An optimized deployment method of sonar buoys for a drone swarm to search for submarines

Abstract

The present invention discloses an optimized sonobuoy deployment method for unmanned aerial vehicle (UAV) swarms to search for submarines, belonging to the technical field of unmanned cluster command and control. The method includes: constructing a probability field model of the sea area to be searched based on the prior information of the sea area to be searched, and rasterizing the sea area to be searched based on the probability field model to obtain the probability field of the rasterized sea area to be searched; based on the probability field where submarines may be distributed and the limited number of sonobuoys, optimizing the deployment of sonobuoys in the sea area to be searched to obtain the deployment positions of each sonobuoy, so that the cumulative search probability meets the preset search probability requirement; using the UAV swarm route planning method to plan the routes of multiple UAVs to cover the deployment positions of each sonobuoy, and the total flight distance meets the preset distance requirement. The present invention solves the problem of optimized sonobuoy deployment for UAV swarms to search for submarines.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned cluster command and control, and more specifically, relates to a method for simulating the mission planning of an unmanned aerial vehicle (UAV) swarm for submarine search. Background Art

[0002] Inspection anti-submarine refers to conducting inspection and search in a specified sea area within a specific time to find out whether there are enemy submarines in the specified sea area. Using passive sonar and forming a corresponding buoy group to detect submarine information is an important means of inspection anti-submarine. The traditional inspection anti-submarine has the following problems: First, it is assumed that the submarine distribution in the specified sea area is evenly distributed. In actual situations, due to various factors, the distribution probability of submarines in the specified sea area does not follow a uniform distribution, and the probabilities of submarines appearing in different regions are often very different; Second, only when the number of buoys meets the requirements can the probability of successful submarine search reach the requirements. When the number of buoys does not meet the formation requirements, it is difficult to meet the specified requirements; Finally, the problem of multi-aircraft collaborative deployment is rarely considered. In actual situations, in order to deploy sonar buoys faster, multi-aircraft collaborative deployment is often carried out to shorten the deployment time as much as possible. Summary of the Invention

[0003] In view of the above-mentioned defects or improvement requirements of the prior art, the present invention proposes an optimized deployment method for sonar buoys in UAV swarm submarine search to solve the problem of optimized deployment of sonar buoys in UAV swarm submarine search.

[0004] To achieve the above object, the present invention provides an optimized deployment method for sonar buoys in UAV swarm submarine search, including:

[0005] (1) Based on the prior information of the sea area to be searched, construct a probability field model of the sea area to be searched, and rasterize the sea area to be searched based on the probability field model to obtain the probability field of the rasterized sea area to be searched;

[0006] (2) Based on the probability field where submarines may be distributed and the limited number of sonar buoys, optimize the deployment of sonar buoys in the sea area to be searched to obtain the deployment positions of each sonar buoy, so that the cumulative search probability meets the preset search probability requirements;

[0007] (3) Use the UAV swarm route planning method to plan the routes of multiple UAVs to cover the deployment positions of each sonar buoy, and the total flight distance meets the preset distance requirements.

[0008] In some alternative embodiments, step (1) includes:

[0009] (1.1) Assume that the sea area to be searched is Φ, the length of the sea area Φ to be searched is L, and the width is W. The sea area to be searched is divided into square grids with a length of d0. na = round(L / d0), nb = round(W / d0), where round() is a rounding and integer-taking function. na is the number of grids divided along the long side of the sea area to be searched, and nb is the number of grids divided along the wide side of the sea area to be searched. The sea area Φ to be searched is divided into n grids with a length and width of d0, where n = na × nb. Denote the divided grids as N = {1, 2, 3, …, n}.

[0010] (1.2) Through other prior information, it is known that the probability of the submarine appearing in the i-th grid is Pi. Denote the probability of the submarine appearing in each grid as P = {P1, P2, P3, …, Pn}.

