Let be the distance between vertices A and B. d ( A Let C be the distance between vertices A and C. d ( B ,C) is the distance between vertices B and C.

[0024] Furthermore, after employing the Delaunay triangulation strategy to find coverage gaps in the wireless sensor network composed of the static nodes, the method further includes:

[0025] Calculate variables Cl To measure the size of the area of ​​the cover hole inside each triangle, variables Cl The calculation formula is:

[0026]

[0027] In the formula, For the area of ​​the corresponding triangle, The total area of ​​the three sectors inside the triangle. This represents the total area of ​​the repeated sectors inside the triangle;

[0028] according to Cl The values ​​are used to sort the coverage vulnerabilities inside each triangle, and the top ones are selected. N The location of the coverage vulnerability is used as the initial sparrow population.

[0029] Furthermore, the step of dividing individuals in the sparrow population into different non-dominant ranks and calculating the crowding distance entropy for each individual includes:

[0030] The individuals in the sparrow population are sorted by non-dominance level to obtain the non-dominance level results;

[0031] Based on the non-dominated rank results, the crowding distance entropy for all individuals in each non-dominated rank is calculated, wherein the formula for calculating the crowding distance entropy is:

[0032]

[0033] In the formula, h For the first i The number of individuals in the non-dominant hierarchy of a population. j Is with the first i Individuals in the same dominance hierarchy within a population To solve i +1 corresponds to the reciprocal of the coverage of the node deployment scheme. , To solve i -1 is the reciprocal of the coverage of the node deployment scheme. To solve i The average node movement distance for the node deployment scheme corresponding to +1. To solve i -1 corresponds to the average node movement distance of the node deployment scheme. To solvej The reciprocal of the coverage of the corresponding node deployment scheme. To solve j The average node movement distance of the corresponding node deployment scheme, To solve i The reciprocal of the coverage of the corresponding node deployment scheme. , To solve i The average node movement distance of the corresponding node deployment scheme, To solve i The crowding distance entropy.

[0034] Furthermore, the step of dividing individuals in the sparrow population into different non-dominant ranks includes:

[0035] 1) Current round number t When =1, find all n ( p Store the solutions where ) = 0 in a set. F (1), n ( p ) is the dominant solution p Quantity;

[0036] 2) For sets F ( t All individuals in ) j Examine individuals j The set under control S ( j ), S ( j All individuals in ) k Attributes n ( k ) = n ( k -1;

[0037] 3) t = t +1, will n ( p The solutions where ) = 0 are stored in a set. F ( t );

[0038] 4) Repeat steps 2) to 3) until all solutions are stored in the corresponding set, and obtain the non-dominated level result.

[0039] Furthermore, the process of ranking individuals based on their non-dominance level and crowding distance entropy, and determining explorers, followers, and vigilant individuals based on the ranking results, includes:

[0040] Individuals are ranked according to their non-dominance level and crowding distance entropy. The ranking criteria are as follows: for two individuals with different non-dominance levels, the one with the lower non-dominance level is better than the one with the higher non-dominance level; for individuals with the same dominance level, the one with the larger crowding distance entropy is better than the one with the smaller crowding distance entropy.

[0041] Based on the sorting results, explorers, followers, and vigilants are determined in the population according to a set ratio.

[0042] Furthermore, the method also includes:

[0043] The proportion of explorers in the population is dynamically adjusted. PD ,in:

[0044]

[0045]

[0046] In the formula, T The maximum number of iterations, t This represents the current iteration number.

[0047] Furthermore, the position update formula for the follower is:

[0048]

[0049] In the formula, This is the optimal position within the current sparrow population. This is the second-best position in the current sparrow population. This is the worst position in the current sparrow population. It is currently the second worst-performing species in the sparrow population. When the number of iterations is t When +1 is added, it represents the optimal position within the current sparrow population. When the number of iterations is t At that time, the first i The sparrow in the first j Location information in the dimension 1× d A matrix, where each element is randomly assigned the value 1 or -1, and , When the number of iterations is t At +1, the first i The sparrow in the first j Location information in the dimension Q For random numbers that follow a normal distribution, N Let be the population size for the sparrow search algorithm. i The number is the number of the sparrow. j As a dimension,r This is the step size control parameter, and .

[0050] Furthermore, this invention also proposes a wireless sensor network coverage optimization system, the system comprising:

[0051] The initial population determination module is used to randomly deploy static and dynamic nodes within the monitoring area and determine each coverage gap in the wireless sensor network composed of the static nodes as the initial sparrow population.

[0052] The ranking module is used to divide individuals in the sparrow population into different non-dominant ranks and calculate the crowding distance entropy for each individual.

[0053] The sorting module is used to sort individuals based on their non-dominance level and crowding distance entropy, and to determine explorers, followers, and vigilant individuals based on the sorting results.

[0054] The location update module is used to update the positions of the explorer, the follower, and the vigilant, wherein the position update of the follower adopts a two-sample learning strategy.

[0055] The output module is used to output the globally optimal deployment strategy, optimal coverage, and node movement distance when the set iteration conditions are met.

[0056] The advantages of this invention are:

[0057] (1) This invention locates the initial coverage gaps in the network after random deployment of static nodes, and uses them as the initial population of the Sparrow Search algorithm. This can quickly locate the population near the target area, improving the convergence speed and accuracy of the algorithm. By sorting individuals based on non-dominant level and crowding distance entropy, the quality and quantity of the explorer population in the Sparrow Search algorithm are optimized, improving the global search capability of the algorithm and solving the problem of the algorithm's difficulty in balancing multiple targets. Furthermore, a two-sample learning strategy is used during position updates to improve the follower position update method, enhancing the algorithm's ability to escape local optima. Thus, the target area coverage and node movement distance can be effectively balanced, reducing the cost of network deployment and improving the overall network performance.

[0058] (2) The approximate location of the coverage vulnerability in the current network can be quickly located based on the relationship between the Delaunay triangle and the uncovered area. However, considering that the size of the coverage vulnerability inside the divided triangular network is significantly different, this invention filters the coverage vulnerability according to the area size of the coverage vulnerability to ensure the accuracy of the initial population.

[0059] (3) By utilizing an improved method for calculating crowding distance entropy, this invention can balance the solution with coverage and movement distance in the ranking process.

[0060] (4) In traditional algorithms, the number of explorers is generally set to a fixed value between 10% and 20% of the total population. A fixed number cannot maximize the role of explorers in the entire population. In this invention, the proportion of explorers to the total population is set as a dynamic weight coefficient. In the early stage of the algorithm, the number of explorers is 20% of the entire population. As the algorithm iterates, it is gradually reduced to 10%, which can effectively give full play to the role of the explorer population.

