Multi-uav formation cooperative combat task allocation method and system

Abstract

This invention provides a task allocation method and system for multi-aircraft formation cooperative operations. The method involves initializing information about formation members and targets; initializing sparrow search algorithm parameters to obtain an initial sparrow population; using Logistic chaotic mapping to initialize the sparrow population; improving the fitness function of the sparrow search algorithm based on the sparrow fitness function index; obtaining an initial allocation scheme based on the initialized information about formation members and targets and an allocation scheme encoding; calculating an initial optimal allocation scheme based on the initial allocation scheme and the improved sparrow search algorithm; and dynamically tracking the target to be executed by the UAV according to the initial optimal allocation scheme. This invention introduces chaotic operators and a Gaussian walk strategy, improves its position update formula, and proposes an improved sparrow search algorithm to solve the problem of multi-target task allocation in multi-aircraft formation cooperative attacks.

Description

Technical Field

[0001] This invention relates to the technical field of task allocation, specifically to a task allocation method and system for multi-aircraft formation cooperative operations. More particularly, it preferably relates to a task allocation method for multi-aircraft formation cooperative operations based on an improved sparrow search algorithm. Background Technology

[0002] In the face of a multidimensional battlefield environment, aircraft possess significant advantages such as high cost-effectiveness and large-scale deployment, enabling them to assist or even replace humans in completing high-risk missions. However, individual UAVs have limited payload and weak information processing capabilities, making multi-aircraft formation cooperative operations a focus of widespread attention worldwide. In multi-aircraft formation cooperative operations, the task allocation problem among multiple targets is one of the important issues requiring further research.

[0003] In the task allocation problem, commonly used algorithms include Sparrow Search Algorithm (SSA), Genetic Algorithm (GA), Ant Colony Optimization (ACO), and Particle Swarm Optimization (PSO). Traditional heuristic algorithms perform well in solving allocation model problems.

[0004] Chinese invention patent document CN103279793A discloses a method for task allocation in unmanned aerial vehicle (UAV) formations under a given environment, belonging to the field of UAV technology. The method includes the following steps: determining the encoding sequence of the task allocation algorithm; determining the advantage function for UAV formations to perform tasks; determining the velocity update formula and position update formula for the discrete particle swarm optimization algorithm; determining the tabu search process; and determining the hybrid optimization process.

[0005] Regarding the aforementioned technologies, the inventors believe that existing task allocation model platforms and task characteristics are relatively simple, and most algorithms suffer from problems such as complex structure and slow convergence, which cannot meet the real-time requirements of task allocation in dynamic battlefield environments. Summary of the Invention

[0006] To address the shortcomings of existing technologies, the purpose of this invention is to provide a task allocation method and system for multi-aircraft formation cooperative operations.

[0007] According to the present invention, a task allocation method for multi-aircraft formation cooperative operations is provided, which improves the sparrow search algorithm; and the improved sparrow search algorithm is used to allocate tasks for multi-aircraft formation cooperative operations.

[0008] Preferably, the method includes the following steps:

[0009] Information initialization steps: Initialize the information of the formation combat members and targets;

[0010] Steps to define time slice length: Define the task time slice interval t s ;

[0011] Initialize algorithm parameters: Initialize the sparrow search algorithm parameters to obtain the initial sparrow population;

[0012] Initial scheme acquisition steps: Initialize the sparrow population using Logistic chaotic mapping, improve the fitness function of the sparrow search algorithm based on the sparrow fitness function index, and obtain the initial allocation scheme based on the information of the initial formation combat members and targets and the allocation scheme encoding.

[0013] Steps for obtaining the initial optimal solution: Calculate the initial optimal allocation scheme based on the initial allocation scheme and the improved sparrow search algorithm;

[0014] Dynamic tracking steps: The UAV dynamically tracks the mission target to be executed according to the initial optimal allocation scheme;

[0015] Preferably, the method further includes the following steps:

[0016] Change judgment steps: At the beginning of the first time slice, obtain the drone failure information and whether the target has changed;

[0017] Steps for finding the optimal solution: Within the time slice, based on the changed information, randomly generate an allocation scheme, repeat the initial optimal solution acquisition steps, and seek the optimal allocation scheme after the information changes;

[0018] Result output steps: at the next time slice or at t+t s At each step, repeat the decision-making and optimal solution-seeking steps until t > t. min Where t represents time, t min This represents the estimated minimum mission flight time;

[0019] At the end of the time slice, based on the optimal allocation scheme of the optimal solution search steps, the drone's position information is updated using the tracking method, and the optimal allocation result is output.

[0020] Preferably, in the information initialization step, before the allocation begins, the aircraft platform information, weapon information, and kill probability of enemy mission targets are acquired; information on enemy mission targets is detected, including target location, quantity, type, and mobility, and the threat probability of enemy targets to our personnel is calculated.

[0021] Preferably, in the initialization algorithm parameter step, the basic parameters of the sparrow search algorithm are set, with the number of sparrows in the flock being N, and the number of finders being P. num The number of scouts is S num The problem to be optimized is D-dimensional, and the maximum number of iterations is Gen. max The lower bound of the search interval is lb; the upper bound of the search interval is ub; the sparrow flock, discoverer, and scout are all internal parameters of the sparrow search algorithm.

[0022] Preferably, in the initial scheme acquisition step, a Logistic mapping is introduced to initialize the sparrow colony:

[0023] x n+1 =μx n (1-x n ),x n ∈(0,1)····· (1)

[0024] In the formula: μ is the chaos control parameter; x n Let x be the value of the nth iteration. n+1 This is the value of the (n+1)th iteration;

[0025] Define a sparrow population as having M individuals and a dimension of D.

[0026] Randomly generate the initial value x1 of the chaotic mapping;

[0027] Substituting x1 into equation (1) yields the remaining M-1 sparrows in the population under the Logistic mapping;

[0028] The variable values ​​mapped by the Logistic Regression plot are converted to the actual values ​​X of the sparrow flock using the following formula:

[0029] X i =X lb +(X lb -X ub )×(x i +1)×0.5 (2)

[0030] In the formula, x i Let X be the value of the i-th iteration of the Logistic mapping; lb Search the lower bound for the sparrow; X ub Search for the upper bound for the sparrow; X i This represents the actual value of the sparrow population in the i-th iteration.