[0011] In some alternative embodiments, step (2) includes:

[0012] (2.1) Select chromosomes. Among them, the chromosomes are fixed-length chromosomes. One locus of a chromosome is the placement position (x, y) of a sonobuoy, where x represents the abscissa of the sonobuoy, and y represents the ordinate of the sonobuoy. The fixed length is the number of sonobuoys.

[0013] (2.2) Use random numbers to generate an initial population. Each sonobuoy generated by random numbers is evenly distributed in the grids of the sea area to be searched. Among them, each individual in the population represents a chromosome.

[0014] (2.3) Obtain the fitness value of each individual in the current population, and determine whether each fitness value meets the fitness requirements. If not, enter the next step (2.4). Among them, the larger the fitness value, the better the sonobuoy placement plan.

[0015] (2.4) Select individuals from the previous generation population that meet the preset fitness requirements.

[0016] (2.5) For the optimal individual in the current population, no crossover operation is performed. For the remaining individuals, a crossover operation is adopted.

[0017] (2.6) Perform a mutation operation on the offspring generated by the selection and crossover operations.

[0018] (2.7) Execute step (2.3) on the mutated current population to continue calculating the fitness value of each individual in the current population until each fitness value meets the fitness requirements, and obtain the optimal population. Among them, each individual in the optimal population corresponds to the optimal placement position of each sonobuoy.

[0019] In some alternative embodiments, let the number of sonar buoys be M. If the i-th sonar buoy is at the position (x, y) and the detection range of the i-th buoy is r, denote the number of grids within the range r as m, and the set of all grids within the range r as The probability that the submarine appears in the grids within the range r is denoted as Then the fitness corresponding to the chromosome is:

[0020] In some alternative embodiments, step (2.6) includes:

[0021] From Determine the displacement in mutation, where the fitness of the optimal individual is fmax, the fitness of the i-th individual in the population is fi, and k is a constant mutation coefficient, is the constant term coefficient;

[0022] Obtain the mutated position from xj = xj + step and yj = yj + step. xj represents the x-coordinate of the j-th buoy, and yj represents the y-coordinate of the j-th buoy.

[0023] In some alternative embodiments, step (3) includes:

[0024] Use L anti-submarine drones, with each drone carrying K sonar buoys, so that each drone throws K sonar buoys to cover the deployment positions of each sonar buoy, and the total flight distance of the L anti-submarine drones is the shortest.

[0025] In some alternative embodiments, step (3) includes:

[0026] (3.1) Select a chromosome. Each locus in the chromosome represents the serial number of a deployment position. The first position of each gene sequence is the starting point, and the ending position is the end point. Each chromosome has a set of split points. The size of the set of split points is L - 1, where L is the number of drones, and the split points represent the interval points for different drones to fly;

[0027] (3.2) The fitness is the total flight distance of the drone swarm. If the fitness of an individual is greater, it means the route allocation plan is better;

[0028] (3.3) Obtain the fitness value of each individual in the current population, and determine whether each fitness value meets the fitness requirement. If not, perform selection, crossover, and mutation operations to obtain the next generation population, calculate the fitness value of the next generation population, and continue until the fitness value of the current population meets the fitness requirement to obtain an optimized multi-aircraft route planning plan.

[0029] In some alternative embodiments, in step (3.2), from Determine the fitness value, where L is the number of L UAVs, and di is the total flight distance of the i-th UAV.

[0030] Generally speaking, compared with the prior art, the above technical solution conceived by the present invention can achieve the following beneficial effects:

[0031] For a given prior probability field, optimize the sonobuoy placement so that the cumulative search probability for the area to be detected is maximized as much as possible. For the obtained optimized sonobuoy placement plan, perform the route planning of the UAV swarm so that the total flight distance is minimized as much as possible. Brief Description of the Drawings

[0032] Figure 1 It is a flowchart for calculating the sonobuoy placement plan for optimizing the sonobuoy placement of UAV swarm anti-submarine search provided by an embodiment of the present invention;

[0033] Figure 2 It is a schematic diagram of the prior probability field obtained through prior information provided by an embodiment of the present invention;

