A method and device for multi-UAV collaborative target allocation

Abstract

The present application provides a method and device for cooperative target allocation of multiple UAVs, which relates to the field of automatic control technology. In the process of solving any round of target allocation, the situation evaluation function of one UAV relative to the other UAV is determined according to the current situation information of the two UAVs; the attack benefit values ​​of the two UAVs under all strategy combinations in the strategy space are determined to form the payment matrix of the two UAVs; the expected benefits and expected total benefits of the two UAVs under each strategy are determined according to the strategy space, probability space and payment matrix to form a set of replicated dynamic equations; the evolutionary stable strategies of the two UAVs are solved; the UAVs are controlled to fight according to the attack targets assigned in the evolutionary stable strategies, and the situation information is updated; when the predetermined combat conditions are met, the next round of target allocation is solved. In this way, the dynamic interaction of the two UAVs can be predicted, the changes in the battlefield situation can be adapted, and a more global and reasonable target allocation strategy can be obtained.

Description

Technical Field

[0001] The present application relates to the field of automatic control technology, and in particular to a method and device for multi-UAV collaborative target allocation. Background Art

[0002] Coordinated target allocation for multiple UAVs is one of the hot issues in the current UAV research field, especially in military, surveillance, search and rescue applications. In the process of performing air combat missions, the purpose of coordinated target allocation for multiple UAVs in air combat is to accurately evaluate multiple related factors such as air combat situation, fighter performance and tactical intentions, and to use currently available information and resources to adopt scientific and effective strategies to optimize the allocation of our UAV resources, maximize the damage to enemy UAVs while minimizing the loss of our UAVs, so that our UAVs are in an advantageous position during air combat.

[0003] At present, the commonly used methods for the intelligent agent target allocation problem are optimization methods and heuristic methods, but the above methods all have their own limitations. Specifically, although the heuristic method has good adaptive ability, the traditional genetic algorithm and particle swarm optimization algorithm have the defect of easily falling into the local optimum; the optimization method has the defect that only by maximizing or minimizing the objective function under given constraints can the global optimal solution or the local optimal solution under certain constraints be found under ideal conditions. On the other hand, the existing technology mostly establishes combat models or functions based on fixed situations, which belongs to quasi-static one-time selection, and has weak response ability to the dynamic evolution of strategies over time, especially in complex interactive scenarios between individual behaviors and group patterns. Summary of the invention

[0004] In view of this, the purpose of this application is to provide a method and device for collaborative target allocation of multiple UAVs. Based on the evolutionary game theory, a replicated dynamic equation group of the UAVs of both sides is constructed to describe the evolution process of the strategy distribution over time. This can fully consider the strategies that the enemy may take against its own actions, and then predict the dynamic interaction between the UAVs of both sides, adapt to changes in the battlefield situation, and have a stronger ability to bypass the "local optimal trap", so as to obtain a more global and reasonable target allocation strategy for collaborative combat of multiple UAVs.

[0005] The present application provides a method for allocating coordinated targets among multiple UAVs, the method comprising:

[0006] Construct the strategy space of the evolutionary game between the two drones and the probability space of the two drones choosing each strategy;

[0007] In the process of solving any round of target allocation, a situation evaluation function of one UAV relative to the other UAV is determined according to the current situation information of the two UAVs; wherein the situation evaluation function is used to evaluate the attack effectiveness of one UAV on the other UAV; the situation evaluation function is obtained by weighting at least one combat advantage function of one UAV relative to the other UAV;

[0008] According to the situation evaluation function of one party's UAV relative to the other party's UAV, the attack benefit values ​​of both parties' UAVs under all strategy combinations in the strategy space are determined to form the payment matrix of both parties' UAVs; wherein, the attack benefit value of any party's UAV is defined by the ratio between the attack effectiveness of its own UAV on the enemy UAV and the attack effectiveness of the enemy UAV on its own UAV;

[0009] According to the strategy space of the evolutionary game between the two drones, the probability space of selecting each strategy and the payment matrix of the two drones, the expected benefits of the two drones under each strategy and the expected total benefits of the two drones are determined;

[0010] According to the probability space of the two drones selecting each strategy, the expected benefits of the two drones under each strategy combination, and the expected total benefits of the two drones, all strategies in the strategy space are traversed to form a replication dynamic equation group of the two drones; wherein the replication dynamic equation group is used to describe the evolution process of the strategy distribution of the two drones over time;

[0011] Solving the replication dynamic equations of both drones to obtain the evolutionary stable strategies of both drones; wherein the benefits of both drones reach a balanced state under the evolutionary stable strategies;

[0012] The UAVs of both sides are controlled to fight according to the attack targets assigned in the evolutionary stability strategy, and the situation information is updated; until the UAVs of both sides meet the predetermined combat conditions, the next round of target allocation is solved.

[0013] The embodiment of the present application also provides a multi-UAV cooperative target allocation device, the device comprising:

[0014] A construction module is used to construct the strategy space of the evolutionary game between the two drones and the probability space of the two drones choosing each strategy;

[0015] The first determination module is used to determine the situation evaluation function of one party's UAV relative to the other party's UAV according to the current situation information of the two parties' UAVs during the solution process of any round of target allocation; wherein the situation evaluation function is used to evaluate the attack effectiveness of one party's UAV on the other party's UAV; the situation evaluation function is obtained by weighting at least one combat advantage function of one party's UAV relative to the other party's UAV;

[0016] The second determination module is used to determine the attack benefit values ​​of the drones of both parties under all strategy combinations in the strategy space according to the situation evaluation function of the drone of one party relative to the drone of the other party, so as to form a payment matrix of the drones of both parties; wherein the attack benefit value of the drone of any party is defined by the ratio between the attack effectiveness of the drone of the own party against the drone of the enemy and the attack effectiveness of the drone of the enemy against the drone of the own party;

