A sensor node deployment method for forest monitoring

Abstract

The present invention belongs to the technical field of sensor optimization layout, and specifically relates to a sensor node layout method for forest monitoring, the steps of which include setting the sensing radius, communication radius and size of the monitoring area of ​​the sensor node, and generating random distribution information of the sensor node in the monitoring area; establishing a coverage rate mathematical model of the sensor node, combining it with the tangent reward item to form the objective function of the wireless sensor network layout optimization method; introducing the gray wolf optimization algorithm GWO and improving it; on the basis of random deployment, collecting the random distribution information of the sensor node, and dynamically adjusting the position distribution of the sensor node according to the deployment strategy generated by the improved GWO. The present invention reduces the redundancy of the coverage area between the sensor nodes while ensuring a high coverage rate, thereby reducing the energy consumption of sensor data processing and improving the overall performance of the wireless sensor network.

Description

Technical Field

[0001] The invention belongs to the technical field of sensor optimization arrangement, and in particular relates to a sensor node arrangement method for forest monitoring. Background Art

[0002] With the increasing awareness of environmental protection, forest monitoring has become an important research field. By monitoring key parameters in the forest environment (such as temperature, humidity, soil moisture content, etc.), it is possible to effectively assess the health of the forest and timely warn of potential ecological threats. In this field, Wireless Sensor Networks (WSN) are widely used in large-scale environmental monitoring due to their low power consumption, flexible deployment and self-organizing capabilities. Traditional forest management methods, such as fixed fire prevention measures or pest control, are difficult to cope with the dynamic changes in the forest environment. Modern wireless sensor network technology can capture the changing parameters in the forest environment in real time, which is crucial for dealing with risks such as extreme temperature fluctuations and potential fires. Through the collected data, the management system can analyze and make corresponding adjustments in a timely manner, thereby implementing more accurate protection measures, which not only improves resource utilization efficiency, but also effectively protects the forest ecology.

[0003] However, in practical applications, how to reasonably arrange (layout) sensor nodes while ensuring high coverage of the monitoring area and minimize the number of nodes to reduce costs and energy consumption has always been an important challenge. At present, the layout methods of sensor nodes mostly rely on random deployment or drone deployment. Although these methods are simple, they often lead to uneven node distribution, resulting in monitoring blind spots and coverage redundancy, which in turn affects the accuracy of data collection. In order to solve these problems, researchers have proposed a variety of sensor node layout optimization algorithms, including virtual force algorithms, computational geometry algorithms, and intelligent optimization algorithms. Although these algorithms improve network coverage and data transmission efficiency and reduce resource waste by optimizing node positions, they each have their limitations. The virtual force algorithm can adjust the node position by simulating the gravitational and repulsive forces between nodes to achieve better coverage, but it is highly dependent on initial conditions, and as the number of nodes increases, the computational complexity increases significantly, which limits the application effect of this method in large-scale monitoring scenarios. Computational geometry algorithms (such as Voronoi diagrams and Delaunay triangulations) optimize by analyzing the geometric relationships of sensor nodes, but they have poor adaptability when dealing with dynamically changing monitoring needs, and it is especially difficult to achieve real-time adjustments to maintain a high coverage rate. Intelligent optimization algorithms (such as genetic algorithms and particle swarm optimization algorithms) have attracted widespread attention due to their powerful global search capabilities, especially in solving complex optimization problems. However, these algorithms usually require a lot of parameter tuning and are prone to falling into local optimal solutions during the search process, resulting in reduced optimization efficiency. In particular, in the node deployment problem of large-scale wireless sensor networks, how to reduce algorithm complexity and avoid premature convergence while ensuring coverage has always been a key problem.

[0004] Therefore, the existing wireless sensor network deployment methods have many shortcomings in complex forest monitoring scenarios, and a new sensor node deployment method is urgently needed. Summary of the invention

[0005] In view of the above deficiencies in the prior art, the purpose of the present invention is to provide a sensor node deployment method for forest monitoring, which simplifies the node deployment process and reduces energy consumption while ensuring high coverage, thereby improving the overall performance of the wireless sensor network.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is a sensor node deployment method for forest monitoring, comprising the following steps:

[0007] S1, setting the sensing radius, communication radius and size of the monitoring area of ​​the sensor node, and generating random distribution information of the sensor node in the monitoring area, the sensor node is a wireless sensor node;

[0008] S2. Based on the Boolean perception model, calculate the coverage of each sensor node to the monitoring area, and then establish a mathematical model of the coverage rate of the sensor nodes;

[0009] S3, introducing the coverage rate mathematical model and the tangency reward item into the arrangement process of the sensor nodes to form the objective function of the optimal arrangement of the wireless sensor network;

[0010] S4, introduce the gray wolf optimization algorithm GWO and improve it, take the objective function in S3 as the fitness function, and gradually optimize the position of sensor nodes by simulating the hunting behavior of gray wolves; the improvements include replacing the linear convergence factor of GWO with a nonlinear convergence factor, introducing a genetic mutation strategy and a mutation mechanism based on Cauchy distribution in GWO;

[0011] S5. On the basis of random deployment, by collecting the random distribution information of sensor nodes, and then according to the deployment strategy generated by the improved GWO, the location distribution of sensor nodes is dynamically adjusted to improve the coverage of the wireless sensor network and optimize the energy consumption of sensor nodes; after multiple iterations, the monitoring area is covered to the maximum extent. When the iteration termination condition is met, the final position of each sensor node is output to obtain the final optimized sensor node layout plan.

[0012] A wireless sensor network (WSN) is a network composed of all sensor nodes. Node tangency rewards usually refer to the incentives provided to participants who run nodes in a blockchain network. A Boolean perception model refers to a model that analyzes the sensitivity of input variables in logical operations and Boolean algebra.