[0034] Figure 3 It is a schematic diagram of the result of calculating the sonobuoy placement plan by a genetic algorithm provided by an embodiment of the present invention;

[0035] Figure 4 It is a schematic diagram of the placement path of the planned UAV swarm provided by an embodiment of the present invention. Detailed Embodiment

[0036] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0037] The problem of optimizing the deployment of sonar buoys for anti-submarine search by a swarm of unmanned aerial vehicles (UAVs) is as follows: First, based on the prior information of the sea area to be searched, the prior probability field where submarines may appear in the sea area to be searched is obtained. From the generated probability field of the sea area to be searched, the present invention is used to calculate the deployment positions of sonar buoys. The traditional deployment scheme of inspection anti-submarine buoys is generally based on empirical models, and specific deployment schemes are adopted for specific situations. For example, in the case of a large area and few buoys, the "X" formation is adopted, and in the case of a small area and many buoys, a dense formation is adopted. Different from the traditional deployment scheme of inspection anti-submarine buoys, the present invention is based on the probability field of possible submarine distributions and a limited number of sonar buoys, and uses the method proposed by the present invention to obtain an optimized deployment scheme of sonar buoys. This deployment scheme is specifically the deployment position of the sonar buoy, which is an (x, y) coordinate on a two-dimensional coordinate system. After calculation by this method, it contains N coordinate points to be deployed, and this deployment scheme can achieve a relatively high cumulative search probability. For the sonar buoy deployment scheme, the route planning of the UAV swarm is carried out. The goal of the UAV swarm route planning is that each part of the aircraft deployment scheme finally makes the total flight distance as short as possible.

[0038] A method for optimizing the deployment of sonar buoys for anti-submarine search by a swarm of UAVs proposed by the present invention mainly includes several steps such as constructing a probability field model of the sea area to be searched, a deployment scheme of sonar buoys, and route planning of the UAV swarm.

[0039] Step 1: Construct a probability field model. The sea area to be searched is gridded. N = {1, 2, 3,..., n} represents each grid after the gridding of the sea area to be searched, and P = {P1, P2, P3,..., Pn} represents the probability that a submarine appears in the corresponding grid. For example, if the probability that a submarine appears in grid 1 is 0.7, then P1 = 0.7.

[0040] Step 2: Use the prior probability field model obtained in Step 1. Under the condition of limited buoy numbers, the optimized deployment of sonar buoys for anti-submarine search by a swarm of UAVs is used for calculation to find a buoy deployment scheme that makes the cumulative search probability as high as possible when the number of buoys is limited and determined.

[0041] Step 3: Based on the deployment scheme obtained in Step 2, use the route planning of the UAV swarm to deploy sonar buoys in the sea area to be searched, so that the planned path of the UAV swarm is as short as possible.

[0042] Further, in step 1, assume that the sea area to be searched is Φ, the length of the sea area to be searched Φ is L, and the width is W. The sea area to be searched is divided into square grids with a length of d0. na = round(L / d0), nb = round(W / d0). In the above formula, round() is a rounding and integer-taking function. na is the number of grids divided along the long side of the sea area to be searched, and nb is the number of grids divided along the wide side of the sea area to be searched. The sea area to be searched Φ is divided into n grids with a length and width of d0, where n = na × nb. Denote the divided grids as N = {1, 2, 3, …, n}.

[0043] Further, in step 1, through other prior information, the probability of the submarine appearing in the i-th grid is Pi. Denote the probability of the submarine appearing in each grid as P = {P1, P2, P3, …, Pn}. If the submarine must appear in grid i, then Pi = 1. If the submarine must not appear in grid i, then Pi = 0. That is, if the probability of the submarine appearing in grid i is P, then Pi = P.

[0044] As Figure 2 shown,[[]]END]] Figure 2 is the prior probability field after rasterization. The density of the scatter points represents the probability of the submarine appearing here. The denser the scatter points, the greater the probability of the submarine appearing here, and the sparser the scatter points, the smaller the probability of the submarine appearing here.