[0017] The third determination module is used to determine the expected benefits of both drones under each strategy combination and the expected total benefits of both drones according to the strategy space of the evolutionary game of both drones, the probability space of selecting each strategy, and the payment matrix of both drones;

[0018] The fourth determination module is used to traverse all strategies in the strategy space according to the probability space of the two drones selecting each strategy, the expected benefits of the two drones under each strategy, and the expected total benefits of the two drones, to form a replication dynamic equation group of the two drones; wherein the replication dynamic equation group is used to describe the evolution process of the strategy distribution of the two drones over time;

[0019] A solution module is used to solve the replication dynamic equations of the two UAVs to obtain the evolutionary stable strategy of the two UAVs; wherein the benefits of the two UAVs reach a balanced state under the evolutionary stable strategy;

[0020] The control module is used to control the drones of both sides to fight according to the attack targets assigned in the evolutionary stability strategy and update the situation information; until the drones of both sides meet the predetermined combat conditions, the first determination module is called to solve the next round of target allocation.

[0021] An embodiment of the present application also provides an electronic device, including: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, the steps of a multi-UAV collaborative target allocation method as described above are performed.

[0022] An embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the multi-UAV collaborative target allocation method as described above are executed.

[0023] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0025] Figure 1 One of the flow charts of a multi-UAV cooperative target allocation method provided in an embodiment of the present application is shown;

[0026] Figure 2 A schematic diagram showing the evolution trend of a red team drone strategy provided by an embodiment of the present application is shown;

[0027] Figure 3 A schematic diagram showing the motion trajectory of two drones provided in an embodiment of the present application is shown;

[0028] Figure 4 A schematic diagram showing a target change trend of a red team UAV provided in an embodiment of the present application is shown;

[0029] Figure 5 A second flowchart of a multi-UAV cooperative target allocation method provided in an embodiment of the present application is shown;

[0030] Figure 6 A schematic diagram of the structure of a multi-UAV cooperative target allocation device provided in an embodiment of the present application is shown;

[0031] Figure 7 A schematic structural diagram of an electronic device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0032] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application usually described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, each other embodiment obtained by those skilled in the art without making creative work belongs to the scope of protection of the present application.

[0033] The research found that multi-UAV coordinated target allocation is one of the hot issues in the current UAV research field, especially in military, monitoring, search and rescue applications. In the process of performing air combat missions, the purpose of multi-UAV air combat coordinated target allocation is to accurately evaluate multiple related factors such as air combat situation, fighter performance and tactical intentions, and use the currently available information and resources to adopt scientific and effective strategies to optimize our UAV resource allocation, maximize the damage to enemy UAVs while minimizing the loss of our UAVs, so that our UAVs are in an advantageous position during air combat.

[0034] Multi-Agent: refers to a collection of multiple agents. By studying the architecture and behavior theory of the collection, research and development are conducted on the interactive communication and collaboration capabilities of the agents in the collection, and a mutually collaborative integrated system is formed. The drone cluster in the embodiment of this application is a typical multi-agent application.

[0035] At present, the commonly used methods for the intelligent agent target allocation problem are optimization methods and heuristic methods, but the above methods all have their own limitations. Specifically, although the heuristic method has good adaptive ability, the traditional genetic algorithm and particle swarm optimization algorithm have the defect of easily falling into the local optimum; the optimization method has the defect that only by maximizing or minimizing the objective function under given constraints can the global optimal solution or the local optimal solution under certain constraints be found under ideal conditions. On the other hand, the existing technology mostly establishes combat models or functions based on fixed situations, which belongs to quasi-static one-time selection, and has weak response ability to the dynamic evolution of strategies over time, especially in complex interactive scenarios between individual behaviors and group patterns.

[0036] Based on this, an embodiment of the present application provides a multi-UAV collaborative target allocation method and device, which can obtain a more global and reasonable target allocation strategy for multi-UAV collaborative combat based on evolutionary game theory.

[0037] Among them, Evolutionary Game Theory is a branch of game theory that combines traditional game theory with evolutionary theory in biology. It is used to analyze how individuals continuously optimize their strategy choices through the evolutionary process in repeated games, thereby adapting to the environment and increasing their probability of survival or success.

[0038] See also Figure 1 , Figure 1This is one of the flow charts of a multi-UAV cooperative target allocation method provided by an embodiment of the present application. First, the application scenario of the embodiment of the present application is introduced. The embodiment of the present application is applied to the scenario of cooperative combat of the drone clusters of both parties, hereinafter referred to as the red drone and the blue drone, and each drone cluster includes one or more drone units. The method can be applied to the target allocation of the drones of either party. Preferably, in order to ensure the authenticity of the combat scenario, the method can be applied to the target allocation of the drones of both parties at the same time.

[0039] To facilitate understanding, the concepts defined in the embodiments of the present application are first introduced below.

[0040] 1. Situation Assessment: Situation assessment in air combat refers to the perception and assessment of elements in the air combat environment. Correctly judging the enemy and friendly situation is an important part of air combat decision-making. The embodiment of the present application can construct angle advantage function, speed advantage function, height advantage function and distance advantage function for situation assessment. The following is an example of the red side drone.