[0013] The traditional Grey Wolf Optimizer (GWO) is a meta-heuristic algorithm inspired by the social behavior and hunting strategy of grey wolves. The algorithm simulates the hunting process of grey wolves, and grey wolves adjust the positions of individuals by encircling prey to achieve the global optimization goal. The grey wolf group consists of four different levels: α, β, γ and ω, among which α wolf is responsible for leading the wolf group, β and γ wolves assist α wolf in decision-making, and ω wolf is an ordinary individual. Through the collaboration of these four roles, the algorithm continuously updates the position of individuals and gradually approaches the optimal solution. A significant advantage of the Grey Wolf Optimizer is that it has relatively few parameter settings, making it easier to implement and debug than other complex algorithms. Its core parameters only include convergence factors and constants related to hunting behavior, which greatly simplifies the implementation of the algorithm in different application scenarios and reduces the complexity of parameter tuning. Therefore, the Grey Wolf Optimizer shows high robustness and ease of use in practical applications. However, the algorithm is prone to fall into the local optimal solution during the local search process, resulting in limited global search capabilities, especially when dealing with complex optimization problems. The performance is not ideal. Therefore, the present invention improves it.

[0014] As a preferred solution of the present invention, in S1, it is assumed that the sensor nodes in the wireless sensor network are homogeneous, and the sensing radius of the sensor nodes is R p , the communication radius is R c , the sensor node set V is represented as V={v 1 , v 2 , v 3 , …, v n}, n is the number of sensor nodes, any sensor node v in V i The two-dimensional coordinates of i ,y i ), i=1, 2, 3, …, n; monitoring nodes are set in the monitoring area, and the monitoring node set Q is expressed as Q={q 1 ,q 2 ,q 3 ,…,q m}, m is the number of monitoring nodes, any monitoring node q in Q j The coordinates of (x j ,y j ), j = 1, 2, 3, …, m; the monitoring area is defined as a rectangular area of ​​size L×W.

[0015] As a preferred solution of the present invention, in S1, R c =2R p , ensuring that sensor nodes can communicate effectively and maintain a stable connection of the entire wireless sensor network.

[0016] As a preferred solution of the present invention, in said S2, the method for establishing the coverage mathematical model of the sensor nodes is:

[0017] S21. Calculate the distance matrix D between the sensor node and the monitoring node. The formula is as follows:

[0018] ;

[0019] Where D ij Indicates v i With q j The distance between each D ij The overall distance matrix D can be obtained;

[0020] S22, calculate the distance matrix A between each sensor node, the formula is as follows:

[0021] ;

[0022] In the formula, A ik Indicates v i With sensor node vk The distance between them, k = 1, 2, 3, ..., n and k ≠ i; calculate each A ik The overall distance matrix A can be obtained;

[0023] S23, based on the Boolean perception model, a single sensor node is j The monitoring probability formula is:

[0024] ;

[0025] In the formula, p er (v i ,q j ) indicates v i Can monitor q j probability;

[0026] S24, all sensor nodes are q j The joint perception probability is expressed as:

[0027] ;

[0028] Where P joint (V, q j ) represents the set of all sensor nodes V for monitoring node q j The joint coverage probability of

[0029] S25. Divide the monitoring area into L×W grids of equal area. The center of each grid represents a monitoring node. Then the coverage rate C r It is expressed as:

[0030] ;

[0031] In the formula, x and y represent the coordinates of the center of the grid, which are continuously increased from the origin of the coordinate axis to represent the center point of L×W grids; That means monitoring nodes; through C r Indicates the proportion of the monitoring area covered by the wireless sensor network.

[0032] As a preferred solution of the present invention, in said S3, the tangent reward item It is expressed as:

[0033] ;

[0034] In the formula, is the tolerance range of the distance; E is the indicator function, when the conditions in the brackets are met, E is equal to 1, otherwise E is equal to 0; (x i ,y i ) is v k The two-dimensional coordinates of

[0035] Objective function for optimal deployment of high coverage wireless sensor networks It is expressed as:

[0036] ;

[0037] In the formula, is a weighting factor that balances the contribution of tangency rewards.

[0038] As a preferred solution of the present invention, in S4, the nonlinear convergence factor factor is expressed as:

[0039] ;

[0040] Where MaxIter is the maximum number of iterations; iter is the current number of iterations;

[0041] The genetic mutation strategy includes the genetic strategy and the mutation strategy. The genetic strategy is expressed as:

[0042] ;

[0043] In the formula, mod means taking the remainder; is the congruence symbol; , , They represent the three wolf individuals with the highest fitness in the population; p represents the pth solution in the population (the wolf individual is the potential solution); X p represents the pth wolf individual in the next generation;

[0044] expression In mathematics, it means that the remainder of p divided by 3 is 1, and so on;

[0045] The mutation strategy is to randomly select dimension d r :

[0046] ;

[0047] Where d is the size of the population dimension; randint is a function call used to generate a random integer; then:

[0048] ;

[0049] In the formula, Indicated in dimension d r The pth wolf individual after the mutation; r is a random number uniformly distributed in [0, 1], used to determine whether a mutation occurs; Q Tis the mutation rate, which indicates the probability of mutation; N (0, 1) is the standard normal distribution, which is used to introduce random disturbances; ms is the mutation intensity, which controls the intensity of mutation applied to individual wolves;

[0050] The variation mechanism of the Cauchy distribution is, and the Cauchy distribution function is expressed as:

[0051] ;

[0052] Where C represents the Cauchy distribution variable used for mutation; is the location parameter indicating the peak of the distribution; is a scale parameter that controls the discreteness of the distribution; rand is a uniformly distributed random number between 0 and 1;

[0053] After obtaining the current global optimal solution, the Cauchy mutation is used to update the current global optimal solution, which is expressed as:

[0054] ;

[0055] In the formula, Represents the new coordinates of the global optimal solution in the qth dimension; Represents the coordinates of the global optimal solution in the qth dimension; represents a random number drawn from a standard Cauchy distribution.

[0056] As a preferred embodiment of the present invention, in the genetic strategy, the genetic strategy is applied to the three wolf individuals with the highest fitness, and their excellent characteristics are passed on to the five wolf individuals of the next generation in turn, so as to ensure the retention of the best individuals and increase the diversity of the population; at this time, p=1, 2, 3, 4, 5, indicating that the genetic strategy acts on the five wolf individuals in the offspring; in the mutation strategy, the wolf individuals ranked from 4 to N in fitness will undergo mutation operation, and then the fitness will be re-sorted, and the individuals with the last two fitness will be eliminated to generate offspring individuals, and N is the population size.