[0031] Preferably, in the initial scheme acquisition step, the fitness function is calculated as follows:

[0032] The fitness of the entire population is represented as:

[0033]

[0034] In the formula, F X Here is the population fitness matrix, where f is the fitness value of a single sparrow; X is the sparrow population; X N,D Let N represent the Nth sparrow, and let D represent the objective function.

[0035] Preferably, the initial optimal solution acquisition step includes the following steps:

[0036] Sparrow population initialization steps: Initialize the sparrow population by randomly generating N sparrows and initializing the number of discoverers P. num The number of followers (NP) num and the number of scouts S num Initialize the value of D; Initialize the values ​​of the upper bound lb and the lower bound ub of the search interval;

[0037] Fitness calculation steps: Calculate the fitness value for each sparrow using the fitness function obtained in the initial scheme step, and select the P-value with the specified fitness. num Only one sparrow acts as the discoverer, and the rest are followers. S is randomly selected. num Only scouts;

[0038] Local Optimality Judgment Steps: Determine if the algorithm is trapped in a local optimum; if the sparrow search algorithm is not stalled, update the individual sparrow positions using the position update formula of the bird flock algorithm; if the algorithm stalls, update the sparrow flock positions using Gaussian walk.

[0039] Boundary control steps: Perform boundary control on sparrows, handle sparrows that have crossed the boundary, and obtain the new sparrow flock location and fitness;

[0040] Search steps: If Gen = Gen max The algorithm ends when the result is reached, where Gen represents the number of iterations. max The maximum number of iterations is set.

[0041] The optimal solution is found by analyzing the fitness values ​​output by the improved sparrow search algorithm under different schemes.

[0042] If Gen <Gen max If the condition is met, proceed to the fitness calculation step.

[0043] Preferably, in the local optimum determination step, the bird flocking algorithm is as follows:

[0044] The locations of the discoverers and followers in the flock are described as follows:

[0045]

[0046]

[0047] In the formula: t represents the current iteration number; j represents the j-th dimension of the objective function; This represents the j-th dimension value of the i-th sparrow when the number of iterations is t; represents the j-th dimension value of the i-th sparrow when the iteration number is t+1; randN(0,1) is a Gaussian random number with expectation of 0 and variance of 1; FL is the probability that the follower finds high-energy food; k represents the discoverer; This represents the j-th dimension value of the k-th discoverer when the number of iterations is t; rand(0,1) represents a random number generated in (0,1);

[0048] When the warning value is less than the safety value, substitute equation (4) into the finder position update formula in the sparrow search algorithm:

[0049]

[0050] In the formula: R2 is the warning value; ST is the safety value; Q is a random number from a standard normal distribution; L is a 1×D matrix with all elements being 1;

[0051] Meanwhile, substituting equation (5) into the follower position update formula in the sparrow search algorithm:

[0052]

[0053] in, Let k be the position of the sparrow with the worst fitness value at iteration number t; k represents the discoverer.

[0054] Furthermore, the scout's position update formula is:

[0055]

[0056] In the formula: β is a Gaussian random number with expected value of 0 and variance of 1; K∈[-1,1] is a random number; f i f represents the current fitness level of the sparrow; g The current optimal fitness; The position of the sparrow with the highest fitness value at iteration number t; where the sparrow with scouting behavior detects approaching danger, f i ≠f g When the sparrow was in a dangerous area on the edge, it moved to a safe area; i =f g This indicates that the sparrow is in the optimal position X. bj At this point, the sparrow moves closer to its companions; β and K are parameters that control the step size, allowing the sparrow to update to a safe position;

[0057] During the algorithm's iteration process, if the fitness value of the optimal solution at the (t+1)th generation selection is lower than that at the tth generation, the optimal position of the sparrow search algorithm is not updated; otherwise, the fitness value of the optimal solution at the (t+1)th generation selection is higher than or equal to that at the tth generation, and the optimal position and corresponding value are updated.

[0058] According to the present invention, a mission allocation system for multi-aircraft formation cooperative operations, and a mission allocation method for multi-aircraft formation cooperative operations, include the following modules:

[0059] Information initialization module: Initializes information about the formation's combat members and targets;

[0060] Initialize algorithm parameters module: Initialize the sparrow search algorithm parameters to obtain the initial sparrow population;

[0061] Initial scheme acquisition module: The sparrow population is initialized using Logistic chaotic mapping, the fitness function of the sparrow search algorithm is improved based on the sparrow fitness function index, and the initial allocation scheme is obtained based on the information of the initial formation combat members and targets and the allocation scheme encoding.

[0062] Initial optimal solution acquisition module: Calculates the initial optimal allocation scheme based on the initial allocation scheme and the improved sparrow search algorithm;

[0063] Dynamic tracking module: The UAV dynamically tracks the mission target to be executed based on the initial optimal allocation scheme.

[0064] Preferably, the system also includes the following modules:

[0065] Define time slice length module: Defines the task time slice interval t s ;

[0066] Change detection module: At the beginning of the first time slice, it acquires information on drone failure and whether the target has changed;

[0067] The optimal solution seeking module: Within the time slice, based on the changing information, a random allocation scheme is generated, and the initial optimal solution acquisition module is repeated to seek the optimal allocation scheme after the information changes;

[0068] Results output module: at the next time slice or at t+t s At each step, the decision-making module and the optimal solution-seeking module are repeatedly changed until t > t. min Where t represents time, t min This represents the estimated minimum mission flight time;

[0069] At the end of the time slice, based on the optimal allocation scheme of the optimal solution module, the drone's position information is updated using the tracking method, and the optimal allocation result is output.