[0041] 1. Angle advantage function

[0042] In drone air combat, angle advantage is a key tactical factor, which describes the position and posture advantage of a drone relative to an enemy aircraft when it is in air combat. A drone with angle advantage can more effectively lock onto a target, evade the actions of an enemy aircraft, and take effective attack or defense measures. Angle advantage function S α The expression is as follows:

[0043]

[0044] in, is the angle between the distance vector from the red drone to the blue drone and the velocity vector of the red drone; It is the angle between the distance vector from the blue drone to the red drone and the speed vector of the blue drone. and The range of is [0,π]. When the red drone is directly behind the blue drone, that is, , its angle advantage is the maximum value 1. On the contrary, when the blue drone is directly behind the red drone, that is, When , its angular advantage is the minimum value 0.

[0045] 2. Speed ​​advantage function

[0046] The purpose of the speed advantage function is to quantify the speed advantage of a drone relative to its target. In an air battle, the speed advantage can enable the drone to occupy an advantageous position more quickly to evade enemy attacks or quickly approach the target for attack. Taking the red drone as an example, its speed advantage function S VThe expression is as follows:

[0047]

[0048] Among them, V c is the relative speed of the red drone relative to the blue drone, in m / s. V With V c The maximum value is close to 1 and the minimum value is 0.

[0049] 3. Height advantage function

[0050] The height advantage is mainly related to the height difference between the two drones, which has an important impact on the attack angle and escape strategy in air combat. x is the optimal height difference between the two drones, σ is the optimal height standard deviation. Taking the red drone as an example, the expression of its height advantage function S is as follows:

[0051]

[0052] Where is the height difference between the red drone and the blue drone. When the height difference is in the interval [ x , x +σ], S is the largest, with a maximum value of 1. When the height difference between the two sides is too large, S approaches 0.

[0053] 4. Distance advantage function

[0054] The hit rate of drone weapons is mainly related to the distance between the two drones. The distance advantage indicates the degree of advantage of the blue drone within the attack range of the red drone weapons. Taking the red drone as an example, its distance advantage function S r As shown below:

[0055]

[0056] in, and are the maximum and minimum attack radius of the Red Army’s drone weapons; σ r is the weapon attack distance range, With σ r When the distance between the two drones is between the maximum and minimum attack radius of the weapon, S r Maximum, the maximum value is 1. When the distance between the two drones is too large, S r Approaching 0.

[0057] 5. Situation Assessment Function

[0058] The situation assessment function is used to evaluate the attack effectiveness of the UAV against its target, and can be obtained by selecting one or more of the above-mentioned advantage functions and assigning weights, and these weights are set by experts with rich air combat experience according to the actual influence and importance of each factor. The situation assessment function can provide a quantitative index to reflect the potential success rate of the UAV in performing combat tasks in the current environment. Exemplarily, the situation assessment function S is as follows:

[0059] S = ω α S α + ω V S V + ωS + ω r S r

[0060] Where ω α 、ω V 、ω and ω r are the weights of the angle advantage function, speed advantage function, altitude advantage function and distance advantage function respectively, and ω α + ω V + ω + ω r = 1. The weights of each advantage function are different under different air combat situations. When r ≥ 2000, ω α = 0.2, ω V = 0.3, ω = 0.2, ω r = 0.3; when r < 2000, ω α = 0.4, ω V = 0.2, ω = 0.2, ω r = 0.2.

[0061] II. Target allocation:

[0062] Denote the multi-UAV sets of the red and blue sides as R and B respectively, and construct the red-side UAV target allocation matrix

[0063]

[0064] In a real air combat scenario, there will be differences in the number of UAVs on both sides, which will affect the UAV target allocation results. Assume that the number of red-side UAVs is m and the number of blue-side UAVs is n. For the red-side UAVs, if m ≥ n, it is necessary to ensure that each blue-side UAV is selected as a target by the red-side UAVs; if m < n, it is necessary to ensure that the blue-side UAVs are not selected repeatedly:

[0065]

[0066] III. Other definitions:

[0067] Define the state information of the drone ξ=(x,y,z,v,p,wp), where (x,y,z) represents the location information of the drone, v represents the speed of the drone, p represents the survival probability of the drone, and wp represents the number of remaining weapons of the drone. The state transition of the i-th drone of the red side at time k is:

[0068]

[0069] in, and They represent the speed, position information, survival probability, number of remaining weapons, number of consumed weapons and acceleration of the drone at time k. and represents the speed, three-dimensional coordinates, survival probability and number of remaining weapons of the drone at time k+1. The speed of the red drone follows the following constraints:

[0070]

[0071] in, Represents the maximum speed limit of the Red Team’s drone. Represents the minimum speed limit of the Red Team UAV. is the damage degree of the i-th drone of the red side after being attacked by the drone of the blue side during the air battle, and is defined as Less than the survival probability P T The number of drones on both the red and blue sides will change dynamically as the drones are destroyed in battle.

[0072]

[0073] in, They are the number of weapons launched and the probability of damage by the j-th UAV of the blue side against the i-th UAV of the red side at the k-th moment.

[0074]

[0075] Among them, β w is the environmental factor, is the situation advantage value of the j-th UAV of the blue side over the i-th UAV of the red side at time k obtained by the air combat situation evaluation function, K B It is the ideal attack effectiveness of the blue team's drone weapons.

[0076] When the Red Team's i-th UAV meets certain conditions relative to the Blue Team's j-th UAV's assigned target, it launches weapons to attack it, as shown in the following formula:

[0077]

[0078] in, represents the relative distance between the i-th UAV of the red team and the j-th UAV of the blue team at time k, S T is the situational advantage threshold for attacking enemy aircraft, is the situation advantage value of the red side’s i-th UAV over the blue side’s j-th UAV at time k obtained through the air combat situation evaluation function.

[0079] like Figure 1 As shown in , the target allocation method provided by the embodiment of the present application includes:

[0080] S101. Construct the strategy space of the evolutionary game between the two drones and the probability space of the two drones choosing each strategy.