[0057] As a preferred solution of the present invention, in S5, the specific method for obtaining the final optimized sensor node arrangement solution is:

[0058] S51. In the improved GWO, the set of sensor nodes is expressed as:

[0059] ;

[0060] In the formula, X represents the overall matrix; represents the position of the Nth wolf individual in the search space under dimension d; Represents the dth dimension of the Nth wolf individual; each wolf individual contains the number and location information of all sensor nodes;

[0061] S52, the initial position of the Ith wolf is X I , I = 1, 2, 3, ..., N, N is the population size; each wolf individual represents a potential deployment plan, and the position of each wolf individual in the solution space is defined by a set of candidate variable values, X I Given by:

[0062] ;

[0063] In the formula, , They represent the lower and upper bounds of the d-dimensional variable in the I-th candidate variable respectively; R is a random number uniformly distributed in the range of [0, 1];

[0064] S53, according to the value of fitness, the wolf individuals are divided into four levels, from the highest to the lowest, namely α, β, γ and ω, where α, β and γ represent the top three wolf individuals in fitness, that is, the three alpha wolves, corresponding to the current best solution, the second best solution, and the general solution respectively; the remaining wolf individual ω is an ordinary wolf, representing a candidate solution; the fitness function is the objective function in S3;

[0065] S54. Execute the encirclement strategy in GWO and calculate the distance between the common wolf and the leader wolf:

[0066] ;

[0067] ;

[0068] In the formula, is the vector representing the distance between the common wolf and the alpha wolf Y; Is the impact The coefficient of represents the position vector of the leader wolf Y at time t; represents the position vector of an ordinary wolf at time t; is a random vector uniformly distributed in the range [0, 1];

[0069] S55. Update the position of the common wolf according to the distance between the common wolf and the leader wolf, expressed as:

[0070] ;

[0071] ;

[0072] In the formula, factor is the nonlinear convergence factor in S4; Indicates the current position vector of the wolf leader Y; It is the coefficient that affects the update position. The adjustment amplitude is controlled by factor during calculation. Yes the updated position of the impact; is a random vector uniformly distributed in the range [0, 1];

[0073] S56. The formula for updating the population position is:

[0074] ;

[0075] In the formula, is the position vector updated in the next time step, representing the average position of the current leader;

[0076] S57, perform multi-directional encirclement. The three alpha, beta, and gamma alpha wolves surround the prey from different directions and explore the entire solution space through multi-directional search.

[0077] S58. In the process of iterative search, the encirclement is dynamically adjusted by controlling the size of factor to find the optimal solution. The optimal solution is the sensor node layout plan that covers the monitoring area to the maximum extent.

[0078] As a preferred solution of the present invention, in the above S58, during the iteration process, the factor is gradually reduced, thereby reducing the search range and gradually tightening the encirclement.

[0079] As a preferred solution of the present invention, in S5, the iteration termination condition is that the number of iterations reaches a maximum number of 300 iterations, or the coverage rate of the wireless sensor network for the monitoring area is higher than a set coverage rate threshold.

[0080] The algorithm involved in the present invention can be executed by an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the above algorithm calculation is implemented by executing the software by the processor.

[0081] The beneficial effects of the present invention are:

[0082] The present invention reduces the redundancy of the coverage area between sensor nodes while ensuring high coverage, thereby reducing the energy consumption of sensor data processing and improving the overall performance of the wireless sensor network. The method not only needs to have a stronger global search capability, but also can adapt to dynamic environments, ensuring full coverage of the target area to be monitored with a limited number of sensor nodes, thereby ensuring the accuracy and real-time performance of forest environment monitoring.

[0083] The present invention combines the introduction of the gray wolf optimization algorithm and its multiple improvement strategies, including nonlinear convergence factor, genetic mutation strategy and Cauchy mutation, to effectively optimize the sensor node layout of the forest monitoring wireless sensor network, which not only makes up for the shortcomings of the traditional gray wolf optimization algorithm, but also performs well in convergence, global search capability and robustness. At the same time, based on the coverage mathematical model of the forest monitoring wireless sensor network and the node tangency reward item, a new objective function is constructed. This objective function not only takes into account the coverage, but also attempts to optimize the spatial layout of sensor nodes and reduce redundant coverage to make the sensor node configuration more reasonable. This method effectively combines practicality and theoretical optimization, and realizes the optimization of the overall performance of the network by adjusting the spatial distribution of sensors. The optimized wireless sensor network shows significant improvement in both coverage and energy consumption, and significantly reduces node redundancy.

[0084] The present invention uses the mutant gray wolf optimization algorithm to automatically adjust the position of sensor nodes, and combines the nonlinear convergence factor to control the balance of global and local searches. The genetic mutation strategy increases the diversity of the population, while the Cauchy mutation further improves the global search capability and avoids the risk of falling into the local optimal solution. The introduction of the tangent reward term further reduces node redundancy, optimizes the spatial distribution of sensors, and significantly improves the overall performance of the network. This combined strategy ensures that the system achieves higher coverage and energy consumption control in complex monitoring scenarios.

[0085] The present invention is not only suitable for forest monitoring, but also has a wide range of application prospects and can be extended to other large-scale environmental monitoring, agricultural sensor networks and other fields. By improving the optimization algorithm, the coverage efficiency and energy consumption control of the network are significantly improved, providing an effective solution for complex monitoring scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] Figure 1 It is a schematic diagram of the process of the present invention;

[0087] Figure 2 It is a visual comparison diagram of coverage in the verification process of the present invention; Figure 2 (a), (b), (c), and (d) are schematic diagrams of the coverage of each algorithm under constraint 1, constraint 2, constraint 3, and constraint 4, respectively;

[0088] Figure 3 It is a node deployment comparison diagram under constraint 1 during the verification process of the present invention; Figure 3 (a) is a schematic diagram of the sensor node locations with random deployment; Figure 3 (b), (c), and (d) are schematic diagrams of sensor node locations after optimization by GVGWO, GWO, and IGWO, respectively;

[0089] Figure 4 It is a node deployment comparison diagram under constraint 3 during the verification process of the present invention; Figure 4 (a) is a schematic diagram of the sensor node locations with random deployment; Figure 4 (b), (c), and (d) are schematic diagrams of sensor node locations after optimization by GVGWO, GWO, and IGWO, respectively;

[0090] Figure 5 is a flow chart of sensor node deployment in the verification process of the present invention;