[0061] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0062] Figure 1 This is a flowchart illustrating the wireless sensor network coverage optimization method proposed in an embodiment of the present invention;

[0063] Figure 2 This is a complete flowchart of the wireless sensor network coverage optimization method proposed in this embodiment of the invention;

[0064] Figure 3 This is a schematic diagram of the structure of the triangular network in an embodiment of the present invention, wherein (a) is the triangular network obtained by dividing the static nodes in the monitoring area using the Delaunay triangulation strategy, and (b) is a new triangular network formed by adding the boundary points of the monitoring area to the triangular network composed of static nodes.

[0065] Figure 4 This is a schematic diagram illustrating the coverage vulnerability in an embodiment of the present invention;

[0066] Figure 5 This is a schematic diagram of the non-dominant surface hierarchy in an embodiment of the present invention;

[0067] Figure 6 This is a schematic diagram of the two-sample learning strategy in an embodiment of the present invention;

[0068] Figure 7 This is a schematic diagram of the structure of the wireless sensor network coverage optimization system proposed in an embodiment of the present invention;

[0069] Figure 8 This is a comparative diagram of coverage optimization with a monitoring range of 100m×100m in an embodiment of the present invention. (a) is 100m. Initial deployment diagram of 100m dynamic and static nodes, (b) is 100 The optimized node distribution diagram of the 100m moving node, (c) shows the relationship between network coverage and iteration number, and (d) shows a comparison of the moving distance of each node;

[0070] Figure 9This is a comparative diagram of WSN coverage optimization in a 50m×50m area monitored in this embodiment of the invention. (a) is a 50m×50m area. Initial deployment diagram of 50m dynamic and static nodes, (b) is 50 The node distribution diagram after optimization of the 50m mobile node, (c) is the convergence curve of network coverage optimization, and (d) is a schematic diagram comparing the moving distance of each node.

[0071] Figure 10 This is a comparative diagram of coverage optimization with a monitoring range of 20m×20m in an embodiment of the present invention. (a) is 20m. Initial deployment diagram of 20m dynamic and static nodes, (b) is 20 The node distribution diagram after optimization of the 20m moving node, (c) shows the relationship between network coverage and iteration number, and (d) shows the comparison of the moving distance of each node;

[0072] Figure 11 This is a comparison chart of coverage rates for different numbers of dynamic nodes in embodiments of the present invention. Detailed Implementation

[0073] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0074] like Figures 1 to 2 As shown, the first embodiment of the present invention proposes a method for optimizing the coverage of a wireless sensor network, the method comprising the following steps:

[0075] S10. Randomly deploy static and dynamic nodes within the monitoring area, and determine the locations of coverage gaps in the wireless sensor network composed of the static nodes as the initial sparrow population.

[0076] It should be noted that in biomimetic intelligent algorithms such as SSA, random initialization is typically used to determine the initial population. This is to ensure a thorough search of the solution region in the early stages of the algorithm, aiming to obtain the global optimum. The use of random methods such as chaotic mapping to determine the initial population demonstrates its crucial importance to the overall algorithm performance, influencing both the global search effect and the convergence speed. This embodiment identifies initial coverage vulnerabilities in the network after random deployment of static nodes and uses these vulnerabilities as the initial population for the Sparrow Search algorithm. This allows for rapid localization of the population to the vicinity of the target region, improving the algorithm's convergence speed and accuracy.

[0077] S20. Divide the individuals in the sparrow population into different non-dominant levels and calculate the crowding distance entropy for each individual.

[0078] S30. Sort individuals according to their non-dominance level and crowding distance entropy, and determine explorers, followers and vigilant based on the sorting results;

[0079] It should be noted that in the standard SSA algorithm, explorers are typically composed of high-energy individuals from the population. They are responsible for searching food-rich areas within the population, providing search guidance for followers, and playing a crucial role in the overall search performance. Therefore, both the quality and quantity of the explorer population significantly impact the overall algorithm performance. Higher quality and larger numbers of explorers result in stronger global optimization capabilities but slow down convergence. Only by balancing the quality and quantity of explorers in the SSA algorithm can its performance be optimized. This embodiment optimizes the quality and quantity of the explorer population in the sparrow search algorithm by ranking individuals based on non-dominated levels and crowding distance entropy, thereby improving the algorithm's global search capability and addressing the problem of balancing multiple objectives.

[0080] During the foraging process, 10% to 20% of sparrows will act as watchdogs. When danger approaches, all sparrows will abandon their current food and move to a new location.

[0081] S40. Update the positions of the explorer, the follower, and the vigilant, wherein the position update of the follower adopts a two-sample learning strategy;

[0082] It's important to note that in standard SSA, followers primarily determine their position updates based on the explorers with the highest and lowest fitness values ​​globally. The goal is to randomly find a position near the current best position, or if a follower's current position is close to the worst position, resulting in a low fitness value, it needs to move to another position. While this method of solely learning the current best or worst position can improve the algorithm's search speed to some extent, it also leads to the loss of some important positional information within the population, increasing the likelihood of the algorithm getting trapped in local optima.

[0083] This embodiment uses a two-sample learning strategy to improve the follower position update method, which can improve the algorithm's ability to escape local optima.

[0084] S50. When the set iteration conditions are met, output the global optimal deployment strategy, optimal coverage, and node movement distance.

[0085] Furthermore, in this embodiment, the n sensor nodes in the WSN all have the same structure and attributes, and each sensor node has a defined sensing radius. and communication radius To ensure network connectivity, the communication radius of a node is typically set to twice its sensing radius. Assume the set of wireless sensor nodes is... Any sensor node in the node set The coordinates can be represented as The set of monitoring nodes is Any monitoring point within the monitoring area The coordinates can be represented as The Euclidean distance between the sensor node and the monitoring nodes within the monitoring area is:

[0086]

[0087] The probability sensing model used in this embodiment is a Boolean sensing model, so the monitoring node The node The probability of perception is:

[0088]

[0089] All sensor nodes within the monitoring area are monitored by the monitoring node. The joint sensing probability is:

[0090]

[0091] In the formula: This represents all sensor nodes within the monitoring area. Assume the monitoring area is rectangular, with side lengths of [missing information]. A , B The area of ​​the monitoring area is A B The rectangular monitoring area is divided into... A B A grid of equal area is formed, with monitoring node m located at the center of the grid. The joint sensing probability of the monitoring points can be calculated using the above formula, and the sum of these probabilities gives the coverage area. Coverage rate. The calculation formula is as follows:

[0092]

[0093] Assumption The initial position of node i is The target location is ,but S The total distance traveled by all nodes is:

[0094]

[0095] This embodiment aims to reduce the movement distance of dynamic nodes while ensuring network coverage.

[0096] In one embodiment, step S10, which involves randomly deploying static and dynamic nodes within the monitoring area and identifying coverage gaps in the wireless sensor network composed of the static nodes as the initial sparrow population, includes the following steps:

[0097] S11. Randomly deploy static and dynamic nodes within the monitoring area and obtain the coordinates of all nodes;

[0098] S12. Use the Delaunay triangulation strategy to find each coverage gap in the wireless sensor network composed of the static nodes as the initial sparrow population.