[0081] Aiming at the problem of coordinated target allocation in multi-UAV air combat, an evolutionary game model G = {N, S, U, ∏} is constructed to analyze the strategic interaction between the two UAVs, as follows:

[0082] N = {R, B} is the participants of the evolutionary game. R represents the set of red drones, B represents the set of blue drones, the number of red drones is m, and the number of blue drones is n.

[0083] ∏={∏ R ,∏ B} is the strategy space of the evolutionary game between the two drones, where is the red side’s pure strategy set, is the blue side’s pure strategy set, α R and α B are the number of pure strategies for the red and blue sides respectively.

[0084] U={U R ,U B} is the payoff matrix of the evolutionary game, where is the red side's payment matrix, is the blue side’s payment matrix. and Choose pure strategies for the red team respectively And Blue chooses a pure strategy The specific determination of each element in the payoff matrix of the evolutionary game will be introduced in the following steps.

[0085] I={I R ,I B} is the probability space for both parties in the evolutionary game to choose each strategy, is the red square probability space, is the blue square probability space. Among them, Choose a pure strategy for Red The probability of Choose a pure strategy for Blue probability.

[0086] S102. In the process of solving any round of target allocation, determine the situation evaluation function of one UAV relative to the other UAV based on the current situation information of both UAVs.

[0087] The situation assessment function is used to evaluate the effectiveness of the attack of one party's UAV on the other party's UAV. The situation assessment function is obtained by weighting at least one combat advantage function of one party's UAV relative to the other party's UAV. Multiple combat advantage functions can comprehensively measure the combat advantage of one party's UAV relative to the other party's UAV from different dimensions. Before the first round of solving, the parameters of the UAVs of both parties are initialized first, and then the functional form of the situation assessment function of one party's UAV relative to the other party's UAV is defined.

[0088] In a possible implementation, step S102 may include: determining a combat advantage function of one UAV relative to another UAV based on current situation information of both UAVs; wherein the combat advantage function includes at least one of the following: an angle advantage function, a speed advantage function, a height advantage function, and a distance advantage function; and obtaining a situation assessment function of one UAV relative to another UAV by weighting the combat advantage functions and the corresponding preset weight values.

[0089] S103. According to the situation evaluation function of one drone relative to the other drone, the attack benefit values ​​of the two drones under all strategy combinations in the strategy space are determined to form the payment matrix of the two drones.

[0090] Among them, the attack benefit value of any party's drone is defined by the ratio between the attack effectiveness of the own drone on the enemy drone and the attack effectiveness of the enemy drone on the own drone.

[0091] In a possible implementation, step S103 may include:

[0092] S1031. Determine, based on the situation assessment function of one party's drone relative to the other party's drone, the situation assessment value when each drone unit in one party's drone selects each drone unit in the other party's drone as an attack target.

[0093] In this step, according to the expression of the combat advantage function defined above, when the red side's i-th drone selects the blue side's j-th drone as a target, the specific situation information of the red side's i-th drone and the blue side's j-th drone is substituted into the expression of the combat advantage function to obtain the situation advantage value when the red side's i-th drone selects the blue side's j-th drone as a target. Similarly, we can get the situational advantage value when the j-th UAV of the blue team chooses the i-th UAV of the red team as the target:

[0094] S1032. For each strategy, determine the total attack effectiveness of one party's drone on another party's drone based on the target allocation matrix of the party's drone under the strategy and the situation assessment value.

[0095] Among them, the element value of each element in the target allocation matrix is ​​used to represent whether a drone unit in one party's drone selects a drone unit in the other party's drone as an attack target.

[0096] In this step, the following formula can be used to determine The total attack effectiveness of the red drone against the blue drone in the strategy

[0097]

[0098] in, For strategy The target allocation matrix.

[0099] Similarly, the following formula can be used to determine the strategy The total attack effectiveness of the blue drone against the red drone

[0100]

[0101] in, For strategy The target allocation matrix.

[0102] S1033. Determine the average attack effectiveness of one party's drones under this strategy based on the total attack effectiveness of one party's drones on the other party's drones and the number of drones on one party.

[0103] S1034. For each strategy combination, the ratio of the average attack effectiveness of the drones of one party under this strategy combination to the average attack effectiveness of the drones of the other party is determined as the attack benefit value of the drones of one party under this strategy combination, thereby forming a payment matrix of the drones of one party.

[0104] For the above steps S1033 and S1034, in the strategy combination The red side's attack benefit value is is the ratio of the average attack effectiveness of each drone of the red side to the average attack effectiveness of each drone of the blue side, and the attack benefit value of the blue side The calculation is similar, and it can be expressed as follows:

[0105]

[0106] Traverse the red and blue payoff values ​​under all strategy combinations to construct the payment matrix, as shown below:

[0107]

[0108] Among them, U R represents the payoff matrix of the Red Team’s UAV; U B Represents the payoff matrix of the blue drone.

[0109] S104. According to the strategy space of the evolutionary game between the two drones, the probability space of selecting each strategy, and the payment matrix of the two drones, the expected benefits of the two drones under each strategy and the expected total benefits of the two drones are determined.

[0110] In this step, the expected benefits of both drones under each strategy and the expected total benefits of both drones can be determined by the following formula:

[0111]

[0112] in, The strategies adopted by the Red team are and Blue adopts strategy The expected return, and are the expected total returns of the red party and the blue party respectively; Choose a pure strategy for Red The probability of Choose a pure strategy for Blue probability.

[0113] S105. According to the probability space of the two drones selecting each strategy, the expected benefits of the two drones under each strategy combination, and the expected total benefits of the two drones, all strategies in the strategy space are traversed to form a set of replication dynamic equations for the two drones.