[0045] Further, in step 2, use the sonobuoy optimal deployment method for UAV swarm anti-submarine search proposed in the present invention to calculate the deployment plan of sonobuoys. In the sonobuoy optimal deployment for UAV swarm anti-submarine search proposed in the present invention, the following aspects need to be considered: chromosome, initial population, fitness function, selection operator, crossover operator, mutation operator.

[0046] (1) Chromosome: A chromosome is generally a unit that maintains the characteristics of a species. In the sonobuoy optimal deployment method for UAV swarm anti-submarine search, a single chromosome is a solution to the problem to be solved. Through operations such as selection, crossover, and mutation, find the chromosome required to solve the problem, that is, find an approximate optimal solution to the problem. The quality of chromosome selection also determines the convergence speed of this method. A suitable gene encoding is of great benefit to the implementation of this method. In the present invention, the chromosome is a fixed-length chromosome. A locus of the chromosome is the coordinate of a sonobuoy deployment (two-dimensional coordinate (x, y)), and the fixed length is the number of sonobuoys. For example, if the number of sonobuoys is M, then the abscissa and ordinate of the i-th buoy are a locus of the chromosome. Denote a single chromosome as V = {v1, v2, v3, …, vM}, where vi, i = 1, 2,.., M represents the abscissa and ordinate of the i-th sonobuoy.

[0047] (2) Initial population: The initial population is the initial solution for the optimal deployment of sonar buoys in the search for submarines by an unmanned aerial vehicle (UAV) swarm. Since the initial population has not undergone operations such as selection, crossover, and mutation, it has a high degree of randomness. In the initial population, there may be good solutions as well as bad solutions. In the present invention, since the submarine probability field generated in step 1 is not uniformly distributed, the areas with the highest probability of submarine appearance may be scattered at various positions in the sea area to be searched. Therefore, the initial population is generated using random numbers, and the buoys generated by the random numbers are evenly distributed in the grid of the sea area to be searched. Since the randomly generated initial population enlarges the search scope of the solution of the present method, the present invention uses random numbers to generate the initial population.

[0048] (3) Fitness function: The method for the optimal deployment of sonar buoys in the search for submarines by an unmanned aerial vehicle (UAV) swarm conforms to Darwin's natural law of survival of the fittest, and there needs to be a criterion for judging the good and the bad. In this method, fitness values are used to evaluate the quality of individuals, and the fitness function is used to calculate the fitness values of each individual. In the present invention, the magnitude of the fitness represents the quality of the deployment points of a sonar buoy for an individual. If the fitness of an individual is larger, it represents that the sonar buoy deployment plan is better. If the fitness of a chromosome is smaller, it represents that the sonar buoy deployment plan is worse. In the present invention, the fitness is calculated in the following manner:

[0049] Let the number of sonar buoys be M. If the i-th sonar buoy is at the position (x, y) and the detection range of the i-th buoy is r, denote the set of all grids within the range of r as The probability that a submarine appears in the grids within the range of r is denoted as Then the fitness corresponding to the chromosome is:

[0050] If the fitness corresponding to a chromosome is larger, it represents that the chromosome is closer to the approximate optimal solution.

[0051] (4) Selection operator: Selection is to select individuals with high fitness from the previous generation. More individuals with higher fitness are selected, and at the same time, a small number of algorithms with low fitness are retained. More individuals with higher fitness are retained because individuals with high fitness have a high probability of being approximate optimal solutions. A small number of individuals with low fitness are retained because the currently selected individuals with high fitness are very likely to fall into a local optimal solution, so it is necessary to jump out of the local optimal solution. Common selection operators include roulette wheel algorithm, stochastic universal sampling algorithm, ranking selection algorithm, fitness scaling selection algorithm, tournament selection algorithm. In the present invention, the roulette wheel algorithm is used as the selection operator. In the roulette wheel selection (also known as fitness proportionate selection, FPS) algorithm, the probability of selecting an individual is directly proportional to its fitness value. The selection process in the roulette wheel algorithm is as follows:

[0052] (1) First, calculate the fitness of each individual in the population;

[0053] (2) For a single individual i in the population, the probability of being selected is:

[0054]

[0055] where f is the fitness value and M is the number of individuals in the population.