[0070] Compared with the prior art, the present invention has the following beneficial effects:

[0071] 1. This invention addresses the weakness of the Sparrow Search algorithm, which is prone to getting trapped in local optima. It introduces chaotic operators and Gaussian walk strategy, improves its position update formula, and proposes an improved Sparrow Search algorithm to solve the problem of multi-target task allocation in multi-aircraft formation cooperative attacks.

[0072] 2. This invention designs an allocation code and fitness function. Simulation experiments verify the effectiveness of the improved SSA algorithm. Compared with task allocation simulation experiments based on SSA, GA, ACO, and PSO algorithms, the improved SSA algorithm is verified to be effective and superior. It ensures the optimal allocation scheme while also ensuring search speed.

[0073] 3. This invention outperforms general algorithms such as SSA, GA, ACO, and PSO in task allocation for multi-target coordinated attacks by multi-aircraft formations, and has significant application value. Attached Figure Description

[0074] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0075] Figure 1 This is a flowchart illustrating the task allocation method for multi-aircraft formation cooperative combat based on the improved sparrow search algorithm of the present invention.

[0076] Figure 2 Design index charts for functions;

[0077] Figure 3 A diagram showing the relationship between individual sparrows and the allocation scheme;

[0078] Figure 4 This is the initial battlefield distribution map for the simulation environment;

[0079] Figure 5 A task allocation diagram for drone formations attacking enemy targets;

[0080] Figure 6 A schematic diagram showing the optimal function and average function value for each iteration of the improved sparrow search algorithm;

[0081] Figure 7 The diagram shows the optimization results of drone swarm attack on enemy targets based on different algorithms. Detailed Implementation

[0082] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0083] Embodiment 1 of this invention discloses a task allocation method for multi-aircraft formation cooperative operations, such as... Figure 1 As shown, the sparrow search algorithm is improved; based on the improved sparrow search algorithm, task allocation is performed for multi-aircraft formation cooperative operations.

[0084] The method includes the following steps:

[0085] Information initialization steps: Initialize the information of the formation combat members and targets. Before allocation begins, acquire information on the aircraft platform, weapons, and the probability of kill on enemy mission targets; detect information on enemy mission targets, including target location, quantity, type, and mobility, and calculate the threat probability of enemy targets to our members.

[0086] Steps to define time slice length: Define the task time slice interval t s .

[0087] Initialize algorithm parameters: Initialize the sparrow search algorithm parameters to obtain the initial sparrow population. Set the basic parameters of the sparrow search algorithm: the number of sparrows in the population is N, and the number of finders is P. num The number of scouts is S num The problem to be optimized is D-dimensional, and the maximum number of iterations is Gen. max The lower bound of the search interval is lb; the upper bound of the search interval is ub; the sparrow flock, discoverer, and scout are all internal parameters of the sparrow search algorithm.

[0088] Initial scheme acquisition steps: Initialize the sparrow population using Logistic chaotic mapping, improve the fitness function of the sparrow search algorithm based on the sparrow fitness function index, and obtain the initial allocation scheme based on the information of the initial formation combat members and targets and the allocation scheme encoding.

[0089] Introducing Logistic mapping to initialize a sparrow colony:

[0090] x n+1 =μx n (1-x n ),x n ∈(0,1)····· (1)

[0091] In the formula: μ is the chaos control parameter; x n Let x be the value of the nth iteration.n+1 This is the value of the (n+1)th iteration.

[0092] Define a sparrow population as having M individuals and a dimension of D.

[0093] Randomly generate the initial value x1 for the chaotic mapping.

[0094] Substituting x1 into equation (1) yields the remaining M-1 sparrows in the population under the Logistic mapping.

[0095] The variable values ​​mapped by the Logistic Regression plot are converted to the actual values ​​X of the sparrow flock using the following formula:

[0096] X i =X lb +(X lb -X ub )×(x i +1)×0.5 (2)

[0097] In the formula, x i Let X be the value of the i-th iteration of the Logistic mapping; lb Search the lower bound for the sparrow; X ub Search for the upper bound for the sparrow; X i This represents the actual value of the sparrow population in the i-th iteration.

[0098] The fitness function is calculated as follows:

[0099] The fitness of the entire population is represented as:

[0100]

[0101] In the formula, F X Here is the population fitness matrix, where f is the fitness value of a single sparrow; X is the sparrow population; X N,D Let N represent the Nth sparrow, and let D represent the objective function.

[0102] Steps for obtaining the initial optimal solution: Calculate the initial optimal allocation scheme based on the initial allocation scheme and the improved sparrow search algorithm.

[0103] The steps to obtain the initial optimal solution include the following:

[0104] Sparrow population initialization steps: Initialize the sparrow population by randomly generating N sparrows and initializing the number of discoverers P. num The number of followers (NP) num and the number of scouts S num Initialize the value of D; Initialize the values ​​of the upper bound lb and the lower bound ub of the search interval.

[0105] Fitness calculation steps: Calculate the fitness value for each sparrow using the fitness function obtained in the initial scheme step, and select the P-value with the specified fitness. num Only one sparrow acts as the discoverer, and the rest are followers. S is randomly selected. num Only scouts.

[0106] Local Optimality Judgment Steps: Determine if the algorithm is trapped in a local optimum; if the sparrow search algorithm is not stalled, update the individual sparrow positions using the position update formula of the bird flock algorithm; if the algorithm stalls, update the sparrow flock positions using Gaussian walk.

[0107] The bird flocking algorithm is as follows:

[0108] The locations of the discoverers and followers in the flock are described as follows:

[0109]

[0110]

[0111] In the formula: t represents the current iteration number; j represents the j-th dimension of the objective function; This represents the j-th dimension value of the i-th sparrow when the number of iterations is t; represents the j-th dimension value of the i-th sparrow when the iteration number is t+1; randN(0,1) is a Gaussian random number with expectation of 0 and variance of 1; FL is the probability that the follower finds high-energy food; k represents the discoverer; represents the j-th dimension value of the k-th discoverer when the number of iterations is t; rand(0,1) represents a random number generated in (0,1).