[0114] It should be noted that in evolutionary game theory, when game participants face differences in interests, the party with lower benefits tends to observe and imitate the party with higher benefits, and gradually optimize its own strategy choices through this learning process. Over time, this strategy learning and adjustment leads to dynamic changes in the distribution probability of different strategies in the population. In order to accurately describe and analyze the process of this strategy change, the evolutionary stable strategy can be solved by the replication dynamic equation based on biological evolution theory. The replication dynamic equation is a powerful tool for analyzing and simulating how strategies evolve over time and how to reach the equilibrium state of evolutionary games. It is suitable for studying complex game scenarios such as multi-UAV air combat target allocation.

[0115] In an embodiment of the present application, the replicated dynamic equations are used to describe the evolution of the strategy distribution of both drones over time.

[0116] In the specific implementation, for any strategy, construct the red and blue replication dynamic equation f R (p) and f B (q) is used to represent the change of the red and blue strategy distribution over time. The formula is as follows:

[0117]

[0118] Traversing all strategies in the strategy space, the replication dynamic equations formed to describe the entire evolution process of both drones can be expressed as:

[0119]

[0120] S106. Solve the replication dynamic equations of the two UAVs to obtain the evolutionary stability strategies of the two UAVs.

[0121] In this step, the replication dynamic equations can be solved using the related solution methods in the prior art, and this application does not make any limitation here. The evolutionary stable strategy obtained by the solution should satisfy the equilibrium state of the benefits of both drones under the evolutionary stable strategy, that is, the probability of selecting each strategy of the drones of both parties will remain unchanged, and any unilateral change of the strategy of the drone will only lead to a decrease in its relative benefits, so there is no motivation to deviate from this equilibrium state.

[0122] S107, controlling the UAVs of both parties to fight according to the attack targets assigned in the evolutionary stability strategy, and updating the situation information; until the UAVs of both parties meet the predetermined combat conditions, solving the next round of target allocation.

[0123] In this step, the evolutionary stable strategy obtained sets the strategy that each drone should choose. According to this strategy, each drone determines the attack target assigned to it. After that, each drone can be controlled to fight according to the assigned attack target. During the combat, the drone will fly towards its designated target, fight according to the corresponding combat strategy, and update the situation information in real time.

[0124] Afterwards, when the drones of both sides meet the predetermined combat conditions, the next round of target allocation is solved. Specifically, when a drone unit in any drone of either side shoots down its corresponding attack target, it is determined whether the drones of both sides meet the combat termination conditions.

[0125] Here, the battle termination conditions include that all drone units in any one of the drones are shot down, or the battle between the two drones reaches a preset number of rounds; if not met, the next round of target allocation is solved; if met, the battle between the two drones is ended and the final air combat result is output.

[0126] See also Figure 2 , Figure 2 A schematic diagram of the evolution trend of the Red Team UAV strategy provided by the embodiment of the present application. In an experiment, Figure 2 As shown, the Red Team UAV chose Strategy 5: As a target allocation strategy, this means that in the process of strategy evolution, Strategy 5 provides higher benefits to the Red UAV than other strategies. Therefore, the Red UAV continuously adjusts and learns itself, adjusts its strategy probability distribution, and finally converges to Strategy 5. In this process, the movement trajectories of the Red and Blue UAVs are as follows: Figure 3 As shown in Figure 2, the target change trend of the Red Army’s drone is as follows: Figure 4 shown.

[0127] See also Figure 5 , Figure 5 This is a second flow chart of a multi-UAV cooperative target allocation method provided in an embodiment of the present application. Figure 5 As shown in the figure, first, the situation information of the two drones is initialized, and the air combat scene parameters are set; according to the angle advantage function, speed advantage function, height advantage function and distance advantage function, the function form of the air combat situation evaluation function is obtained by weighting. Secondly, the evolutionary game model of the two drones is constructed, including: game participants, strategy space of evolutionary game, probability space and payment matrix. After that, the replication dynamic equation group is constructed according to the evolutionary game model and the situation evaluation function, and the evolutionary stable strategy solution is obtained by solving, and then the target allocation result of each drone unit is output; then, each drone is controlled to fight according to the assigned attack target, so as to update the situation information; when the drone unit of any drone shoots down its corresponding attack target, it is judged whether the drones of both parties meet the battle termination conditions. If not, it returns to solve the next round of target allocation; if it is satisfied, the battle between the drones of both parties is ended, and the final air combat result is output.

[0128] A multi-UAV collaborative target allocation method provided in an embodiment of the present application forms an evolutionary game model for multi-UAV air combat target allocation and a profit calculation method based on strategy space, payment matrix and probability space and according to a situation assessment function; a replication dynamic equation is constructed based on the expected strategy profit and the expected total profit to realize an evolutionary stable solution method for analyzing and simulating how the strategy evolves over time and how to reach the Nash equilibrium state of the evolutionary game, thereby solving the optimal strategy.

[0129] This method can dynamically evolve in response to time and environment, and has stronger applicability and robustness to the target allocation problem. By introducing evolutionary game theory, we can fully consider the strategies that the enemy may take against our actions, and then predict the dynamic interaction between the two drones, adapt to changes in the battlefield situation, and have a stronger ability to bypass the "local optimal trap", so as to obtain a more global and reasonable target allocation strategy for multi-UAV coordinated operations. As time goes by, the evolutionary game theory continuously adapts to the enemy's decision-making and changes in the battlefield situation, which makes it effective and feasible in the problem of coordinated target allocation in multi-UAV air combat.