[0056] This is equivalent to using a roulette wheel in an entertainment venue and assigning fitness values to parts of each roulette wheel. When the roulette wheel is spun, the probability of each individual being selected is proportional to the size of the corresponding area. For individuals with high fitness, the area they occupy on the roulette wheel is larger, and for individuals with low fitness, the area they occupy on the roulette wheel is smaller. After using the roulette wheel algorithm, it is ensured that the probability of chromosomes with higher fitness being selected is greater than that of chromosomes with lower fitness. At the same time, there is also a certain probability that chromosomes with low fitness will be selected.

[0057] (5) Crossover operator: The crossover process is the process of generating the next generation of chromosomes from the parental chromosomes. In the method for optimizing the deployment of sonar buoys for underwater drone search, common crossover operators include single-point crossover, k-point crossover, and uniform crossover. In the present invention, the selection process is as follows:

[0058] For the optimal individual in the current population, no crossover operation is performed. For the remaining individuals, single-point crossover is used. First, select two individuals, then select a chromosome segment from the two individuals, and finally exchange the corresponding segments of the selected chromosomes to achieve the purpose of crossover.

[0059] (6) Mutation operator: The mutation operator is applied to the offspring generated by the selection and crossover operations. Mutation is based on a certain probability and usually occurs with a (very) low probability because it carries the risk of damaging the performance of the individual to which this operator is applied. Mutation is to prevent the algorithm from converging to a local optimal solution.

[0060] In the present invention, a concept of mutation displacement is introduced. For individuals with higher fitness, the range of mutation displacement is smaller. For individuals with lower fitness, the range of mutation displacement is larger.

[0061] The calculation method of the displacement in mutation is as follows:

[0062]

[0063] Assume that in the current population, the fitness of the optimal individual is fmax, and the fitness of the i-th individual in the population is fi. In the above formula, k is a constant mutation coefficient, and in the present invention, k is taken as 1. is the coefficient of the constant term. To prevent the denominator from being 0, in the present invention, the value is taken as 0.1.

[0064] Under the accuracy determined by the evolutionary generation of the sonar buoy optimal deployment method for the unmanned aerial vehicle (UAV) group to search for submarines, the feasible solutions are distributed at different positions. To enable the individuals close to the local optimal solution to be retained. During mutation, the perturbation range is determined according to the fitness of the individual. The position after mutation can be expressed as:

[0065] xj = xj + step

[0066] yj = yj + step

[0067] xj represents the x coordinate of the j-th buoy, and yj represents the y coordinate of the j-th buoy.

[0068] As shown in the figure, Figure 1 are the implementation steps of the sonar buoy optimal deployment method for the UAV group to search for submarines. Figure 3 represents the sonar buoy deployment position obtained by the sonar buoy optimal deployment method for the UAV group to search for submarines. According to this deployment position, the subsequent route planning is carried out.

[0069] Further, in step 3, the UAV group route planning method is used to allocate UAVs for the deployment plan obtained in step 2.

[0070] The problem to be solved in step 3 is: there are L anti-submarine UAVs, and each UAV carries K sonar buoys, so that each UAV throws K sonar buoys and the total flight distance of the L anti-submarine UAVs is the shortest. This problem can be simplified to an MTSP problem with constraints. The MTSP problem is a traditional NP-hard problem. For the solution of this problem, the UAV group route planning method is used in the present invention to solve it.