[0112] When the warning value is less than the safety value, substitute equation (4) into the finder position update formula in the sparrow search algorithm:

[0113]

[0114] In the formula: R2 is the warning value; ST is the safety value; Q is a random number from a standard normal distribution; L is a 1×D matrix with all elements being 1.

[0115] Meanwhile, substituting equation (5) into the follower position update formula in the sparrow search algorithm:

[0116]

[0117] in, Let t be the position of the sparrow with the worst fitness value when the number of iterations is t; k represents the discoverer.

[0118] Furthermore, the scout's position update formula is:

[0119]

[0120] In the formula: β is a Gaussian random number with expected value of 0 and variance of 1; K∈[-1,1] is a random number; f i f represents the current fitness level of the sparrow; g The current optimal fitness; The position of the sparrow with the highest fitness value at iteration number t; where the sparrow with scouting behavior detects approaching danger, f i ≠f g When the sparrow was in a dangerous area on the edge, it moved to a safe area; i =f g This indicates that the sparrow is in the optimal position X. bj At this point, the sparrow moves closer to its companion; β and K are parameters that control the step size, allowing the sparrow to update to a safe position.

[0121] During the algorithm's iteration process, if the fitness value of the optimal solution at the (t+1)th generation selection is lower than that at the tth generation, the optimal position of the sparrow search algorithm is not updated; otherwise, the fitness value of the optimal solution at the (t+1)th generation selection is higher than or equal to that at the tth generation, and the optimal position and corresponding value are updated.

[0122] Gaussian walk is:

[0123] When the sparrow search algorithm converges and iterates stagnates during the search process, a Gaussian random walk strategy is introduced, utilizing the cosine function; the new sparrow's position update is described as follows:

[0124]

[0125]

[0126] In the formula: σ is the variance. Let t be the position of the sparrow. The position of the sparrow at time t+1; For the random discoverer sparrow; t max is the maximum number of iterations; Gaussian is the Gaussian distribution function.

[0127] Boundary control steps: Perform boundary control on sparrows, handle sparrows that have crossed the boundary, and obtain the new sparrow flock location and fitness.

[0128] Search steps: If Gen = Gen max The algorithm ends when the result is reached, where Gen represents the number of iterations. max This represents the maximum number of iterations. The optimal solution is found by analyzing the fitness values ​​output by the improved sparrow search algorithm under different schemes. If Gen <Gen max If the condition is met, proceed to the fitness calculation step.

[0129] Dynamic tracking steps: The UAV dynamically tracks the mission target to be executed according to the initial optimal allocation scheme.

[0130] Change judgment steps: At the beginning of the first time slice, obtain the drone failure information and whether the target has changed.

[0131] The steps to find the optimal solution are as follows: Within the time slice, based on the changing information, a random allocation scheme is generated, and the initial optimal solution acquisition steps are repeated to find the optimal allocation scheme after the information changes.

[0132] Result output steps: at the next time slice or at t+t s At each step, repeat the decision-making and optimal solution-seeking steps until t > t. min Where t represents time, t min This represents the estimated minimum mission flight time.

[0133] At the end of the time slice, based on the optimal allocation scheme of the optimal solution search steps, the drone's position information is updated using the tracking method, and the optimal allocation result is output.

[0134] Embodiment 1 of this invention discloses a mission allocation system for multi-aircraft formation cooperative operations, such as... Figure 1 As shown, the task allocation method for multi-aircraft formation cooperative operations includes the following modules:

[0135] Information initialization module: Initializes information about the formation's combat members and targets.

[0136] Initialize Algorithm Parameters Module: Initializes the sparrow search algorithm parameters to obtain the initial sparrow population.

[0137] Initial scheme acquisition module: The sparrow population is initialized using Logistic chaotic mapping, the fitness function of the sparrow search algorithm is improved based on the sparrow fitness function index, and the initial allocation scheme is obtained based on the information of the initial formation combat members and targets and the allocation scheme encoding.

[0138] Initial optimal solution acquisition module: Calculates the initial optimal allocation scheme based on the initial allocation scheme and the improved sparrow search algorithm.

[0139] Dynamic tracking module: The UAV dynamically tracks the mission target to be executed based on the initial optimal allocation scheme.

[0140] Define time slice length module: Defines the task time slice interval t s .

[0141] Change detection module: At the beginning of the first time slice, it acquires information on drone failure and whether the target has changed.

[0142] The optimal solution seeking module: Within the time slice, based on the changing information, a random allocation scheme is generated, and the initial optimal solution acquisition module is repeated to seek the optimal allocation scheme after the information changes.

[0143] Results output module: at the next time slice or at t+t s At each step, the decision-making module and the optimal solution-seeking module are repeatedly changed until t > t. min Where t represents time, t min This represents the estimated minimum mission flight time.

[0144] At the end of the time slice, based on the optimal allocation scheme of the optimal solution module, the drone's position information is updated using the tracking method, and the optimal allocation result is output.

[0145] Embodiment 2 of this invention discloses a task allocation method for multi-aircraft formation cooperative operations based on an improved sparrow search algorithm, such as... Figure 1 As shown, the UAV multi-target attack task allocation model is established:

[0146] To verify the effectiveness of the algorithm, a simulation experiment was conducted on the task allocation of 6 drones, each carrying 2 missiles, targeting 8 threatening enemy targets. Figure 4 and Figure 5 As shown in Table 1, solid pentagrams represent drones, solid small circles represent enemy targets, and a drone formation consists of 6 drones, each carrying 2 air-to-ground missiles. Their mission is to destroy the target, and their relevant attributes are shown in Table 1. The enemy consists of three types of targets: radar, anti-aircraft missiles, and anti-aircraft guns. Each target can only be attacked by a maximum of 1 friendly drone, and its relevant attributes are shown in Table 2.

[0147] Table 1. Drone Formation Information

[0148]

[0149] Table 1 shows the platform information and weapon resource information of the UAV formation, including the number of each UAV, its initial position coordinates, the type of weapon it carries, and the quantity, speed, and value of soft-kill weapons. Among these, the air-to-ground missiles carried by our formation members are AGM-1 and AGM-2. The parameters of the AGM-1 are: effective kill range r... M1 =10km, maximum tracking range of the drone R Umax1 =60km; AGM-2's parameters are: effective kill range r M2 =10km, maximum tracking range of the drone R Umax2 =60km.