[0130] Experiments show that the embodiments of the present application can maintain effective and stable performance in various experimental scenarios with advantages, balance, disadvantages and different numbers of enemies with the same strategy, and can further expand the advantages in advantageous scenarios and reduce losses in disadvantageous scenarios. Compared with traditional algorithms such as PSO algorithm, genetic algorithm, GA algorithm, etc., it has a higher average success rate.

[0131] See also Figure 6 , Figure 6 This is a schematic diagram of the structure of a multi-UAV cooperative target allocation device provided in an embodiment of the present application. Figure 6 As shown in FIG. 6 , the target allocation device 600 includes:

[0132] A construction module 610 is used to construct a strategy space of the evolutionary game between the two drones and a probability space of the two drones selecting each strategy;

[0133] The first determination module 620 is used to determine the situation evaluation function of one party's UAV relative to the other party's UAV according to the current situation information of the two parties' UAVs during any round of target allocation solution; wherein the situation evaluation function is used to evaluate the attack effectiveness of one party's UAV on the other party's UAV; the situation evaluation function is obtained by weighting at least one combat advantage function of one party's UAV relative to the other party's UAV;

[0134] The second determination module 630 is used to determine the attack benefit values ​​of the drones of both parties under all strategy combinations in the strategy space according to the situation evaluation function of the drone of one party relative to the drone of the other party, and form a payment matrix of the drones of both parties; wherein the attack benefit value of any drone of the party is defined by the ratio between the attack effectiveness of the drone of the party against the drone of the enemy and the attack effectiveness of the drone of the enemy against the drone of the party;

[0135] The third determination module 640 is used to determine the expected benefits of the two drones under each strategy combination and the expected total benefits of the two drones according to the strategy space of the evolutionary game of the two drones, the probability space of selecting each strategy, and the payment matrix of the two drones;

[0136] The fourth determination module 650 is used to traverse all strategies in the strategy space according to the probability space of the two drones selecting each strategy, the expected benefits of the two drones under each strategy, and the expected total benefits of the two drones, to form a replication dynamic equation group of the two drones; wherein the replication dynamic equation group is used to describe the evolution process of the strategy distribution of the two drones over time;

[0137] A solution module 660 is used to solve the replication dynamic equations of the two UAVs to obtain an evolutionary stable strategy of the two UAVs; wherein the benefits of the two UAVs reach an equilibrium state under the evolutionary stable strategy;

[0138] The control module 670 is used to control the drones of both sides to fight according to the attack targets assigned in the evolutionary stability strategy and update the situation information; until the drones of both sides meet the predetermined combat conditions, the first determination module 620 is called to solve the next round of target allocation.

[0139] Furthermore, when the control module 670 is used to call the first determination module 620 to solve the next round of target allocation when the UAVs of both parties meet the predetermined combat conditions, the control module 670 is used to:

[0140] When a drone unit in any one of the drones shoots down its corresponding attack target, it is determined whether the drones of both parties meet the battle termination condition; wherein the battle termination condition includes that all drone units in any one of the drones are shot down, or the battle between the drones of both parties reaches a preset number of rounds;

[0141] If it is not satisfied, the next round of target allocation is solved;

[0142] If satisfied, the battle between the two drones ends.

[0143] Furthermore, when the second determination module 630 is used to determine the attack benefit values ​​of the two drones under all strategy combinations in the strategy space according to the situation evaluation function of one drone relative to the other drone to form the payment matrix of the two drones, the second determination module 630 is used to:

[0144] According to the situation assessment function of one party's UAV relative to the other party's UAV, determine the situation assessment value when each UAV unit in one party's UAV selects each UAV unit in the other party's UAV as an attack target;

[0145] For each strategy, the total attack effectiveness of one drone on the other drone is determined according to the target allocation matrix of one drone under the strategy and the situation assessment value; wherein the element value of each element in the target allocation matrix is ​​used to characterize whether a drone unit in one drone selects a drone unit in the other drone as an attack target;

[0146] According to the total attack effectiveness of one party's drones on the other party's drones and the number of drones on one party, the average attack effectiveness of one party's drones under this strategy is determined;

[0147] For each strategy combination, the ratio of the average attack effectiveness of one party's drone under this strategy combination to the average attack effectiveness of the other party's drone is determined as the attack benefit value of one party's drone under this strategy combination, forming the payment matrix of one party's drone.

[0148] Furthermore, the second determination module 630 is used to determine the attack benefit values ​​of the two drones under all strategy combinations in the strategy space according to the situation evaluation function of one drone relative to the other drone through the following formula to form the payment matrix of the two drones:

[0149]

[0150]

[0151] In the formula, I={I R ,I B} is the probability space for both parties in the evolutionary game to choose each strategy, is the red square probability space, is the blue square probability space, Choose a pure strategy for Red The probability of Choose a pure strategy for Blue The probability of The situational advantage value when the red team's i-th drone chooses the blue team's j-th drone, For strategy The target allocation matrix; The situational advantage value when the j-th UAV of the blue side selects the i-th UAV of the red side; For strategy The target allocation matrix.

[0152] Furthermore, the third determination module 640 is used to determine the expected benefits of both drones under each strategy combination and the expected total benefits of both drones according to the strategy space of the evolutionary game of both drones, the probability space of selecting each strategy, and the payment matrix of both drones through the following formula:

[0153]

[0154]

[0155] in, The strategies adopted by the Red team are and Blue adopts strategy The expected return, and are the expected total returns of the red party and the blue party respectively; Choose a pure strategy for Red The probability of Choose a pure strategy for Blue probability.