[0071] In step 3, the chromosome encoding scheme is as follows:

[0072] The chromosome encoding method adopts the decimal encoding method. Each locus in the chromosome represents the serial number of a deployment position. The first position of each gene sequence is the starting point, and the ending position is the ending point. Each chromosome has a set of splitting points. The size of the set of splitting points is L - 1, where L is the number of UAVs. The splitting points represent the interval points for different UAVs to fly. The significance of adding splitting points is to distinguish the flight paths of different UAVs. For example: for a chromosome V = {v1, v2, v3,..., v15}, the splitting points are the set B = {v5, v9}, then the corresponding UAV group flight paths are v1->v2->v3->v4->v15, v5->v6->v6->v8->v15, v9->v10->v11->v12->v13->v14->v15.

[0073] The fitness is the total flight distance of the UAV swarm, and the magnitude of the fitness represents the quality of a route allocation scheme. If the fitness of an individual is larger, it means the route allocation scheme is better. If the fitness of an individual is smaller, it means the route allocation scheme is inferior. In the present invention, the fitness is calculated in the following manner:

[0074]

[0075] Wherein, L is L UAVs, and di is the total distance flown by the i-th UAV.

[0076] The selection strategy adopts the roulette wheel algorithm, and the crossover strategy adopts the partially matched crossover. The partially matched crossover ensures that each gene in a chromosome appears only once, and no duplicate genes will appear in a chromosome through this crossover strategy.

[0077] The partially matched crossover is similar to the two-point crossover, and the crossover region is determined by randomly selecting two crossover points. After performing the crossover, two invalid chromosomes are generally obtained, and individual genes may appear repeatedly. To repair the chromosomes, a matching relationship for each chromosome can be established within the crossover region, and then this matching relationship can be applied to the duplicate genes outside the crossover region to eliminate the conflicts.

[0078] The steps of the partially matched crossover are as follows:

[0079] Randomly select the start and end positions of several genes in a pair of chromosomes (the selected positions of the two chromosomes are the same)

[0080] Exchange the positions of these two groups of genes.

[0081] Perform conflict detection, and establish a mapping relationship based on the two groups of exchanged genes. Finally, all conflicting genes will go through the mapping to ensure that the newly formed pair of offspring genes has no conflicts.

[0082] For example:

[0083] There are currently gene sequences A(1,2,3,4,5,6,7,8) and gene sequence B(3,5,8,1,7,4,2,6)

[0084] The gene A segment (2,3,4,5) is exchanged with the gene B segment (5,8,1,7). At this time, a mapping relationship is established:

[0085] 2<->5, 3<->8, 4<->1, 2<->5<->7

[0086] After the exchange, the gene sequences A and B become:

[0087] A(4,5,8,1,7,6,2,3)

[0088] B(8, 2, 3, 4, 5, 1, 7, 6)

[0089] After the above steps, an optimized multi-aircraft route planning scheme is obtained.

[0090] Such as Figure 3 shown is the sonobuoy deployment position calculated in Step 2. After the route planning of the UAV swarm, the result is as Figure 4 shown.

[0091] It should be noted that according to the implementation needs, each step / component described in this application can be split into more steps / components, or two or more steps / components or partial operations of steps / components can be combined into new steps / components to achieve the purpose of the present invention.