[0099] In one embodiment, step S12, which uses the Delaunay triangulation strategy to find coverage gaps in the wireless sensor network composed of the static nodes as the initial sparrow population, includes the following steps:

[0100] S121. The wireless sensor network composed of the static nodes is divided using the Delaunay triangulation strategy to obtain a triangular network.

[0101] S122. Based on the relationship between the triangles in the triangular mesh and the uncovered areas, determine each coverage gap in the triangular mesh as the initial sparrow population, wherein the relationship between the triangles and the uncovered areas is as follows:

[0102] when At that time, there is no uncovered area inside the triangle;

[0103] when When, if and only if the triangle is an acute triangle, there exists an uncovered region inside the triangle that is entirely inside the triangle;

[0104] when If the triangle is acute or right, there is an uncovered area inside the triangle. If the triangle is obtuse and the perpendicular distance from point C to AB is greater than R, there is an uncovered area inside the triangle.

[0105] in, A , B , C These are three vertices of the same triangle in a triangular mesh formed after the statically constructed network is divided using a triangulation strategy. R This represents the sensing radius of the sensor node. d ( A , B Let be the distance between vertices A and B.d ( A Let C be the distance between vertices A and C. d ( B ,C) is the distance between vertices B and C.

[0106] Specifically, after dividing the static nodes within the monitoring area using the Delaunay triangulation strategy, the triangular network is as follows: Figure 3 As shown in (a). Adding the boundary points of the monitoring area to the triangular network formed by the static nodes creates a new triangular network, as shown in (a). Figure 3 As shown in (b), in a network composed of all static nodes, the uncovered areas are mainly located inside some triangles. The relationship between the Delaunay triangle and the uncovered areas can be divided into the three types mentioned above, where the uncovered areas refer to coverage gaps.

[0107] This embodiment uses the Delaunay triangulation strategy to find the uncovered areas in the network composed of static nodes as the initial population of the sparrow search algorithm. This can quickly locate the population near the target area, improving the convergence speed and accuracy of the algorithm.

[0108] In one embodiment, after step S12: finding coverage gaps in the wireless sensor network composed of the static nodes using the Delaunay triangulation strategy, the method further includes the following steps:

[0109] S12', Calculate variables Cl To measure the size of the area of ​​the cover hole inside each triangle, variables Cl The calculation formula is:

[0110]

[0111] In the formula, For the area of ​​the corresponding triangle, The total area of ​​the three sectors inside the triangle. This represents the total area of ​​the repeated sectors inside the triangle. The calculation formula is:

[0112]

[0113] In the formula, d i The distance between two vertices of the triangle. R denoted as the sensing radius of the node.

[0114] S12'', according to Cl The values ​​are used to sort the coverage vulnerabilities inside each triangle, and the top ones are selected. N The location of the coverage vulnerability is used as the initial sparrow population.

[0115] It should be noted that the approximate location of coverage vulnerabilities in the current network can be quickly determined based on the relationship between the Delaunay triangle and the uncovered area. Considering the significant differences in the size of coverage vulnerabilities within the divided triangular network, further filtering of the initial population's location is necessary. The sensor nodes have fixed and identical sensing radii. The size of coverage vulnerabilities within each triangle forming the triangular network is related to the positions of the three vertices of the triangle, as analyzed below:

[0116] The area of ​​the vulnerability is related to the area of ​​the triangle, the area of ​​the sector, and the overlapping area between sectors, respectively. Figure 4 As shown. Since the sum of the interior angles of a triangle is 180 degrees, the sum of the areas of the sectors inside any triangle is the same; the area of ​​a triangle can be calculated from the coordinates of its three nodes; the area of ​​the repeated sectors inside a triangle is related to the distance between the nodes and the perceptual radius of the nodes, and can be calculated using mathematical formulas. This is achieved by defining variables. Cl To measure the size of the coverage area inside each triangle, according to Cl Sort the values ​​by size and select the top ones. N The vulnerability locations are used as the initial population. N denoted as the population size for the IM-DTSSA algorithm.

[0117] It should be noted that, according to Cl The value sorts the triangles, which means sorting the size of the uncovered areas within the triangles. After filtering, the search efficiency of the initial population of the algorithm can be effectively improved because it avoids the algorithm spending a long time on global search in the initial stage.

[0118] In one embodiment, it is set that for the solution P Calculate its two key attributes n ( p )and S ( p ), n ( p ) is the dominant solution P Quantity, S ( p ) is being solved P The set of dominated solutions, in step S20, divides the individuals in the sparrow population into different non-dominated levels, specifically including the following steps:

[0119] 1) Current round number t When =1, find all n ( p Store the solutions where ) = 0 in a set. F (1), n ( p ) is the dominant solutionp Quantity;

[0120] 2) For sets F ( t All individuals in ) j Examine individuals j The set under control S ( j ), S ( j All individuals in ) k Attributes n ( k ) = n ( k -1;

[0121] 3) t = t +1, will n ( p The solutions where ) = 0 are stored in a set. F ( t );

[0122] 4) Repeat steps 2) to 3) until all solutions are stored in the corresponding set, and obtain the non-dominated level result.

[0123] It should be noted that when solving a specific problem, the energy level of an individual in the population depends on its fitness value. In multi-objective optimization, how to evaluate the fitness value is a pressing issue. This embodiment introduces the idea of ​​Non-Dominated Ranking (NSGA2). First, the sparrow population is ranked by non-dominated level. Then, further ranking is performed by calculating the crowding distance entropy of individuals within the same non-dominated level. In NSGA2, for two solutions... x and y If for any i =1,2,…, m All have f i ( x )≤ f i ( y ), and exists i ,make f i ( x )< f i ( y ), then it is called x Dominate y .in, f i ( x ) is in the iFitness values ​​on each objective function. All solutions are partitioned into different dominance surfaces, as illustrated below. Figure 5 As shown. Figure 5 As can be seen, the number of dominant solutions A and B is 0, therefore solutions A and B are both on the first-order non-dominated surface. Solution C is on the second-order non-dominated surface, and it can be seen that both solutions A and B can dominate solution C. Therefore, solutions on different dominated surfaces can be ordered by non-dominated level.

[0124] In one embodiment, in step S20, the crowding distance entropy corresponding to each individual is calculated, and the formula is expressed as follows:

[0125]

[0126] In the formula, h For the first i The number of individuals in the non-dominant hierarchy of a population. j Is with the first i Individuals in the same dominance hierarchy within a population To solve i +1 corresponds to the reciprocal of the coverage of the node deployment scheme. , To solve i -1 is the reciprocal of the coverage of the node deployment scheme. To solve i The average node movement distance for the node deployment scheme corresponding to +1. To solve i -1 corresponds to the average node movement distance of the node deployment scheme. To solve j The reciprocal of the coverage of the corresponding node deployment scheme. To solve j The average node movement distance of the corresponding node deployment scheme, To solve i The reciprocal of the coverage of the corresponding node deployment scheme. , To solve i The average node movement distance of the corresponding node deployment scheme, To solve i The crowding distance entropy.