[0150] Table 2 Task Objective Information

[0151]

[0152] Table 2 presents information about enemy mission targets, including target number, type, initial location, and value. Among these, the radar target performance parameter K... R =10e+12m 4 Maximum effective range R max =100km; Horizontal distance d near the boundary of the anti-aircraft missile kill zone sj =30km, horizontal distance d from the outer edge of the kill zone sy =100km; Effective fire radius R of anti-aircraft gun g =10km, radius of the cone's dead zone r g =0.9km.

[0153] like Figure 1 As shown, the task allocation method includes the following steps:

[0154] Step 1: Initialize the information of the formation combat members and targets. Specific information is shown in Table 1 and Table 2.

[0155] Specifically, before allocation begins, information such as aircraft platform details, weapon information, and kill probability against enemy mission targets are acquired; information on enemy mission targets is detected, including target location, quantity, type, and maneuverability, while calculating the threat probability of enemy targets to our personnel. The initial formation combat personnel and target information is used in step 4.

[0156] Step 2, define the task time slice interval t s The observation task time can be defined as: t0 = 0, t1 = t s t2=2t s ...Based on the estimated minimum mission flight time t... min Divide it into n t A time slice. Task time slice interval t s Used to proceed to the next time slice.

[0157] Specifically, the time for the observation task can be defined as: t i =it s , where i is a natural number, i = 0, 1, 2...; t i This represents the time of the i-th observation task. The task time slice interval t is defined as the observation task status. s .

[0158] Step 3, Initialize Algorithm Parameters. Set the basic parameters of the sparrow search algorithm: the number of sparrows in the flock is N, and the number of finders is P. num The number of scouts is S num The problem to be optimized is D-dimensional, and the maximum number of iterations is Gen.max The lower bound of the search interval is lb; the upper bound of the search interval is ub; the sparrow flock, discoverer, and scout are all internal parameters of the sparrow search algorithm. The initialized algorithm parameters are used in steps 4 and 5.1.

[0159] Specifically, the basic parameters of the sparrow search algorithm are set as follows: the number of sparrows in the flock is N = 30, of which 20% are discoverers, 10% are sparrows that have detected danger, the safety threshold is ST = 0.8, the problem to be optimized is D-dimensional, and the maximum number of iterations is Gen. max =100.

[0160] Step 4: Initialize the sparrow population using a Logistic chaotic mapping. Based on the sparrow fitness function index, derive the fitness function for the improved SSA algorithm. Obtain the initial allocation scheme based on the allocation scheme encoding, setting Gen = 1, where Gen represents the iteration number. Subsequent iterations are performed based on the initial allocation scheme.

[0161] Specifically, the Logistic mapping in step 4 is as follows:

[0162] As the number of iterations increases, the sparrow search algorithm is prone to getting trapped in local optima, leading to a decline in search performance. The Logistic mapping, however, is a typical chaotic mapping; the ergodicity of chaotic motion enables the algorithm to avoid getting trapped in local optima. Therefore, the Logistic mapping is introduced to initialize the sparrow swarm, improving the convergence of the sparrow search algorithm.

[0163] x n+1 =μx n (1-x n ),x n ∈(0,1)····· (1)

[0164] In the formula: μ is the chaos control parameter, which is generally taken between [0,4]; x n Let x be the value of the nth iteration. n+1 This is the value of the (n+1)th iteration.

[0165] Define a sparrow population with M individuals and D dimensions. The initialization steps of the sparrow search algorithm are as follows:

[0166] Sub-step 4.1: Randomly generate the initial value x1 of the chaotic mapping, where all elements in x1 are ±1;

[0167] Sub-step 4.2, substituting x1 into equation (1) yields the remaining M-1 sparrows in the population under the Logistic mapping;

[0168] Sub-step 4.3 converts the variable values ​​of the Logistic mapping to the actual values ​​X of the sparrow flock. The conversion formula is as follows:

[0169] X i =X lb +(X lb -X ub )×(x i +1)×0.5 (2)

[0170] In the formula, x i Let X be the value of the i-th iteration of the Logistic mapping. lb To search the lower bound for the sparrow, X ub To search for the upper bound of the sparrow, X i This represents the actual value of the sparrow population in the i-th iteration.

[0171] The fitness function in step 4 is calculated as follows:

[0172] The fitness of the entire population can be expressed as:

[0173]

[0174] In the formula, F X This is the population fitness matrix, where f is the fitness value of a single sparrow. X represents the sparrow population. N,D Let N represent the Nth sparrow, and let D represent the objective function.

[0175] The allocation scheme for step 4 is coded as follows:

[0176] like Figure 2 As shown, based on the fitness function of the improved SSA algorithm, the variable is Π (assignment scheme). The optimal scheme is found by analyzing the fitness values ​​of different schemes. UAV-Target represents the unmanned aerial vehicle (UAV) target; Missile-Target represents the missile target. X i-order This indicates the execution order of the i-th target.

[0177] In the SSA algorithm, each individual sparrow represents a potential solution to the search algorithm. The optimal position of a sparrow is found by updating its position during the sparrow flock search process. The key to applying the algorithm to task allocation is how to simulate a drone formation using sparrows, thus mapping the optimal position to the optimal allocation scheme.

[0178] When the drone swarm attacks a target, each swarm member carries an L i If there are multiple missiles, then the allocation scheme for drones can also be transformed into the allocation of missiles to targets. Let the set of drone formation members given by our side be U = {U...} i The enemy target set is T = {T, i = 1, 2, ..., q}. i Let i = 1, 2, ..., p. In setting up the improved SSA algorithm, the solution should be a 2×H matrix. q represents the total number of missiles carried by the UAV. H represents the number of missile formations. The 2*H matrix is ​​the subsequent A matrix. The first dimension represents the missile allocation, and the second dimension represents the mission execution order. Each element in the first dimension varies within [1, p], indicating which target p the missile attacks. Here, we assume our kill probability is 1, meaning each target is attacked only once. The elements in the second dimension belong to [1, c]. i ] indicates the drone U i Task execution order, c i For U drones i The number of tasks assigned. For example... Figure 3 The set of tasks to be executed assigned to U3 is {T7,T... 10 Based on the task time model, analyze the task time consumption and determine which task T should be executed first. 10 The cumulative task time is shorter, and the final allocation execution order is {T}. 10 →T7}. T7 and T 10 Both indicate enemy targets.