[0092] Those skilled in the art can easily understand that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An optimized deployment method of sonar buoys for a drone swarm to search for submarines, characterized in that, it includes: (1) Based on the prior information of the sea area to be searched, construct a probability field model of the sea area to be searched, and rasterize the sea area to be searched based on the probability field model to obtain the probability field of the rasterized sea area to be searched; (2) Based on the probability field where submarines may be distributed and the limited number of sonar buoys, optimize the deployment of sonar buoys in the sea area to be searched to obtain the deployment positions of each sonar buoy, so that the cumulative search probability meets the preset search probability requirement; (3) Use the drone swarm route planning method to plan the routes of multiple drones to cover the deployment position of each sonar buoy, and the total flight distance meets the preset distance requirement; Step (2) includes: (2.1) Select chromosomes. Among them, the chromosome is a fixed-length chromosome, and a locus of the chromosome is the deployment position (x, y) of a sonar buoy. x represents the abscissa of the sonar buoy, and y represents the ordinate of the sonar buoy. The fixed length is the number of sonar buoys; (2.2) Use random numbers to generate an initial population. Each sonar buoy generated by random numbers is evenly distributed in the grid of the sea area to be searched. Among them, each individual in the population represents a chromosome; (2.3) Obtain the fitness value of each individual in the current population, and judge whether each fitness value meets the fitness requirement. If not, enter the next step (2.4). Among them, the larger the fitness value, the better the sonar buoy deployment plan; (2.4) Select individuals that meet the preset fitness requirement from the previous generation population; (2.5) For the optimal individual in the current population, no crossover operation is performed, and for the remaining individuals, crossover operations are adopted; (2.6) Perform mutation operations on the offspring generated by selection and crossover operations; (2.7) Execute step (2.3) on the mutated current population to continue calculating the fitness value of each individual in the current population until each fitness value meets the fitness requirement to obtain the optimal population. Among them, each individual in the optimal population corresponds to the optimal deployment position of each sonar buoy; Step (3) includes: (3.1) Select chromosomes. Each locus in the chromosome represents the serial number of a deployment position. The first position of each gene sequence is the starting point, and the ending position is the ending point. Each chromosome has a set of segmentation points. The size S of the set of segmentation points is equal to Q – 1, where Q is the number of drones, and the segmentation points represent the interval points for different drones to fly; (3.2) The fitness is the total flight distance of the drone swarm. If the fitness of an individual is greater, it means the route allocation plan is better; (3.3) Obtain the fitness value of each individual in the current population, and judge whether each fitness value meets the fitness requirement. If not, perform selection, crossover, and mutation operations to obtain the next generation population, calculate the fitness value of the next generation population, until the fitness value of the current population meets the fitness requirement to obtain an optimized multi-drone route planning plan.

2. The method according to claim 1, characterized in that, Step (1) includes: (1.1) Suppose the sea area to be searched is Φ, the length of the sea area Φ to be searched is L, and the width is W. The sea area to be searched is divided into square grids with a length of d0. na = round(L / d0), nb = round(W / d0), where round() is a rounding and integer-taking function. na is the number of grids divided along the long side of the sea area to be searched, and nb is the number of grids divided along the wide side of the sea area to be searched. The sea area Φ to be searched is divided into n grids with a length and width of d0, where n = na × nb. The grids thus divided are denoted as N = {1, 2, 3, …, n}; (1.2) Through other prior information, the probability of the submarine appearing in the i-th grid is known to be Pi. The probabilities of the submarine appearing in each grid are denoted as P = {P1, P2, P3, …, Pn}.

3. The method according to claim 2, characterized in that, Let the number of sonar buoys be M. If the i-th sonar buoy is at the position (x, y) and the detection range of the i-th buoy is r, denote the number of all grids within the range of r as m, and the set of all grids within the range of r as , and the probability that the submarine appears in the grids within the range of r is denoted as . Then the fitness corresponding to the chromosome is: .

4. The method according to claim 3, characterized in that, Step (2.6) includes: Determined by Determine the displacement in the mutation, where the fitness of the optimal individual is fmax, the fitness of the i-th individual in the population is fi, and k is the constant mutation coefficient, Is the constant term coefficient; Obtaining the mutated position from xj = xj + step and yj = yj + step, where xj represents the x-coordinate of the j-th sonobuoy, and yj represents the y-coordinate of the j-th sonobuoy.

5. The method according to claim 4, characterized in that, Step (3) includes: Using Q anti-submarine drones, each carrying K sonobuoys, such that each drone throws K sonobuoys to cover the deployment positions of each sonobuoy, and the total flight distance of the Q anti-submarine drones is the shortest.

6. The method according to claim 5, characterized in that, In step (3.2), it is determined by the fitness value, where Q is the Q-frame UAVs, and di is the total distance flown by the i-th UAV.

Images