[0091] Figure 6 It is the optimization flow chart of GVGWO of the present invention. DETAILED DESCRIPTION

[0092] The embodiments of the present invention are further described below in conjunction with the accompanying drawings:

[0093] like Figure 1 As shown, a sensor node deployment method for forest monitoring includes the following steps:

[0094] S1, setting the sensing radius, communication radius and size of the monitoring area of ​​the sensor node, and generating random distribution information of the sensor node in the monitoring area, the sensor node is a wireless sensor node;

[0095] S2. Based on the Boolean perception model, calculate the coverage of each sensor node to the monitoring area, and then establish a mathematical model of the coverage rate of the sensor nodes;

[0096] S3, introducing the coverage rate mathematical model and the tangency reward item into the arrangement process of the sensor nodes to form the objective function of the optimal arrangement of the wireless sensor network;

[0097] S4, introduce the gray wolf optimization algorithm GWO and improve it, take the objective function in S3 as the fitness function, and gradually optimize the position of sensor nodes by simulating the hunting behavior of gray wolves; the improvements include replacing the linear convergence factor of GWO with a nonlinear convergence factor, introducing a genetic mutation strategy and a mutation mechanism based on Cauchy distribution in GWO;

[0098] S5. On the basis of random deployment, by collecting the random distribution information of sensor nodes, and then according to the deployment strategy generated by the improved GWO, the location distribution of sensor nodes is dynamically adjusted to improve the coverage of the wireless sensor network and optimize the energy consumption of sensor nodes; after multiple iterations, the monitoring area is covered to the maximum extent. When the iteration termination condition is met, the final position of each sensor node is output to obtain the final optimized sensor node layout plan.

[0099] In S1, the sensor nodes in the wireless sensor network are assumed to be homogeneous, and the sensing radius of the sensor nodes is R p , the communication radius is R c , the sensor node set V is represented as V={v 1 , v 2 , v 3 , …, v n}, n is the number of sensor nodes, any sensor node v in V i The two-dimensional coordinates of i ,y i ), i=1, 2, 3, …, n; monitoring nodes are set in the monitoring area, and the monitoring node set Q is expressed as Q={q 1 ,q 2 ,q 3 ,…,q m}, m is the number of monitoring nodes, any monitoring node q in Q j The coordinates of (x j ,y j ), j = 1, 2, 3, ..., m; the monitoring area is defined as a rectangular area of ​​size L × W. Set R c =2R p , ensuring that sensor nodes can communicate effectively and maintain a stable connection of the entire wireless sensor network.

[0100] In S2, the method for establishing the coverage mathematical model of sensor nodes is:

[0101] S21. Calculate the distance matrix D between the sensor node and the monitoring node. The formula is as follows:

[0102] ;

[0103] Where D ij Indicates v i With q j The distance between each D ij The overall distance matrix D can be obtained; by analyzing the distance matrix D, the position of the sensor nodes can be optimized to reduce coverage holes and improve coverage efficiency;

[0104] S22, calculate the distance matrix A between each sensor node, the formula is as follows:

[0105] ;

[0106] In the formula, A ik Indicates v i With sensor node v k The distance between them, k = 1, 2, 3, ..., n and k ≠ i; calculate each A ikThe overall distance matrix A can be obtained; the distance matrix A is a key tool for the design and optimization of wireless sensor networks, which helps to enhance the stability, connectivity and coverage performance of the network;

[0107] S23, based on the Boolean perception model, a single sensor node is j The monitoring probability formula is:

[0108] ;

[0109] In the formula, p er (v i ,q j ) indicates v i Can monitor q j probability;

[0110] If D ij Less than or equal to the sensing radius R p , the monitoring node is considered to be covered; otherwise, the opposite is true; this binary coverage model is often used in wireless sensor networks to simplify the representation of sensing capabilities and facilitate coverage optimization analysis;

[0111] S24, all sensor nodes are q j The joint perception probability is expressed as:

[0112] ;

[0113] Where P joint (V, q j ) represents the set of all sensor nodes V for monitoring node q j The purpose is to calculate the total probability of the monitoring node being covered by multiple sensor nodes.

[0114] S25. Divide the monitoring area into L×W grids of equal area. The center of each grid represents a monitoring node. Then the coverage rate C r It is expressed as:

[0115] ;

[0116] In the formula, x and y represent the coordinates of the center of the grid, which are continuously increased from the origin of the coordinate axis to represent the center point of L×W grids; That means monitoring nodes; through C r Indicates the proportion of the monitoring area covered by the wireless sensor network.

[0117] The present invention deeply analyzes the geometric relationship between the communication range and the perception range of the sensor nodes. Its goal is to achieve the tangential state by precisely adjusting the distance between sensor nodes, thereby maximizing the coverage area and reducing overlap, thereby improving the overall energy efficiency. To achieve this goal, a novel tangency reward term is introduced to incentivize sensor nodes to maintain the optimal distance, thereby optimizing the network layout and ensuring the best utilization of energy resources.

[0118] In S3, tangent reward items It is expressed as:

[0119] ;

[0120] In the formula, is the tolerance range of the distance; E is the indicator function, when the conditions in the brackets are met, E is equal to 1, otherwise E is equal to 0; (x i ,y i ) is v k The two-dimensional coordinates of Quantifying spatial optimization by rewarding configurations where sensor nodes are tangentially aligned, thus enhancing overall coverage efficiency;

[0121] Objective function for optimal deployment of high coverage wireless sensor networks It is expressed as:

[0122] ;

[0123] In the formula, is a weighting factor that balances the contribution of tangency rewards.

[0124] Nonlinear convergence factor: The position update formula in GWO usually relies on linear decay factors, which gradually decrease from a higher value to a lower value as the number of iterations increases. This linear change may cause the algorithm to perform a rough global search in the early stage and a fine local search in the later stage. However, this simple linear decay sometimes leads to too fast convergence speed and may fall into the local optimum too early. The present invention reasonably designs the convergence factor according to the number of iterations. The results show that the designed nonlinear factor can help the improved GWO better balance the capabilities of global search and local search.