[0179] Step 5: When t=0, calculate the initial optimal allocation scheme according to the proposed improved sparrow optimization algorithm.

[0180] Specifically, the improved sparrow search algorithm in step 5 includes the following sub-steps:

[0181] Sub-step 5.1: Initialize the sparrow population, i.e., randomly generate N sparrows and initialize the number P of discoverers, followers, and scouts. num NP num and S num Initialize the value of D; Initialize the values ​​of the upper and lower bounds of the search interval, lb and ub;

[0182] Sub-step 5.2: Calculate the fitness value of each sparrow using the fitness function from step 4, and select the sparrow with the best fitness (P). num Only one sparrow acts as the discoverer, and the rest are followers. S is randomly selected. num Only scouts;

[0183] Sub-step 5.3: Determine if the algorithm is trapped in a local optimum; if the sparrow search algorithm is not stalled, update the individual sparrow positions using the position update formula of the bird flock algorithm; if the algorithm stalls, update the sparrow flock positions using Gaussian walk.

[0184] Sub-step 5.4 involves boundary control of the sparrows. This involves handling sparrows that have crossed the boundary and obtaining the new sparrow flock's location and fitness.

[0185] Sub-step 5.5, if Gen = Gen max The algorithm then terminates, where Gen represents the number of iterations; Gen maxTo determine the maximum number of iterations, the optimal solution is found by analyzing the fitness values ​​output by the improved sparrow search algorithm under different schemes; Gen <Gen max Then proceed to sub-step 5.2.

[0186] The bird flocking algorithm in sub-step 5.3 is as follows:

[0187] The Bird Swarm Algorithm (BSA) is based on the social behavior and interactions within bird flocks and exhibits good stability. The BSA algorithm is similar to the SSA algorithm in that the positions of discoverers and followers within the flock are described as follows:

[0188]

[0189]

[0190] In the formula: t represents the current iteration number, j represents the j-th dimension objective function, j∈[1,2,...,D], This represents the j-th dimension value of the i-th sparrow when the number of iterations is t. denoted as j-th dimension value of the i-th sparrow when the iteration number is t+1; randN(0,1) is a Gaussian random number with expectation of 0 and variance of 1; FL is the probability that the follower finds high-energy food, FL∈[0,2]; k represents the discoverer. Let represent the j-th dimension value of the k-th discoverer when the iteration number is t; rand(0,1) represents generating a random number (real number) in (0,1); the problem to be optimized is D-dimensional.

[0191] When the warning value is less than the safety value, as the number of iterations of the sparrow search algorithm increases, the global search ability of the discoverers weakens, and the performance of their position update strategy is not as good as that of the bird flock algorithm. Substituting equation (4) into the position update formula of the discoverer in the sparrow search algorithm:

[0192]

[0193] In the formula: R2 is the warning value, R2∈[0,1]; ST is the safety value, ST∈[0.5,1]; Q is a random number from a standard normal distribution; L is a 1×D matrix with all elements being 1.

[0194] Meanwhile, substituting equation (5) into the follower position update formula in the sparrow search algorithm:

[0195]

[0196] Let be the position of the sparrow with the worst fitness value at iteration number t; k represents the discoverer and k ≠ i. Furthermore, the scout's position update formula is:

[0197]

[0198] In the formula: β is a Gaussian random number with expected value of 0 and variance of 1; K∈[-1,1] is a random number; f i f represents the current fitness level of the sparrow; g The current optimal fitness; Let f be the position of the sparrow with the highest fitness value at iteration number t. Among these, sparrows exhibiting scouting behavior detect approaching danger. i ≠f g The sparrow was in a relatively dangerous area on the edge, and needed to be moved to a safer area. i =f g This indicates that the sparrow is in the optimal position X. bj However, sparrows need to move closer to their companions to reduce the danger. Here, β and K are parameters that control the step size, with the aim of getting the sparrow to a safe position.

[0199] During the algorithm's iteration process, if the fitness value of the optimal solution at the (t+1)th generation is lower than that at the tth generation, then the optimal position of the SSA algorithm is not updated; otherwise, the optimal position and its corresponding value are updated.

[0200] The Gaussian walk in sub-step 5.3 is as follows:

[0201] When the sparrow search algorithm converges and its iteration stalls during the search process, a Gaussian random walk strategy is introduced to avoid falling into local optima and ensure that the algorithm finds the global optimum. A cosine function is used to reduce perturbation, thereby enhancing the algorithm's stability and search capability. The new sparrow position update is described as follows:

[0202]

[0203]

[0204] In the formula: σ is the variance. Let t be the position of the sparrow. Let t+1 be the position of the sparrow. For the random discoverer sparrow; t max The maximum number of iterations; Gaussian is the Gaussian distribution function.

[0205] Step 6: The drone dynamically tracks the mission target to be executed according to the initial allocation plan;

[0206] Step 7: At the beginning of the first time slice, obtain information on drone failure and whether the target has changed;

[0207] Step 8: Within the time slice, randomly generate a set of allocation schemes based on the changed information, and repeat step 5 to find the optimal allocation scheme after the information changes.

[0208] Step 9, in the next time slice or at t+t s At time t, repeat steps 7 and 8 until t > t min At the end of the time slice, based on the optimal solution in step 8, the drone's position information is updated using the tracking method, and the optimal allocation result is output.

[0209] The effectiveness of this invention can be further demonstrated through the following simulation experiments.

[0210] The simulation was performed using MATLAB 2018b on a Windows 10 computer with an Intel(R) Core(TM) i5-7300HQ CPU@2.50GHz.