[0127] It should be noted that the NSGA2 algorithm uses the crowding distance entropy CD for each solution. i The density of solutions on the same dominant surface is analyzed by measuring the balance and diffusion of solutions. However, this embodiment modifies the crowding distance entropy to prioritize solutions that balance coverage and movement distance during the sorting process.

[0128] In one embodiment, step S30, which involves ranking individuals based on their non-dominance level and crowding distance entropy, and determining explorers, followers, and vigilant individuals based on the ranking results, includes the following steps:

[0129] S31. Sort individuals according to their non-dominance level and crowding distance entropy to obtain the sorting results. The sorting criteria are as follows: for two individuals with different non-dominance levels, the one with the lower non-dominance level is better than the one with the higher non-dominance level; for individuals with the same dominance level, the one with the larger crowding distance entropy is better than the one with the smaller crowding distance entropy.

[0130] S32. Based on the sorting results, determine explorers, followers, and vigilants in the population according to the set proportions.

[0131] It should be noted that in this embodiment, the explorer population can be set to consist of the top PD individuals sorted according to the non-dominated sorting algorithm, and the remaining individuals can form the follower population.

[0132] Furthermore, in the standard SSA algorithm, the number of explorers is generally set to a fixed value between 10% and 20% of the total population. However, a fixed number cannot maximize the role of explorers in the entire population. Therefore, this embodiment sets the proportion of explorers to the total population as a dynamic weighting coefficient to achieve dynamic adjustment, specifically as follows:

[0133]

[0134]

[0135] In the formula, T The maximum number of iterations, t This represents the current iteration number.

[0136] It should be noted that in the early stages of the algorithm's operation, the requirements for global search are higher, requiring more explorers to explore globally optimal positions to improve the algorithm's early global optimization ability. In the middle and later stages of the algorithm's operation, a smaller number of explorers can ensure a certain level of global exploration ability while also improving the algorithm's local optimization ability in the later stages. (Based on the above formula...) PD The adjustment formula can achieve this: in the early stages of the algorithm, the number of explorers is 20% of the entire population, and as the algorithm iterates, it is gradually reduced to 10%, which can effectively utilize the role of the explorer population.

[0137] In one embodiment, in step S40: updating the positions of the explorer, the follower, and the watcher, wherein the position update of the follower employs a two-sample learning strategy, including:

[0138] Explorer position update: Update the explorer's position using the explorer position update formula, which is expressed as:

[0139]

[0140] In the formula, t is the current iteration number. The maximum number of iterations, For the t generation i Only sparrows in the first j Dimensional position, random number , For random numbers that follow a normal distribution, for All-1 matrix, warning threshold Safety value ,like When the bird indicates that no predators have been found in its current foraging environment, it can conduct a wide search; if When this occurs, it indicates that the watchdog has spotted the predator and issued an alarm, and the sparrows need to move to a safe location.

[0141] Follower position update: The follower position is updated using a two-sample learning strategy, expressed by the formula:

[0142]

[0143] In the formula, This is the optimal position within the current sparrow population. This is the second-best position in the current sparrow population. This is the worst position in the current sparrow population. It is currently the second worst-performing species in the sparrow population. When the number of iterations is t When +1 is added, it represents the optimal position within the current sparrow population. When the number of iterations is t At that time, the first i The sparrow in the first j Location information in the dimension 1× d A matrix, where each element is randomly assigned the value 1 or -1, and , When the number of iterations is t At +1, the first i The sparrow in the first j Location information in the dimension Q For random numbers that follow a normal distribution, N Let be the population size for the sparrow search algorithm. i The number is the number of the sparrow. j As a dimension, r This is the step size control parameter, and .

[0144] Watchdog Location Update: During foraging, 10% to 20% of sparrows will act as watchdogs. When danger approaches, all sparrows will abandon their current food and move to a new location. The watchdog update process is as follows:

[0145]

[0146] In the formula, This indicates the current globally optimal position. It is a step size control parameter that follows a standard normal distribution. It is a random number. , , These represent the current fitness value of the sparrow, the fitness value of the global optimal solution, and the fitness value of the global worst solution, respectively. It is a very small constant, and we need to avoid the denominator being 0.

[0147] It should be noted that, in order to enable followers to better learn the location information of the optimized explorer population, this paper improves the follower location update formula in the standard SSA by using a two-sample learning strategy. For example... Figure 6 As shown, followers i Learning solely from the current global best position p or the current global second best position m will both lead to a decrease in the diversity of the current population and result in getting stuck in a local optimum.

[0148] In order to enable followers i Capable of effectively exploring to the global optimum. k ,make i At the same time to p and m Learning involves using a combination of information from the two currently better explorer positions as the direction for the next position update, giving the algorithm a strong ability to escape local optima in the mid-to-late stages. On the other hand, to help followers with lower fitness values ​​avoid the worst positions in the global search space when flying to other locations to forage, the algorithm can consider both the worst and second-worst global positions simultaneously, preventing some followers from flying from one poor position to another. At this point, followers can comprehensively consider the positional information of the current explorer population, increasing the probability of exploring unsearched locations in the current search space and further improving their search capabilities for target areas.

[0149] It should be noted that, as Figure 2 As shown, the overall process of the wireless sensor network coverage optimization method proposed in this embodiment is as follows:

[0150] (1) Initialize algorithm parameters, including population size N Number of sensor nodesn Explorer ratio PD Maximum number of iterations Safety values ​​ST, r1, r2, r3, and r4.

[0151] (2) The line sensor network randomly deploys static and dynamic nodes within the monitoring area, obtains the coordinates of all nodes, and calculates the initial coverage rate.

[0152] (3) Establish a triangular network and calculate the corresponding triangles. Cl The values ​​are sorted, and the centroid positions of the first N triangles are selected as the initial population for the algorithm.

[0153] (4) Based on the idea of ​​the non-dominated sorting algorithm, individuals are divided into different non-dominated levels, and the crowding distance entropy corresponding to each individual is calculated. The individuals are sorted according to their non-dominated levels and crowding distance entropy, and the top PD individuals are selected as explorers.

[0154] (5) Update the location of the explorer and the location of the follower.

[0155] (6) If the watcher detects danger, the position of all individuals in the population is updated.

[0156] (7) Determine whether the algorithm has reached the maximum number of iterations. If it has, output the global optimal deployment strategy, optimal coverage and node movement distance; otherwise, return to step (4).