[0156] Furthermore, when the fourth determination module 650 is used to traverse all strategies in the strategy space to form a replication dynamic equation group of the two drones according to the probability space of the two drones selecting each strategy, the expected benefits of the two drones under each strategy, and the expected total benefits of the two drones, the fourth determination module 650 is used to:

[0157] For any strategy, the replication dynamic equation of both drones is determined by the following formula:

[0158]

[0159] Traversing all strategies in the strategy space, the replication dynamic equations of both drones are expressed as:

[0160]

[0161] Furthermore, when the first determination module 620 is used to determine the situation evaluation function of one drone relative to the other drone according to the current situation information of both drones in any round of target allocation solution, the first determination module 620 is used to:

[0162] Determine the combat advantage function of one UAV relative to the other UAV based on the current situation information of the two UAVs; wherein the combat advantage function includes at least one of the following: angle advantage function, speed advantage function, height advantage function and distance advantage function;

[0163] According to each combat advantage function and the corresponding preset weight value, a weighted situation assessment function of one UAV relative to the other UAV is obtained.

[0164] See also Figure 7 , Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 7As shown in , the electronic device 700 includes a processor 710 , a memory 720 and a bus 730 .

[0165] The memory 720 stores machine-readable instructions executable by the processor 710. When the electronic device 700 is running, the processor 710 communicates with the memory 720 via the bus 730. When the machine-readable instructions are executed by the processor 710, the above-mentioned Figure 1 The steps of a multi-UAV collaborative target allocation method in the method embodiment shown are specifically implemented in accordance with the method embodiment and will not be described in detail here.

[0166] The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 1 The steps of a multi-UAV collaborative target allocation method in the method embodiment shown are specifically implemented in accordance with the method embodiment and will not be described in detail here.

[0167] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0168] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0169] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0170] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0171] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application can essentially be embodied in the form of a software product, or in other words, the part that contributes to the prior art or the part of the technical solution. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0172] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A multi-UAV cooperative target allocation method, characterized in that: The method comprises: Construct the strategy space of the evolutionary game between the two drones and the probability space of the two drones choosing each strategy; In the process of solving any round of target allocation, a situation evaluation function of one UAV relative to the other UAV is determined according to the current situation information of the two UAVs; wherein the situation evaluation function is used to evaluate the attack effectiveness of one UAV on the other UAV; the situation evaluation function is obtained by weighting at least one combat advantage function of one UAV relative to the other UAV; According to the situation evaluation function of one party's UAV relative to the other party's UAV, the attack benefit values ​​of both parties' UAVs under all strategy combinations in the strategy space are determined to form the payment matrix of both parties' UAVs; wherein, the attack benefit value of any party's UAV is defined by the ratio between the attack effectiveness of its own UAV on the enemy UAV and the attack effectiveness of the enemy UAV on its own UAV; According to the strategy space of the evolutionary game between the two drones, the probability space of selecting each strategy and the payment matrix of the two drones, the expected benefits of the two drones under each strategy and the expected total benefits of the two drones are determined; According to the probability space of the two drones selecting each strategy, the expected benefits of the two drones under each strategy combination, and the expected total benefits of the two drones, all strategies in the strategy space are traversed to form a replication dynamic equation group of the two drones; wherein the replication dynamic equation group is used to describe the evolution process of the strategy distribution of the two drones over time; Solving the replication dynamic equations of both drones to obtain the evolutionary stable strategies of both drones; wherein the benefits of both drones reach a balanced state under the evolutionary stable strategies; Control the UAVs of both parties to fight according to the attack targets assigned in the evolutionary stability strategy, and update the situation information; until the UAVs of both parties meet the predetermined combat conditions, solve the next round of target allocation; The following formula is used to determine the attack benefit values ​​of both drones under all strategy combinations in the strategy space based on the situation evaluation function of one drone relative to the other drone, forming the payment matrix of both drones: In the formula, is the probability space for both parties in the evolutionary game to choose each strategy, is the red square probability space, is the blue square probability space, Choose a pure strategy for Red The probability of Choose a pure strategy for Blue The probability of For the Red The first drone selected Blue The situational advantage value when flying a drone, For strategy The target allocation matrix; For Lan Fangdi The first drone selected by the Red Army Situational advantage value when flying a drone; For strategy The target allocation matrix; For The total attack effectiveness of the Red UAV against the Blue UAV in the strategy; For the strategy The total attack effectiveness of the blue drone against the red drone; For the strategy combination The red side's attack benefit value under ; For the strategy combination The blue team's attack benefit value under ; The payoff matrix for the Red Team’s drone; The payoff matrix for the blue drone; The following formula is used to determine the expected benefits of both drones under each strategy and the expected total benefits of both drones based on the strategy space of the evolutionary game between the two drones, the probability space of selecting each strategy, and the payment matrix of the two drones: in, , The strategies adopted by the Red team are and Blue adopts strategy The expected return, and are the expected total returns of the red party and the blue party respectively; Choose a pure strategy for Red The probability of Choose a pure strategy for Blue probability; According to the probability space of the two drones selecting each strategy, the expected benefits of the two drones under each strategy combination, and the expected total benefits of the two drones, all strategies in the strategy space are traversed to form the replication dynamic equation group of the two drones, including: For any strategy, the replication dynamic equation of both drones is determined by the following formula: in, and Replicated dynamic equations for both red and blue drones; Traversing all strategies in the strategy space, the replication dynamic equations of both drones are expressed as: 。 2. The method according to claim 1, characterized in that When the UAVs of both parties meet the predetermined combat conditions, the next round of target allocation is solved, including: When a drone unit in any one of the drones shoots down its corresponding attack target, it is determined whether the drones of both parties meet the battle termination condition; wherein the battle termination condition includes that all drone units in any one of the drones are shot down, or the battle between the drones of both parties reaches a preset number of rounds; If it is not satisfied, the next round of target allocation is solved; If satisfied, the battle between the two drones ends.