[0211] To verify the effectiveness of the improved SSA algorithm, the experiment was run 100 times to obtain the optimal allocation scheme, algorithm time, and average optimal function value.

[0212] Simulation results:

[0213]

[0214] The improved SSA algorithm takes 0.2241 seconds, and the optimal function value is 0.7421. The first row of matrix A represents the optimal allocation scheme, and the second row represents the order of the attack targets. The specific allocation is as follows: Figure 5 As shown.

[0215] from Figure 6 It can be seen that as the number of iterations increases, the optimal function value gradually stabilizes. After the 30th iteration, the optimal function value basically no longer changes, and its value is 0.7421. The average function value also tends to stabilize as the number of iterations increases. Although there are slight fluctuations, the overall trend no longer changes.

[0216] To verify the solution accuracy and search speed of the improved SSA algorithm, simulation experiments on task allocation based on SSA, Genetic Algorithm (GA), Ant Colony Algorithm (ACO), and Particle Swarm Optimization (PSO) were compared, and the simulation time and optimal function values ​​were analyzed and compared. Each algorithm was run 100 times for different task allocation scales to obtain simulation results.

[0217] Table 3 Comparison of Simulation Results Based on Different Algorithms

[0218]

[0219] from Figure 7Analysis reveals that, with minimal changes in the allocation scheme, the optimal function values ​​of the five algorithms gradually stabilize with increasing iterations, differing only in their solution accuracy and speed. Specifically, the SSA, ACO, and PSO algorithms converged prematurely due to getting trapped in local optima, failing to reach optimal fitness. While the GA algorithm ultimately achieved a relatively low optimal fitness value, its solution accuracy was too slow, resulting in overall low efficiency. In comparison, the improved SSA algorithm demonstrated superior performance, exhibiting not only shorter simulation time but also higher simulation accuracy compared to the traditional intelligent algorithms. The comparison of the five algorithms fully demonstrates the superior performance of the improved SSA algorithm.

[0220] This invention presents a task allocation method for multi-aircraft formation cooperative operations based on an improved Sparrow Search Algorithm, addressing the dynamic task allocation problem in multi-aircraft formations cooperating to perform ground attack missions. This method improves upon the Sparrow Search Algorithm (SSA) by introducing a Logistic mapping from a chaotic optimization algorithm and replacing its position update with a Bird Swarm Algorithm (BSA) update formula similar to SSA. A Gaussian walk strategy is employed to complete the improvement of the SSA algorithm. An allocation code and fitness function are designed. Simulation experiments on task allocation for six UAVs carrying two missiles attacking eight targets verify the effectiveness of the improved SSA algorithm. Comparison with task allocation simulation experiments based on SSA, GA, ACO, and PSO algorithms confirms the effectiveness and superiority of the improved SSA algorithm, ensuring both optimal allocation and search speed. This invention outperforms general algorithms such as SSA, GA, ACO, and PSO in task allocation for multi-aircraft formation cooperative attacks on multiple targets, demonstrating significant application value.

[0221] Based on traditional task allocation algorithms, this invention proposes a dynamic task allocation method based on the improved Sparrow Search Algorithm (SSA). Because the SSA algorithm is prone to getting trapped in local optima, this method addresses this deficiency by using chaotic optimization operators, position updates from the BSA algorithm, and Gaussian walks to improve the SSA algorithm. The target task sequence encoding and fitness function are designed, and a dynamic task allocation process based on the improved SSA algorithm is presented.

[0222] Those skilled in the art will understand that, besides implementing the system and its various devices, modules, and units provided by this invention in the form of purely computer-readable program code, the same functions can be achieved entirely through logical programming of the method steps, making the system and its various devices, modules, and units of this invention function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by this invention can be considered as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices, modules, and units for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0223] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for task allocation of multi-UAV formation cooperative combat, characterized in that, The method comprises the following steps: An information initialization step: initializing information of formation combat members and targets; An algorithm parameter initialization step: initializing parameters of the sparrow search algorithm to obtain an initialized sparrow population; An initial scheme acquisition step: initializing the sparrow population by using a Logistic chaotic mapping, improving a fitness function of the sparrow search algorithm according to a sparrow fitness function index, and obtaining an initial allocation scheme based on an allocation scheme code according to the initialized information of the formation combat members and targets; An initial optimal scheme acquisition step: calculating an initial optimal allocation scheme according to the initial allocation scheme and the improved sparrow search algorithm; A dynamic tracking step: tracking a task target to be executed by a UAV according to the initial optimal allocation scheme; The initial optimal scheme acquisition step comprises the following steps: Sparrow population initialization step: initialize the sparrow population, randomly generate N sparrows, initialize the number of discoverers P num , the number of followers N-P num , and the number of scouts S num ; initialize the value of D; initialize the values of the upper bound lb and the lower bound ub of the search interval; Fitness calculation step: the fitness value of each sparrow is calculated by using the fitness function in the initial scheme acquisition step, and P num Only sparrow as discoverer, the rest as follower, and S is randomly selected num Only scout; A local optimal judgment step: judging whether the algorithm falls into a local optimum; when the sparrow search algorithm is not stagnant, updating a position of a sparrow individual by using a position updating formula of a bird swarm algorithm; when the algorithm is stagnant, updating a position of the sparrow swarm by using a Gaussian walk; A boundary control step: controlling a boundary of the sparrow, processing a sparrow out of the boundary, and obtaining a new position and fitness of the sparrow swarm; Finding step: if Gen = Gen max then the algorithm ends, where Gen represents the number of iterations; Gen max is the maximum number of iterations set. An optimal scheme seeking step: seeking an optimal scheme by analyzing fitness values output by the improved sparrow search algorithm under different schemes; If Gen < Gen max then go to fitness calculation step.