[0125] In S4, the nonlinear convergence factor is expressed as:

[0126] ;

[0127] Where MaxIter is the maximum number of iterations; iter is the current number of iterations;

[0128] In the early stages of iteration, , when iter keeps increasing, it will appear ; By introducing nonlinear terms, a balance is achieved between global exploration and local development, enhancing the adaptability and flexibility of the algorithm. It can provide a strong global search capability in the early stage of iteration, gradually reduce the convergence factor to improve the local search accuracy, thereby avoiding falling into the local optimal trap and improving the overall search efficiency. Ultimately, this design not only speeds up the convergence speed, but also increases the probability of finding the global optimal solution.

[0129] GWO optimizes node positions by simulating the hunting behavior of wolves, but it is easily limited by the initial position, resulting in a local optimal solution. To overcome this limitation, the present invention introduces a genetic mutation mechanism. First, a genetic strategy is applied to the three wolves with the highest fitness, and their characteristics are passed on to the other five wolves in turn to ensure that the optimal individuals are retained and the diversity of the population is increased. Secondly, by randomly mutating the positions of the remaining individuals during the optimization process, the diversity of node distribution is further increased, thereby improving the coverage. The mutation operation can adjust the position of the node so that it better covers key areas or uncovered blind areas, thereby improving the coverage of the overall network.

[0130] The genetic mutation strategy includes the genetic strategy and the mutation strategy. The genetic strategy is expressed as:

[0131] ;

[0132] In the formula, mod means taking the remainder; is the congruence symbol; , , They represent the three wolf individuals with the highest fitness in the population; p represents the pth solution in the population; X p represents the pth wolf individual in the next generation;

[0133] The mutation strategy is to randomly select dimension d r :

[0134] ;

[0135] In the formula, d is the size of the population dimension; then:

[0136] ;

[0137] In the formula, Indicated in dimension d r The pth wolf individual after the mutation; r is a random number uniformly distributed in [0, 1], used to determine whether a mutation occurs; Q Tis the mutation rate, which indicates the probability of mutation; N (0, 1) is the standard normal distribution, which is used to introduce random disturbances; ms is the mutation intensity, which controls the intensity of mutation applied to individual wolves;

[0138] In order to solve the problem that traditional GWO is prone to fall into local optimum, the present invention introduces a mutation mechanism based on Cauchy distribution to enhance the global search capability of the algorithm by increasing population diversity. The characteristic of Cauchy distribution is that there is a clear peak at the center and heavy tails at both ends. Using this characteristic, Cauchy mutation will introduce significant disturbances in the neighborhood of the solution, greatly expanding the search range and increasing the possibility of escaping the local optimum.

[0139] The variation mechanism of the Cauchy distribution is, and the Cauchy distribution function is expressed as:

[0140] ;

[0141] Where C represents the Cauchy distribution variable used for mutation; is the location parameter indicating the peak of the distribution; is a scale parameter that controls the dispersion of the distribution; rand is a uniformly distributed random number between 0 and 1; π is used in the tangent function to introduce variability in the mutation process;

[0142] After obtaining the current global optimal solution, the Cauchy mutation is used to update the current global optimal solution, which is expressed as:

[0143] ;

[0144] In the formula, Represents the new coordinates of the global optimal solution in the qth dimension; Represents the coordinates of the global optimal solution in the qth dimension; represents a random number drawn from a standard Cauchy distribution. This approach introduces randomness into the optimization process, enhancing the algorithm's ability to explore the solution space extensively and reducing the risk of premature convergence to a local optimum.

[0145] In the genetic strategy, the genetic strategy is applied to the three wolf individuals with the highest fitness, and their excellent characteristics are passed on to the five wolf individuals of the next generation in turn to ensure that the best individuals are retained and the diversity of the population is increased; at this time, p=1, 2, 3, 4, 5, indicating that the genetic strategy acts on the five wolf individuals in the offspring; in the mutation strategy, the wolf individuals ranked 4 to N in fitness will undergo mutation operations, and then the fitness will be re-ranked, and the individuals with the last two fitnesses will be eliminated to generate offspring individuals, and N is the population size.

[0146] In the study of coverage optimization of forest monitoring wireless sensor networks based on the improved gray wolf optimization algorithm (GVGWO), the key task is to determine how to deploy these sensor nodes under the premise of a known number of sensors in order to maximize the coverage of the network. This involves finding the optimal location of sensors in the target area in order to achieve complete coverage of the area as much as possible.

[0147] In S5, the specific method for obtaining the final optimized sensor node deployment solution is:

[0148] S51. In the improved GWO, the set of sensor nodes is expressed as:

[0149] ;

[0150] In the formula, X represents the overall matrix; represents the position of the Nth wolf individual in the search space under dimension d; Represents the dth dimension of the Nth wolf individual; each wolf individual contains the number and location information of all sensor nodes;

[0151] That is, the position of the first wolf individual in the search space under dimension d; This is the first dimension of the first wolf individual; the rest of the symbols can be deduced in the same way;

[0152] S52, the initial position of the Ith wolf is X I , I = 1, 2, 3, ..., N, N is the population size; each wolf individual represents a potential deployment plan, and the position of each wolf individual in the solution space is defined by a set of candidate variable values, X I Given by:

[0153] ;

[0154] In the formula, , Respectively represent the lower and upper bounds of the d-dimensional variable in the I-th candidate variable; R is a random number uniformly distributed in the range [0, 1]; This formula ensures that the initial position of each wolf individual is randomly distributed in the predetermined search space, and introduces a certain fluctuation through the sine function, which enhances the diversity of the population and thus improves the global search capability;

[0155] S53, according to the value of fitness, the wolf individuals are divided into four levels, from the highest to the lowest, namely α, β, γ and ω, where α, β and γ represent the top three wolf individuals in terms of fitness, i.e. the three alpha wolves, corresponding to the current best solution, the second best solution and the general solution respectively; the remaining wolf individual ω is an ordinary wolf, representing a candidate solution; the fitness function is the objective function in S3; the level of each wolf individual is not fixed, but is redistributed according to the fitness value;

[0156] GWO's encirclement strategy is designed based on the hunting behavior of gray wolves. Wolves approach and capture prey through cooperation and hierarchical behaviors (α, β, γ and ω). GWO effectively explores and utilizes the solution space and gradually approaches the optimal solution by simulating this complex encirclement strategy. The core of the encirclement strategy is to dynamically adjust the position of individual wolves so that the entire wolf pack gradually encircles the prey and continuously tightens the encirclement.