3. The method according to claim 1, characterized in that According to the situation evaluation function of one drone relative to the other drone, the attack benefit values ​​of both drones under all strategy combinations in the strategy space are determined to form the payment matrix of both drones, including: According to the situation assessment function of one party's UAV relative to the other party's UAV, determine the situation assessment value when each UAV unit in one party's UAV selects each UAV unit in the other party's UAV as an attack target; For each strategy, the total attack effectiveness of one drone on the other drone is determined according to the target allocation matrix of one drone under the strategy and the situation assessment value; wherein the element value of each element in the target allocation matrix is ​​used to characterize whether a drone unit in one drone selects a drone unit in the other drone as an attack target; According to the total attack effectiveness of one party's drones on the other party's drones and the number of drones on one party, the average attack effectiveness of one party's drones under this strategy is determined; For each strategy combination, the ratio of the average attack effectiveness of one party's drone under this strategy combination to the average attack effectiveness of the other party's drone is determined as the attack benefit value of one party's drone under this strategy combination, forming the payment matrix of one party's drone.

4. The method according to claim 1, characterized in that: According to the current situation information of the two UAVs, the situation evaluation function of one UAV relative to the other UAV is determined, including: Determine the combat advantage function of one UAV relative to the other UAV based on the current situation information of the two UAVs; wherein the combat advantage function includes at least one of the following: angle advantage function, speed advantage function, height advantage function and distance advantage function; According to each combat advantage function and the corresponding preset weight value, a weighted situation assessment function of one UAV relative to the other UAV is obtained.

5. A multi-UAV cooperative target allocation device, characterized in that: The device comprises: A construction module is used to construct the strategy space of the evolutionary game between the two drones and the probability space of the two drones choosing each strategy; The first determination module is used to determine the situation evaluation function of one party's UAV relative to the other party's UAV according to the current situation information of the two parties' UAVs during the solution process of any round of target allocation; wherein the situation evaluation function is used to evaluate the attack effectiveness of one party's UAV on the other party's UAV; the situation evaluation function is obtained by weighting at least one combat advantage function of one party's UAV relative to the other party's UAV; The second determination module is used to determine the attack benefit values ​​of the drones of both parties under all strategy combinations in the strategy space according to the situation evaluation function of the drone of one party relative to the drone of the other party, so as to form a payment matrix of the drones of both parties; wherein the attack benefit value of the drone of any party is defined by the ratio between the attack effectiveness of the drone of the own party against the drone of the enemy and the attack effectiveness of the drone of the enemy against the drone of the own party; The third determination module is used to determine the expected benefits of both drones under each strategy combination and the expected total benefits of both drones according to the strategy space of the evolutionary game of both drones, the probability space of selecting each strategy, and the payment matrix of both drones; The fourth determination module is used to traverse all strategies in the strategy space according to the probability space of the two drones selecting each strategy, the expected benefits of the two drones under each strategy, and the expected total benefits of the two drones, to form a replication dynamic equation group of the two drones; wherein the replication dynamic equation group is used to describe the evolution process of the strategy distribution of the two drones over time; A solution module is used to solve the replication dynamic equations of the two UAVs to obtain the evolutionary stable strategy of the two UAVs; wherein the benefits of the two UAVs reach a balanced state under the evolutionary stable strategy; A control module is used to control the UAVs of both parties to fight according to the attack targets assigned in the evolutionary stability strategy and update the situation information; until the UAVs of both parties meet the predetermined combat conditions, the first determination module is called to solve the next round of target allocation; The second determination module is used to determine the attack benefit values ​​of the two drones under all strategy combinations in the strategy space according to the situation evaluation function of one drone relative to the other drone through the following formula to form the payment matrix of the two drones: In the formula, is the probability space for both parties in the evolutionary game to choose each strategy, is the red square probability space, is the blue square probability space, Choose a pure strategy for Red The probability of Choose a pure strategy for Blue The probability of For the Red The first drone selected Blue The situational advantage value when flying a drone, For strategy The target allocation matrix; For Lan Fangdi The first drone selected by the Red Army Situational advantage value when flying a drone; For strategy The target allocation matrix; For The total attack effectiveness of the Red UAV against the Blue UAV in the strategy; For the strategy The total attack effectiveness of the blue drone against the red drone; For the strategy combination The red side's attack benefit value under ; For the strategy combination The blue team's attack benefit value under ; The payoff matrix for the Red Team’s drone; The payoff matrix for the blue drone; The third determination module is used to determine the expected benefits of both drones under each strategy and the expected total benefits of both drones according to the strategy space of the evolutionary game of both drones, the probability space of selecting each strategy, and the payment matrix of both drones through the following formula: in, , The strategies adopted by the Red team are and Blue adopts strategy The expected return, and are the expected total returns of the red party and the blue party respectively; Choose a pure strategy for Red The probability of Choose a pure strategy for Blue probability; When the fourth determination module is used to traverse all strategies in the strategy space to form a replication dynamic equation group of the two drones according to the probability space of the two drones selecting each strategy, the expected benefits of the two drones under each strategy combination, and the expected total benefits of the two drones, the fourth determination module is used to: For any strategy, the replication dynamic equation of both drones is determined by the following formula: in, and Replicated dynamic equations for both red and blue drones; Traversing all strategies in the strategy space, the replication dynamic equations of both drones are expressed as: 。 6. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and the machine-readable instructions are executed by the processor to execute the steps of a multi-UAV collaborative target allocation method as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the multi-UAV cooperative target allocation method as described in any one of claims 1 to 4 are executed.

Images