2. The method of claim 1, wherein, The method further comprises the following steps: Define time slice length step: define task time slice interval t s ; A change judgment step: obtaining information about whether a UAV failure and a target change at the beginning of a first time slice; The optimal scheme seeking step: generating an allocation scheme randomly according to the changed information, repeating the initial optimal scheme acquisition step, and seeking an optimal allocation scheme after the information changes; Result output step: at the next time slice or at the time t+t s , repeat the change judgment step and the optimal solution seeking step until t>t min ; where t represents the time, t min represents the estimated minimum mission flight time; At the end of the time slice, updating position information of the UAV by using a tracking method according to the optimal allocation scheme of the optimal scheme seeking step, and outputting an optimal allocation result.

3. The method of claim 1, wherein, In the information initialization step, aircraft platform information, weapon information, and a damage probability to enemy task targets are obtained before the allocation starts; information about the enemy task targets is detected, including target positions, quantities, types, and maneuverability, and a threat probability of the enemy targets to the members is calculated.

4. The method of claim 1, wherein, In the initialization algorithm parameter step, the basic parameters of the sparrow search algorithm are set, the number of sparrow swarm individuals is N, the number of discoverers is P num , the number of scouts is S num , the problem to be optimized is D-dimensional, the maximum number of iterations is Gen max , the lower limit of the search interval is lb; the upper limit of the search interval is ub; the sparrow swarm, the discoverer and the scout are all internal parameters of the sparrow search algorithm.

5. The method of claim 1, wherein, In the initial scheme acquisition step, the Logistic mapping is introduced to initialize the sparrow swarm: x n+1 = μx n (1 - x n ), x n ∈(0,1)·····(1) where: μ is a chaos control parameter; x n is the value of the nth iteration, x n+1 is the value of the n+1 iteration; The sparrow swarm has M individuals, and the dimension is D; Chaotic mapping initial values x1 are randomly generated; The remaining M-1 sparrows in the swarm under the Logistic mapping are obtained by substituting x1 into formula (1); A variable value of the Logistic mapping is converted into an actual value X of the sparrow swarm, and the conversion formula is: X i = X lb + (X lb - X ub ) x (x i + 1) x 0.5 (2) where x i is the value of the Logistic map at the i-th iteration; X lb is the lower bound of the sparrow search; X ub is the upper bound of the sparrow search; X i is the actual value of the sparrow swarm at the i-th iteration.

6. The method of claim 1, wherein, In the initial scheme acquisition step, the fitness function is calculated as: The fitness of the whole swarm is expressed as: In the formula, F X is the population fitness matrix, and f is the fitness value of a single sparrow; X is the sparrow population; X N,D represents the Nth sparrow, and the Dth dimension target function.

7. The method of claim 1, wherein, In the local optimal judgment step, the bird swarm algorithm is: Positions of the discoverer and the follower in the bird swarm are described as: where t represents the current iteration number; j represents the jth dimension of the objective function; represents the jth dimension of the ith sparrow at iteration t; represents the jth dimension of the ith sparrow at iteration t+1; randN(0,1) is a Gaussian distributed random number with expectation 0 and variance 1; FL is the probability of the follower finding high-energy food; k represents the finder; represents the jth dimension of the kth finder at iteration t; rand(0,1) represents a random number generated within (0,1); When the early warning value is less than the safety value, formula (4) is substituted into a discoverer position updating formula in the sparrow search algorithm: In the formula, R2 is the early warning value, ST is the safety value, Q is a random number of a standard normal distribution, and L is a 1×D matrix with elements all being 1; Meanwhile, formula (5) is substituted into a follower position updating formula in the sparrow search algorithm: wherein, is the worst sparrow position for the fitness value at iteration t; k represents the discoverer; In addition, a position updating formula of the scout is: wherein: β is a Gaussian distribution random number with expectation of 0 and variance of 1; K ∈ [-1, 1] is a random number; f i is the fitness of the current sparrow; f g is the current optimal fitness; is the sparrow position with the highest fitness value at the iteration number t; wherein, when the sparrow with the scout behavior finds that danger is coming, f i ≠ f g , the sparrow is in the marginal dangerous area, and the sparrow is transferred to a safe area; f i = f g indicates that the sparrow is in the optimal position X bj , and the sparrow is close to the companion; β and K are both parameters for controlling the step length, and the sparrow is updated to a safe position; In the iteration process of the algorithm, if the optimal solution fitness value of the t+1 generation is lower than the t generation, the optimal position of the sparrow search algorithm is not updated, if the optimal solution fitness value of the t+1 generation is higher than or equal to the t generation, the optimal position and the corresponding value are updated.

8. A multi-UAV formation cooperative combat task allocation system, characterized in that, The task allocation method for multi-aircraft formation cooperative combat according to any one of claims 1-7 comprises the following modules: An information initialization module for initializing information of formation combat members and targets; An initialization algorithm parameter module for initializing sparrow search algorithm parameters to obtain an initialized sparrow population; An initial scheme acquisition module for initializing the sparrow population by using a Logistic chaotic mapping, improving a sparrow search algorithm fitness function according to a sparrow fitness function index, and obtaining an initial allocation scheme based on allocation scheme coding according to the initialized information of formation combat members and targets; An initial optimal scheme acquisition module for calculating an initial optimal allocation scheme according to the initial allocation scheme and the improved sparrow search algorithm; A dynamic tracking module for dynamically tracking a task target to be executed by the unmanned aerial vehicle according to the initial optimal allocation scheme.

9. The multi-UAV formation cooperative engagement task allocation system of claim 8, wherein, The system further comprises the following modules: Define time slice length module: define task time slice interval t s ; A change judgment module for obtaining unmanned aerial vehicle failure information and whether the target has changed at the beginning of the first time slice; A best scheme seeking module for randomly generating an allocation scheme according to the changed information, repeating the initial optimal scheme acquisition module, and seeking an optimal allocation scheme after the information changes within the time slice; Result output module: at the next time slice or at time t+t s , repeat the change judgment module and the optimal solution module until t>t min ; wherein t represents the time, t min represents the estimated minimum task flight time; At the end of the time slice, the position information of the unmanned aerial vehicle is updated by using a tracking method according to the optimal allocation scheme of the best scheme seeking module, and an optimal allocation result is output.

Images