[0157] S54. Execute the encirclement strategy in GWO and calculate the distance between the common wolf and the leader wolf:

[0158] ;

[0159] ;

[0160] In the formula, is the vector representing the distance between the common wolf and the alpha wolf Y; Is the impact The coefficient of , which is usually used to adjust the step size or weight; represents the position vector of the leader wolf Y at time t; represents the position vector of an ordinary wolf at time t; is a random vector uniformly distributed in the range [0, 1]; randomness and diversity are introduced to avoid local optimality; the expression Will Zooming to a larger range further affects the extent and direction of position updates;

[0161] S55. Update the position of the common wolf according to the distance between the common wolf and the leader wolf, expressed as:

[0162] ;

[0163] ;

[0164] In the formula, factor is the nonlinear convergence factor in S4; Indicates the current position vector of the wolf leader Y; It is the coefficient that affects the update position. The adjustment amplitude is controlled by factor during calculation. Yes the updated position of the impact; is a random vector uniformly distributed in the range [0, 1], which introduces variability for updates;

[0165] S56. The formula for updating the population position is:

[0166] ;

[0167] In the formula, is the position vector updated in the next time step, representing the average position of the current alpha wolf; this averaging method facilitates the optimization process by guiding the search to the center of the current best solution;

[0168] S57, perform multi-directional encirclement, the three alpha, beta, and gamma alpha wolves encircle the prey from different directions, and explore the entire solution space through multi-directional search; such multi-directional search can effectively explore the entire solution space and reduce the possibility of falling into a local optimum;

[0169] S58. In the process of iterative search, by controlling the size of factor, dynamically adjusting the encirclement, and finding the optimal solution, the optimal solution is the sensor node layout scheme that covers the monitoring area to the maximum extent. In the iterative process, gradually reduce the factor, thereby narrowing the search range and gradually tightening the encirclement.

[0170] In the early stage of iteration, a larger factor is used to maintain a wide search range (global search); in the later stage of iteration, a smaller factor is used to narrow the search range (local search), gradually tightening the encirclement. By simulating the hunting behavior of gray wolves, GWO can effectively adjust the layout of sensor nodes to cope with complex geographical conditions and environmental changes, and improve the performance and stability of the entire wireless sensor network.

[0171] In S5, the iteration termination condition is that the number of iterations reaches the maximum number of iterations of 300, or the coverage rate of the wireless sensor network for the monitoring area is higher than the set coverage rate threshold. The coverage rate threshold is set as needed, for example, it can be set to be greater than or equal to 95%.

[0172] The verification process is:

[0173] Through simulation experiments, the improved GWO (GVGWO) of the present invention is compared with the traditional GWO and the GWO variant (IGWO); the simulation experiment compares the optimized coverage and sensor node distribution, which better reflects the practicality of GVGWO. The experiment is carried out on a computer equipped with an i5 dual-core CPU, with a main frequency of 2.4GHz, 8GB of memory, and Windows 11 system. The experimental simulation software uses matlab 2018a.

[0174] In order to compare the performance of several methods in forest monitoring wireless sensor network (FMWSN) coverage optimization, unified public parameters are set to ensure the fairness and objectivity of the experiment. The deployment process is as follows: Figure 5 As shown in the figure, it includes network parameter initialization, random deployment of location information of each sensor (i.e. sensor node), establishing communication, and using algorithm for deployment optimization.

[0175] The initial population size N of all algorithms is set to 40, the population dimension is set to n (n is the number of sensor nodes, that is, d=n at this time), the number of iterations MaxIter is set to 300, the number of experiments is 20 times, and on this basis, it is compared with other swarm intelligence algorithms (IGWO, GWO). The simulation limits of different monitoring areas, number of node sensor points, sensor node perception radius and communication radius are shown in Table 1, and the unit of coverage is m. Based on the limits in Table 1, the network coverage of each method after optimization is shown in Table 2. When applied to FMWSN, in the actual forest sensor network coverage, due to environmental occlusion factors, the communication distance between sensor nodes is limited. Therefore, in the simulation, limit 1 (that is, the abbreviation of limit condition 1, the same below), limit 2, and limit 3 are set within the communication distance range of 8±2 (m), and limit 4 is set within the communication distance range of 4±1 (m). As long as the communication distance in the simulation result is within the limited range, the communication is considered to be effective. In order to study the response of sensor nodes under fault conditions, a uniform failure rate of 5% is set when deploying (that is, arranging) sensor nodes.

[0176] Table 1 Network parameter settings

[0177]

[0178] Table 2 Coverage of each deployment algorithm

[0179]

[0180] Figure 2The simulation results of GVGWO, IGWO and GWO algorithms under four different constraints are presented, which is a visual comparison of coverage. The results show that the GVGWO algorithm achieves the highest coverage under all test conditions and becomes the preferred solution for optimizing sensor network deployment. Specifically, in constraint 1, the coverage of GVGWO reaches 99.70%, which is 1.07%, 6.17% and 17.4% higher than IGWO, GWO and random deployment, respectively. After 80 iterations, the coverage of GVGWO increases to 97.20%, which is 2.00% and 14.72% higher than IGWO and GWO, respectively. Under constraint 2, although the number of nodes is reduced by 10, GVGWO still achieves a coverage of 96.73%, which is 1.26% and 20.38% higher than IGWO and GWO, respectively. In addition, after 140 iterations, GVGWO successfully avoids the local optimum and shows higher convergence accuracy, which verifies the effectiveness of the Cauchy mutation strategy.

[0181] Under restriction 4, the number of sensor nodes is reduced by 15 compared with restriction 1. Experimental data show that the coverage of GVGWO reaches 94.36%, which is 2.97% and 22.44% higher than IGWO and GWO respectively. In addition, after 70 iterations, the coverage of GVGWO is as high as 90.44%, while the coverage of the other two algorithms is much lower than that of GVGWO, indicating that GVGWO can still maintain high coverage efficiency and stability while reducing sensor nodes and coverage area.