[0157] It should be noted that the initial Sparrow Search algorithm solves the problem by searching and updating explorers, followers, and watchdogs. Compared with intelligent algorithms such as Particle Swarm Optimization and Ant Colony Optimization, it has the advantages of high search accuracy, fast convergence speed, and good stability. However, in solving multi-objective problems, the Sparrow Search algorithm struggles to balance multiple objectives, and its diversity gradually decreases in the later stages of iteration. This embodiment proposes a WSN coverage optimization algorithm to improve the Sparrow Search algorithm with multiple strategies. First, the Delaunay triangulation strategy is used to find uncovered areas in the network composed of static nodes, which are then used as the initial population for the Sparrow Search algorithm. Second, the idea of ​​non-dominated sorting is introduced to improve the quality of the explorer population and solve the problem of the algorithm struggling to balance multiple objectives. Finally, a two-sample learning strategy is used to optimize the update formula of the followers, avoiding the problem of the algorithm easily getting trapped in local optima in the later stages of operation.

[0158] In addition, such as Figure 7 As shown, the second embodiment of the present invention also proposes a wireless sensor network coverage optimization system, the system comprising:

[0159] The initial population determination module 10 is used to randomly deploy static and dynamic nodes within the monitoring area and determine each coverage gap in the wireless sensor network composed of the static nodes as the initial sparrow population.

[0160] The ranking module 20 is used to divide individuals in the sparrow population into different non-dominant ranks and calculate the crowding distance entropy for each individual.

[0161] The sorting module 30 is used to sort individuals based on their non-dominance level and crowding distance entropy, and to determine explorers, followers and vigilant individuals based on the sorting results.

[0162] The location update module 40 is used to update the positions of the explorer, the follower and the vigilant, wherein the position update of the follower adopts a two-sample learning strategy.

[0163] The output module is used to output the globally optimal deployment strategy, optimal coverage, and node movement distance when the set iteration conditions are met.

[0164] This embodiment identifies initial coverage vulnerabilities in the network after random deployment of static nodes, filters them, and uses them as the initial population for the Sparrow Search algorithm. This allows for rapid positioning of the population near the target area, improving the algorithm's convergence speed and accuracy. By ranking individuals based on non-dominated levels and crowding distance entropy, the quality and quantity of the explorer population in the Sparrow Search algorithm are optimized, enhancing the algorithm's global search capability and addressing the problem of balancing multiple objectives. Furthermore, a two-sample learning strategy is used during position updates to improve the follower position update method, enhancing the algorithm's ability to escape local optima.

[0165] In one embodiment, the initial population determination module 10 specifically includes:

[0166] The coordinate acquisition unit is used to randomly deploy static and dynamic nodes within the monitoring area and acquire the coordinate positions of all nodes.

[0167] The coverage vulnerability location determination unit is used to use the Delaunay triangulation strategy to find each coverage vulnerability in the wireless sensor network composed of the static nodes as the initial sparrow population.

[0168] In one embodiment, the coverage vulnerability location determination unit is specifically used to perform the following steps:

[0169] The wireless sensor network composed of static nodes is divided using the Delaunay triangulation strategy to obtain a triangular network;

[0170] Based on the relationship between the triangles in the triangular mesh and the uncovered areas, each coverage gap in the triangular mesh is determined as the initial sparrow population. The relationship between the triangles and the uncovered areas is as follows:

[0171] when At that time, there is no uncovered area inside the triangle;

[0172] when When, if and only if the triangle is an acute triangle, there exists an uncovered region inside the triangle that is entirely inside the triangle;

[0173] when If the triangle is acute or right, there is an uncovered area inside the triangle. If the triangle is obtuse and the perpendicular distance from point C to AB is greater than R, there is an uncovered area inside the triangle.

[0174] in, A , B , C These are three vertices of the same triangle in a triangular mesh formed after the statically constructed network is divided using a triangulation strategy. R This represents the sensing radius of the sensor node. d ( A , B Let be the distance between vertices A and B. d ( A Let C be the distance between vertices A and C. d ( B ,C) is the distance between vertices B and C.

[0175] In one embodiment, the system further includes a vulnerability screening module for performing the following steps:

[0176] Calculate variables Cl To measure the size of the area of ​​the cover hole inside each triangle, variables Cl The calculation formula is:

[0177]

[0178] In the formula, For the area of ​​the corresponding triangle, The total area of ​​the three sectors inside the triangle. This represents the total area of ​​the repeated sectors inside the triangle;

[0179] according to Cl The values ​​are used to sort the coverage vulnerabilities inside each triangle, and the top ones are selected. N The location of the coverage vulnerability is used as the initial sparrow population.

[0180] In one embodiment, the level classification module 20 includes:

[0181] The non-dominant ranking unit is used to rank the individuals of the sparrow population according to their non-dominant ranking to obtain the non-dominant ranking result.

[0182] The crowding distance calculation unit is used to calculate the crowding distance entropy of all individuals in each non-dominated level based on the non-dominated level results. The formula for calculating the crowding distance entropy is as follows:

[0183]

[0184] In the formula, h For the first i The number of individuals in the non-dominant hierarchy of a population. j Is with the first i Individuals in the same dominance hierarchy within a population To solve i +1 corresponds to the reciprocal of the coverage of the node deployment scheme. , To solve i -1 is the reciprocal of the coverage of the node deployment scheme. To solve i The average node movement distance for the node deployment scheme corresponding to +1. To solve i -1 corresponds to the average node movement distance of the node deployment scheme. To solve j The reciprocal of the coverage of the corresponding node deployment scheme. To solve j The average node movement distance of the corresponding node deployment scheme, To solve i The reciprocal of the coverage of the corresponding node deployment scheme. , To solve i The average node movement distance of the corresponding node deployment scheme, To solve i The crowding distance entropy.

[0185] In one embodiment, the non-dominant ranking unit is configured to perform the following steps:

[0186] 1) Current round number t When =1, find all n ( p Store the solutions where ) = 0 in a set. F (1), n ( p ) is the dominant solution p Quantity;

[0187] 2) For sets F ( t All individuals in ) j Examine individuals j The set under control S ( j ), S ( j All individuals in ) k Attributes n ( k ) = n ( k -1;

[0188] 3) t = t +1, will n ( p The solutions where ) = 0 are stored in a set. F ( t );

[0189] 4) Repeat steps 2) to 3) until all solutions are stored in the corresponding set, and obtain the non-dominated level result.

[0190] In one embodiment, the sorting module 30 is specifically used to perform the following steps:

[0191] Individuals are ranked according to their non-dominance level and crowding distance entropy. The ranking criteria are as follows: for two individuals with different non-dominance levels, the one with the lower non-dominance level is better than the one with the higher non-dominance level; for individuals with the same dominance level, the one with the larger crowding distance entropy is better than the one with the smaller crowding distance entropy.