[0182] Figure 3 and Figure 4 The simulation results of GVGWO, IGWO and GWO algorithms for optimizing sensor node deployment under restrictions 1 and 3 are shown (“*” represents the location of the sensor node, and the circle is the sensing area). The results show that under the same conditions, increasing the number of sensor nodes can improve coverage and reduce coverage redundancy, but blind areas still exist. Figure 3 The random deployment under the restriction 1 in the above results in uneven node distribution, large coverage redundancy and blind areas; while IGWO and GWO improve the distribution, there are still uncovered areas and redundant nodes. In contrast, GVGWO optimizes the significantly uniform distribution of nodes, almost completely covers the monitoring area, and shows the best coverage effect. The simulation shows that GVGWO has great potential in improving the coverage performance of FMWSN and is of great significance for building an efficient forest monitoring system. At the same time, the faulty nodes enter the dormant state, which reduces the number of sensor nodes actually participating in the optimization to Figure 3 50 of them and Figure 4 30 of them (the limit in Table 2 is the actual number of sensor nodes after considering the failure rate).

[0183] By combining the improved gray wolf optimization algorithm with the tangency reward item, the present invention simulates the sensor node layout in forest monitoring. The experimental results are shown in Table 2, which verifies the effectiveness of this method in the deployment of complex wireless sensor networks. The simulation results show that the proposed new algorithm significantly improves the network coverage and reduces node redundancy. In particular, by introducing the tangency reward item, the redundancy of the sensor nodes is effectively controlled, showing good application prospects. The present invention provides a new idea for achieving high coverage of sensor node layout, and also lays a solid foundation for further research in related fields.

[0184] During the verification process, the GVGWO algorithm flow chart is as follows Figure 6 As shown. The complete algorithm flow is as follows: after the parameters are initialized, the fitness value is calculated and sorted, and the three wolf leaders α, β, and γ are recorded according to the sorting, and inheritance and mutation are performed according to the corresponding strategy. After inheritance and mutation, the fitness value is calculated again and sorted, and then the nonlinear convergence factor is calculated, the position of the ordinary wolf is updated, the fitness value is calculated, and the Cauchy mutation strategy is applied to return the wolf individual with the best fitness, completing an iteration, and ending when the number of iterations reaches the preset value (maximum number of iterations). The iteration termination condition can also be set to achieve the expected coverage rate.

[0185] also, Figure 2 , Figure 3 and Figure 4 A visual analysis of coverage and sensor node distribution is demonstrated. Experimental results show that the algorithm proposed in this invention not only performs well in improving the coverage of forest monitoring networks, but also greatly reduces redundant coverage through the tangent reward mechanism, significantly improves the overall performance and stability of the wireless sensor network, and further proves the feasibility and practicality of the algorithm.

[0186] In the simulation experiment, the results of the optimization algorithm proposed in this invention combined with the tangency bonus item show that the new algorithm is effective in dealing with complex node layout optimization problems. Compared with the traditional random deployment method, the coverage rate of the optimized layout method has increased by more than 17%, and the redundancy of sensor nodes has been greatly reduced, indicating that the improved algorithm has significantly improved coverage efficiency, resource utilization and redundancy control. This optimized layout provides a more efficient and intelligent solution for forest monitoring, and the simulation study provides important theoretical support and practical reference for the further optimization and application of the algorithm.

Claims

1. A sensor node deployment method for forest monitoring, characterized in that The following steps are involved: S1, setting the sensing radius, communication radius and size of the monitoring area of ​​the sensor node, and generating random distribution information of the sensor node in the monitoring area, the sensor node is a wireless sensor node; S2. Based on the Boolean perception model, calculate the coverage of each sensor node to the monitoring area, and then establish a mathematical model of the coverage rate of the sensor nodes; S3, introducing the coverage rate mathematical model and the tangency reward item into the arrangement process of the sensor nodes to form the objective function of the optimal arrangement of the wireless sensor network; S4, introduce the gray wolf optimization algorithm GWO and improve it, take the objective function in S3 as the fitness function, and gradually optimize the position of sensor nodes by simulating the hunting behavior of gray wolves; the improvements include replacing the linear convergence factor of GWO with a nonlinear convergence factor, introducing a genetic mutation strategy and a mutation mechanism based on Cauchy distribution in GWO; S5. On the basis of random deployment, by collecting random distribution information of sensor nodes, and then dynamically adjusting the location distribution of sensor nodes according to the deployment strategy generated by the improved GWO, the coverage of the wireless sensor network is improved and the energy consumption of sensor nodes is optimized; After multiple iterations, the monitoring area is covered to the maximum extent. When the iteration termination condition is met, the final position of each sensor node is output to obtain the final optimized sensor node layout plan; In S1, it is assumed that the sensor nodes in the wireless sensor network are homogeneous and the sensing radius of the sensor nodes is R. p , the communication radius is R c , the sensor node set V is represented as V={v1, v2, v3, …, v n }, n is the number of sensor nodes, any sensor node v in V i The two-dimensional coordinates of i ,y i ), i=1, 2, 3, …, n; monitoring nodes are set in the monitoring area, and the monitoring node set Q is expressed as Q={q1, q2, q3, …, q m }, m is the number of monitoring nodes, any monitoring node q in Q j The coordinates of (x j ,y j ), j = 1, 2, 3, …, m; the monitoring area is defined as a rectangular area of ​​size L × W; In the above S2, the method for establishing the coverage mathematical model of the sensor nodes is: S21. Calculate the distance matrix D between the sensor node and the monitoring node. The formula is as follows: ； Where D ij Indicates v i With q j The distance between each D ij The overall distance matrix D can be obtained; S22, calculate the distance matrix A between each sensor node, the formula is as follows: ； In the formula, A ik Indicates v i With sensor node v k The distance between them, k = 1, 2, 3, ..., n and k ≠ i; calculate each A ik The overall distance matrix A can be obtained; S23, based on the Boolean perception model, a single sensor node is j The monitoring probability formula is: ； In the formula, p er (v i ,q j ) indicates v i Can monitor q j The probability of S24, all sensor nodes are q j The joint perception probability is expressed as: ； Where P joint (V, q j ) represents the set of all sensor nodes V for monitoring node q j The joint coverage probability of S25. Divide the monitoring area into L×W grids of equal area. The center of each grid represents a monitoring node. Then the coverage rate C r It is expressed as: ； In the formula, x and y represent the coordinates of the center of the grid, which are continuously increased from the origin of the coordinate axis to represent the center point of L×W grids; That means monitoring nodes; through C r Indicates the proportion of the monitoring area covered by the wireless sensor network; In S3, the tangent reward item It is expressed as: ； In the formula, is the tolerance range of the distance; E is the indicator function, when the conditions in the brackets are met, E is equal to 1, otherwise E is equal to 0; (x i ,y i ) is v k The two-dimensional coordinates of Objective Function for Optimal Deployment of Wireless Sensor Networks It is expressed as: ； In the formula, is a weight factor that balances the contribution of tangency rewards; In the above S4, the nonlinear convergence factor is expressed as: ； Where MaxIter is the maximum number of iterations; iter is the current number of iterations; The genetic mutation strategy includes the genetic strategy and the mutation strategy. The genetic strategy is expressed as: ； In the formula, mod means taking the remainder; is the congruence symbol; , , They represent the three wolf individuals with the highest fitness in the population; p represents the pth solution in the population; X p represents the pth wolf individual in the next generation; The mutation strategy is to randomly select dimension d r : ； In the formula, d is the size of the population dimension; then: ； In the formula, Indicated in dimension d r The pth wolf individual after the mutation; r is a random number uniformly distributed in [0, 1], used to determine whether a mutation occurs; Q T is the mutation rate, which indicates the probability of mutation; N (0, 1) is the standard normal distribution, which is used to introduce random disturbances; ms is the mutation intensity, which controls the intensity of mutation applied to individual wolves; The variation mechanism of the Cauchy distribution is, and the Cauchy distribution function is expressed as: ； Where C represents the Cauchy distribution variable used for mutation; is the location parameter indicating the peak of the distribution; is a scale parameter that controls the discreteness of the distribution; rand is a uniformly distributed random number between 0 and 1; After obtaining the current global optimal solution, the Cauchy mutation is used to update the current global optimal solution, which is expressed as: ； In the formula, Represents the new coordinates of the global optimal solution in the qth dimension; Represents the coordinates of the global optimal solution in the qth dimension; represents random numbers drawn from a standard Cauchy distribution; In the above S5, the specific method for obtaining the final optimized sensor node arrangement scheme is: S51. In the improved GWO, the set of sensor nodes is expressed as: ； In the formula, X represents the overall matrix; represents the position of the Nth wolf individual in the search space under dimension d; Represents the dth dimension of the Nth wolf individual; each wolf individual contains the number and location information of all sensor nodes; S52, the initial position of the Ith wolf is X I , I = 1, 2, 3, ..., N, N is the population size; each wolf individual represents a potential deployment plan, and the position of each wolf individual in the solution space is defined by a set of candidate variable values, X I Given by: ； In the formula, , They represent the lower and upper bounds of the d-dimensional variable in the I-th candidate variable respectively; R is a random number uniformly distributed in the range of [0, 1]; S53, according to the value of fitness, the wolf individuals are divided into four levels, from the highest to the lowest, namely α, β, γ and ω, where α, β and γ represent the top three wolf individuals in fitness, that is, the three alpha wolves, corresponding to the current best solution, the second best solution, and the general solution respectively; the remaining wolf individual ω is an ordinary wolf, representing a candidate solution; the fitness function is the objective function in S3; S54. Execute the encirclement strategy in GWO and calculate the distance between the common wolf and the leader wolf: ； ； In the formula, is the vector representing the distance between the common wolf and the alpha wolf Y; Is the impact The coefficient of represents the position vector of the leader wolf Y at time t; represents the position vector of an ordinary wolf at time t; is a random vector uniformly distributed in the range [0, 1]; S55. Update the position of the common wolf according to the distance between the common wolf and the leader wolf, expressed as: ； ； In the formula, factor is the nonlinear convergence factor in S4; Indicates the current position vector of the wolf leader Y; It is the coefficient that affects the update position. The adjustment amplitude is controlled by factor during calculation. Yes the updated position of the impact; is a random vector uniformly distributed in the range [0, 1]; S56. The formula for updating the population position is: ； In the formula, is the position vector updated in the next time step, representing the average position of the current leader; S57, perform multi-directional encirclement. The three alpha, beta, and gamma alpha wolves surround the prey from different directions and explore the entire solution space through multi-directional search. S58, in the process of iterative search, by controlling the size of factor, dynamically adjusting the encirclement, and finding the optimal solution, the optimal solution is the sensor node arrangement scheme that covers the monitoring area to the maximum extent; The complete algorithm flow of the improved GWO is as follows: after the parameters are initialized, the fitness value is calculated and sorted, the three head wolves α, β, and γ are recorded according to the sorting, and inheritance and mutation are performed according to the corresponding strategies. After inheritance and mutation, the fitness value is calculated again and sorted, and then the nonlinear convergence factor is calculated, the position of the ordinary wolf is updated, the fitness value is calculated, and the Cauchy mutation strategy is applied to return the wolf individual with the best fitness, completing an iteration, and ending when the iteration termination condition is reached.

2. The sensor node deployment method for forest monitoring according to claim 1, characterized in that: In S1, set R c =2R p , ensuring that sensor nodes can communicate effectively and maintain a stable connection of the entire wireless sensor network.

3. The sensor node deployment method for forest monitoring according to claim 1, characterized in that: In the genetic strategy, the genetic strategy is applied to the three wolf individuals with the highest fitness, and their excellent characteristics are passed on to the five wolf individuals of the next generation in turn to ensure that the best individuals are retained and the diversity of the population is increased; at this time, p=1, 2, 3, 4, 5, indicating that the genetic strategy acts on the five wolf individuals in the offspring; in the mutation strategy, the wolf individuals ranked 4 to N in fitness will undergo mutation operations, and then the fitness will be re-ranked, and the individuals with the last two fitnesses will be eliminated to generate offspring individuals, and N is the population size.

4. The sensor node deployment method for forest monitoring according to claim 1, characterized in that: In the above-mentioned S58, during the iteration process, the factor is gradually reduced, thereby reducing the search range and gradually tightening the encirclement.

5. The sensor node deployment method for forest monitoring according to claim 1, characterized in that: In the above S5, the iteration termination condition is that the number of iterations reaches the maximum number of iterations of 300 times, or the coverage rate of the wireless sensor network for the monitoring area is higher than the set coverage rate threshold.

Images