[0192] Based on the sorting results, explorers, followers, and vigilants are determined in the population according to a set ratio.

[0193] In one embodiment, the system further includes an adjustment module, specifically used for:

[0194] The proportion of explorers in the population is dynamically adjusted. PD ,in:

[0195]

[0196]

[0197] In the formula, T The maximum number of iterations, t This represents the current iteration number.

[0198] In one embodiment, the location update module 40 is specifically used for:

[0199] The position of the follower is updated using a two-sample learning strategy, expressed by the formula:

[0200]

[0201] In the formula, This is the optimal position within the current sparrow population. This is the second-best position in the current sparrow population. This is the worst position in the current sparrow population. It is currently the second worst-performing species in the sparrow population. When the number of iterations is t When +1 is added, it represents the optimal position within the current sparrow population. When the number of iterations is t At that time, the first i The sparrow in the first j Location information in the dimension 1× d A matrix, where each element is randomly assigned the value 1 or -1, and , When the number of iterations is t At +1, the first i The sparrow in the first j Location information in the dimension Q For random numbers that follow a normal distribution, N Let be the population size for the sparrow search algorithm. i The number is the number of the sparrow. j As a dimension, r This is the step size control parameter, and .

[0202] It should be noted that other embodiments or implementation methods of the wireless sensor network coverage optimization system described in this invention can refer to the above-described method embodiments, and will not be repeated here.

[0203] The following analysis examines the time complexity of the WSN coverage optimization algorithm (IM-DTSSA) proposed in this embodiment, which is a multi-strategy improved sparrow search algorithm:

[0204] The algorithm is set to a maximum of T iterations, a population size of N, and n nodes in the network. First, the algorithm triangulates the static nodes according to the Delaunay triangulation strategy, calculates the Cl values ​​corresponding to the triangles, and sorts them, selecting the top N individuals as the explorer population. Second, a non-dominated sorting algorithm is used to classify individuals into different dominance levels, and the crowding distance entropy of the individuals is calculated and sorted. Finally, a two-sample learning strategy is used to update the positions of the followers. Therefore, the time complexity of the above three steps is analyzed as follows:

[0205] (1) The time complexity of triangulation of static nodes is O( nlogn The time complexity for calculating the Cl value corresponding to the triangle is O(nlogn), and the time complexity for initializing the population is O(nlogn). N ).

[0206] (2) The time complexity for calculating the non-dominated level is O( 2N 2 The time complexity for calculating the crowding distance entropy is O( ). NlogN ).

[0207] (3) The time complexity of updating the positions of various groups in the algorithm is O( N ).

[0208] Therefore, after introducing the above strategy to improve the algorithm, the time complexity of the IM-DTSSA algorithm is O(n). TN 2 Although the IM-DTSSA algorithm increases the computational cost, it outperforms other algorithms in terms of coverage and node movement cost.

[0209] To fully verify the effectiveness of the IM-DTSSA algorithm in optimizing WSN node coverage, simulation experiments were conducted using MATLAB 2021a software. Under the same network parameters, the results were compared with DPSO, ESSA, and NESSA. To make the experimental data more convincing, each algorithm's simulation was run independently 30 times, and the average value was used as the valid data for comparison. The simulation results are analyzed as follows:

[0210] (1) Analysis of Coverage Optimization Results

[0211] (1-1) Comparison of coverage optimization with a monitoring range of 100m×100m

[0212] Within a 100m x 100m monitoring area, 30 static nodes and 20 dynamic nodes are randomly deployed. The sensing radius of each node is set to 10m, and the maximum number of algorithm iterations is 500. Figure 8 (a) gives the initial node distribution when the number of static nodes is 30 and the number of dynamic nodes is 20. The initial nodes are all randomly distributed within the target area. Figure 8 In (a), the red dots represent dynamic nodes, the red circles represent the perception areas of dynamic nodes, the black "+" signs represent static nodes, and the black circles represent the perception areas of static nodes. Figure 8 (b) shows the final deployment of sensor nodes with 30 static nodes and 20 dynamic nodes. It can be clearly seen that most of the dynamic nodes are deployed in the target area that the static nodes did not perceive, and the dynamic nodes are eventually evenly distributed in the area not covered by the static nodes.

[0213] In addition, from Figure 8 As shown in (c), in terms of coverage improvement, the IM-DTSSA algorithm improved the network coverage from the initial 59.60% to 95.63%, achieving a 36.03% increase in network coverage. The NESSA algorithm's coverage gradually increased, eventually reaching 92.21%. The ESSA algorithm achieved a node coverage of 89.32% after 40 generations, but got stuck in a local optimum in the subsequent 200 iterations, only escaping the local optimum at generation 240, finally reaching a coverage of 90.59%. Although the DPSO algorithm's coverage continued to increase, it only reached a coverage of 88.89% in the end. In terms of convergence speed, the DPSO algorithm converged slowly, increasing by only 29.91% from an initial coverage of 58.98% after 500 iterations. Although the ESSA algorithm reached convergence quickly, it was clearly trapped in a local optimum. While the NESSA algorithm achieved a high coverage rate and avoided getting trapped in local optima, its convergence speed was slow, with the coverage rate increasing only gradually. The IM-DTSSA algorithm, on the other hand, quickly found a better solution from a lower initial coverage rate without getting trapped in local optima, demonstrating strong optimization capabilities and the ability to escape local optima.

[0214] from Figure 8 (d) As shown in Table 1, the IM-DTSSA algorithm has the smallest average node movement distance and relatively uniform movement distances among nodes, indicating that the algorithm can effectively control the movement distance cost of nodes. The length of a node's movement distance determines the remaining energy of that node after network deployment. Therefore, the difference in movement distances among different nodes also determines whether the energy is balanced among nodes after network deployment. Nodes with longer movement distances may experience premature energy depletion during network operation, thus affecting network monitoring performance and the selection of data transmission paths. Therefore, based on the experimental results of node movement distance and network coverage, the IM-DTSSA algorithm demonstrates good performance.

[0215] Table 1 Comparison of WSN coverage optimization in a 100m×100m area

[0216]

[0217] (1-2) Comparison of coverage optimization with a monitoring range of 50m×50m

[0218] Within a 50m×50m monitoring area, 15 static nodes and 20 dynamic nodes are randomly deployed. The sensing radius of each node is set to 5m, and the maximum number of iterations of the algorithm is 500. Figure 9(a) gives the initial node distribution when the static number of nodes is 15 and the dynamic number of nodes is 20. The initial nodes are all randomly distributed within the target area. Figure 9 (b) shows the distribution of network nodes after optimization by the IM-DTSSA algorithm. Figure 9 (c) Network coverage convergence curve.

[0219] Depend on Figure 9 (b) shows that after network coverage optimization, the IM-DTSSA algorithm achieves a more uniform node distribution, significantly improving node redundancy in the initial stage. Figure 9 (c) It can be seen that the DPSO algorithm can improve the coverage relatively quickly in the early stages, but after 60 iterations, it begins to get stuck in a local optimum. Although it escapes the local optimum twice in subsequent iterations, the increase in coverage is small, and it finally reaches a coverage of 83.81%. The ESSA algorithm gradually increases the coverage during the optimization process, but due to the slow increase in coverage in the early stages, it eventually reaches a coverage of 89.00%. The NESSA algorithm gradually increases the coverage during the iteration process and has a strong ability to escape local optima, but it still lags behind the IM-DTSSA algorithm in the initial stage. The network coverage of the IM-DTSSA algorithm after optimization is increased from the initial 73.32% to 91.89%, achieving a network coverage increase of 18.57%, which is significantly higher than the DPSO algorithm's 10.50%, the ESSA algorithm's 14.76%, and the NESSA algorithm's 16.88%.

[0220] from Figure 9 (d) As shown in Table 2, the node movement distance in the IM-DTSSA algorithm is less than that in the other three algorithms, and the fluctuation of the movement distance between nodes is smaller. The DPSO algorithm shows a larger difference in the movement distance between nodes, and the ESSA and NESSA algorithms have certain shortcomings in controlling the node movement distance.

[0221] Table 2 Comparison of WSN coverage optimization in a 50m×50m area

[0222]

[0223] (1-3) Comparison of coverage optimization with a monitoring range of 20m×20m

[0224] Within a 20m×20m monitoring area, 10 static nodes and 15 dynamic nodes are randomly deployed. The sensing radius of each node is set to 10m, and the maximum number of iterations of the algorithm is 500. Figure 10 (a) gives the initial node distribution when the number of static nodes is 10 and the number of dynamic nodes is 15. The initial nodes are all randomly distributed within the target area. Figure 10(b) shows the distribution of network nodes after optimization by the IM-DTSSA algorithm. Figure 10 (c) Network coverage convergence curve.

[0225] from Figure 10 (c) clearly shows that the DPSO algorithm achieved a coverage rate of 77% around generation 60, but then became stuck in local optima. The ESSA algorithm consistently improved network coverage, but at a relatively slow pace, reaching 83.90% coverage around generation 40, but ultimately only reaching 87.53%. The NESSA and IM-DTSSA algorithms were similar in coverage improvement speed, but the IM-DTSSA algorithm achieved a higher coverage rate faster in the early stages. This is because the IM-DTSSA algorithm uses the Delaunay triangulation strategy to quickly locate the initial population near the final target region. Ultimately, the optimized network coverage rate of the IM-DTSSA algorithm increased from the initial 47.62% to 91.84%, achieving a 44.22% increase in network coverage. This demonstrates that the IM-DTSSA algorithm has a strong ability to find global optimization and escape local optima.

[0226] In the 20m×20m monitoring area, the node movement distance was significantly reduced compared to the 100m×100m and 50×50m monitoring areas. Figure 10 (d) As can be seen from the experimental results in Table 3, IM-DTSSA still has a good effect on controlling the distance of node movement.

[0227] Table 3 Comparison of WSN coverage optimization in a 20m×20m area

[0228]

[0229] (2) Comparison of coverage with different numbers of dynamic nodes

[0230] This embodiment compares and analyzes the algorithm under different scenarios with a static number of nodes (30) and a varying number of dynamic nodes. Figure 11 It can be seen that when the number of dynamic nodes is 10, the coverage rate of the DPSO algorithm is 64.27%, the coverage rate of the ESSA algorithm is 73.44%, the coverage rate of the NESSA algorithm is 74.32%, and the coverage rate of the IM-DTSSA algorithm is 79.36%. When the number of dynamic nodes reaches 40, the coverage rate of the DPSO algorithm is 88.67%, the coverage rate of the ESSA algorithm is 92.35%, the coverage rate of the NESSA algorithm is 95.62%, and the coverage rate of the IM-DTSSA algorithm reaches 98.89%.

[0231] In summary, when the number of dynamic nodes is less than 45, the original IM-DTSSA algorithm has better coverage optimization compared to ESSA and MSSA algorithms. However, when the number of dynamic nodes is greater than 45, the gap between the algorithms gradually narrows. This is because when the number of network nodes is large, the coverage rate can reach more than 85%.

[0232] This embodiment compares the IM-DTSSA algorithm with three other algorithms under three different experimental environments, demonstrating the feasibility and superiority of the IM-DTSSA algorithm in WSN coverage optimization. Simultaneously, by setting different numbers of dynamic nodes, the impact of the number of dynamic nodes on network coverage is observed, and the experiments demonstrate the algorithm's stability. In summary, the experimental results of the four algorithms on the dynamic node movement distance and network coverage during coverage optimization show that the IM-DTSSA algorithm can reduce the average movement distance of dynamic nodes while maintaining network coverage.

[0233] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0234] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0235] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for optimizing coverage in a wireless sensor network, characterized in that, The method includes: Static and dynamic nodes are randomly deployed within the monitoring area, and each coverage gap in the wireless sensor network composed of the static nodes is identified as the initial sparrow population. Individuals in the sparrow population are divided into different non-dominant levels, and the crowding distance entropy corresponding to each individual is calculated. Individuals are ranked according to their non-dominance level and crowding distance entropy, and explorers, followers, and vigilants are determined based on the ranking results. Update the positions of the explorer, the follower, and the vigilant, wherein the position update of the follower employs a two-sample learning strategy; When the set iteration conditions are met, output the globally optimal deployment strategy, optimal coverage, and node movement distance; The step of dividing individuals in the sparrow population into different non-dominant classes and calculating the crowding distance entropy for each individual includes: The individuals in the sparrow population are sorted by non-dominance level to obtain the non-dominance level results; Based on the non-dominated rank results, the crowding distance entropy for all individuals in each non-dominated rank is calculated, wherein the formula for calculating the crowding distance entropy is: In the formula, h For the first i The number of individuals in the non-dominant hierarchy of a population. j Is with the first i Individuals in the same dominance hierarchy within a population To solve i +1 corresponds to the reciprocal of the coverage of the node deployment scheme. , To solve i -1 is the reciprocal of the coverage of the node deployment scheme. To solve i The average node movement distance for the node deployment scheme corresponding to +1. To solve i -1 corresponds to the average node movement distance of the node deployment scheme. To solve j The reciprocal of the coverage of the corresponding node deployment scheme. To solve j The average node movement distance of the corresponding node deployment scheme, To solve i The reciprocal of the coverage of the corresponding node deployment scheme. , To solve i The average node movement distance of the corresponding node deployment scheme, To solve i Crowded distance entropy, For coverage; The process of dividing individuals in the sparrow population into different non-dominant ranks includes: 1) Current round number t When =1, find all n ( p Store the solutions where ) = 0 in a set. F (1), n ( p ) is the dominant solution p Quantity; 2) For sets F ( t All individuals in ) j Examine individuals j The set under control S ( j ), S ( j All individuals in ) k Attributes n ( k ) = n ( k -1; 3) t = t +1, will n ( p The solutions where ) = 0 are stored in a set. F ( t ); 4) Repeat steps 2) to 3) until all solutions are stored in the corresponding set, and obtain the non-dominated level result.

2. The wireless sensor network coverage optimization method as described in claim 1, characterized in that, The process of randomly deploying static and dynamic nodes within the monitoring area and identifying coverage gaps in the wireless sensor network composed of the static nodes as the initial sparrow population includes: Randomly deploy static and dynamic nodes within the monitoring area and obtain the coordinates of all nodes; The Delaunay triangulation strategy is used to find coverage gaps in the wireless sensor network composed of the static nodes as the initial sparrow population.

3. The wireless sensor network coverage optimization method as described in claim 2, characterized in that, The process of using the Delaunay triangulation strategy to find coverage gaps in the wireless sensor network composed of static nodes as the initial sparrow population includes: The wireless sensor network composed of static nodes is divided using the Delaunay triangulation strategy to obtain a triangular network; Based on the relationship between the triangles in the triangular mesh and the uncovered areas, each coverage gap in the triangular mesh is determined as the initial sparrow population. The relationship between the triangles and the uncovered areas is as follows: when At that time, there is no uncovered area inside the triangle; when When, if and only if the triangle is an acute triangle, there exists an uncovered region inside the triangle that is entirely inside the triangle; when If the triangle is acute or right, there is an uncovered area inside the triangle. If the triangle is obtuse and the perpendicular distance from point C to AB is greater than R, there is an uncovered area inside the triangle. in, A , B , C These are three vertices of the same triangle in a triangular mesh formed after the statically constructed network is divided using a triangulation strategy. R This represents the sensing radius of the sensor node. d ( A , B Let be the distance between vertices A and B. d ( A Let C be the distance between vertices A and C. d ( B ,C) is the distance between vertices B and C.

4. The wireless sensor network coverage optimization method as described in claim 2 or 3, characterized in that, After employing the Delaunay triangulation strategy to find coverage gaps in the wireless sensor network composed of the static nodes, the method further includes: Calculate variables Cl To measure the size of the area of ​​the cover hole inside each triangle, variables Cl The calculation formula is: In the formula, For the area of ​​the corresponding triangle, The total area of ​​the three sectors inside the triangle. This represents the total area of ​​the repeated sectors inside the triangle; according to Cl The values ​​are used to sort the coverage vulnerabilities inside each triangle, and the top ones are selected. N The location of the coverage vulnerability is used as the initial sparrow population.

5. The wireless sensor network coverage optimization method as described in claim 1, characterized in that, The process of ranking individuals based on their non-dominance level and crowding distance entropy, and determining explorers, followers, and vigilant individuals based on the ranking results, includes: Individuals are ranked according to their non-dominance level and crowding distance entropy. The ranking criteria are as follows: for two individuals with different non-dominance levels, the one with the lower non-dominance level is better than the one with the higher non-dominance level; for individuals with the same dominance level, the one with the larger crowding distance entropy is better than the one with the smaller crowding distance entropy. Based on the sorting results, explorers, followers, and vigilants are determined in the population according to a set ratio.

6. The wireless sensor network coverage optimization method as described in claim 1 or 5, characterized in that, The method further includes: The proportion of explorers in the population is dynamically adjusted. PD ,in: In the formula, T The maximum number of iterations, t This represents the current iteration number.

7. The wireless sensor network coverage optimization method as described in claim 1, characterized in that, The position update formula for the follower is: In the formula, This is the optimal position within the current sparrow population. This is the second-best position in the current sparrow population. This is the worst position in the current sparrow population. It is currently the second worst-performing species in the sparrow population. When the number of iterations is t When +1 is added, it represents the optimal position within the current sparrow population. When the number of iterations is t At that time, the first i The sparrow in the first j Location information in the dimension 1× d A matrix, where each element is randomly assigned the value 1 or -1, and , When the number of iterations is t At +1, the first i The sparrow in the first j Location information in the dimension Q Let L be a random number that follows a normal distribution, and L be 1× d A matrix of all ones. N Let be the population size for the sparrow search algorithm. i The number is the number of the sparrow. j As a dimension, r This is the step size control parameter, and .

8. A wireless sensor network coverage optimization system, characterized in that, The system includes: The initial population determination module is used to randomly deploy static and dynamic nodes within the monitoring area and determine each coverage gap in the wireless sensor network composed of the static nodes as the initial sparrow population. The ranking module is used to divide individuals in the sparrow population into different non-dominant ranks and calculate the crowding distance entropy for each individual. The sorting module is used to sort individuals based on their non-dominance level and crowding distance entropy, and to determine explorers, followers, and vigilant individuals based on the sorting results. The location update module is used to update the positions of the explorer, the follower, and the vigilant, wherein the position update of the follower adopts a two-sample learning strategy. The output module is used to output the globally optimal deployment strategy, optimal coverage, and node movement distance when the set iteration conditions are met. The grade classification module 20 includes: The non-dominant ranking unit is used to rank the individuals of the sparrow population according to their non-dominant ranking to obtain the non-dominant ranking result. The crowding distance calculation unit is used to calculate the crowding distance entropy of all individuals in each non-dominated level based on the non-dominated level results. The formula for calculating the crowding distance entropy is as follows: In the formula, h For the first i The number of individuals in the non-dominant hierarchy of a population. j Is with the first i Individuals in the same dominance hierarchy within a population To solve i +1 corresponds to the reciprocal of the coverage of the node deployment scheme. , To solve i -1 is the reciprocal of the coverage of the node deployment scheme. To solve i The average node movement distance for the node deployment scheme corresponding to +1. To solve i -1 corresponds to the average node movement distance of the node deployment scheme. To solve j The reciprocal of the coverage of the corresponding node deployment scheme. To solve j The average node movement distance of the corresponding node deployment scheme, To solve i The reciprocal of the coverage of the corresponding node deployment scheme. , To solve i The average node movement distance of the corresponding node deployment scheme, To solve i Crowded distance entropy, For coverage; The non-dominant ranking unit is used to perform the following steps: 1) Current round number t When =1, find all n ( p Store the solutions where ) = 0 in a set. F (1), n ( p ) is the dominant solution p Quantity; 2) For sets F ( t All individuals in ) j Examine individuals j The set under control S ( j ), S ( j All individuals in ) k Attributes n ( k ) = n ( k -1; 3) t = t +1, will n ( p The solutions where ) = 0 are stored in a set. F ( t ); 4) Repeat steps 2) to 3) until all solutions are stored in the corresponding set, and obtain the non-dominated level result.

Images