A Multi-Rule-Based Method for Assigning Multi-Target Strike Missions by Unmanned Aerial Vehicle (UAV) Swarms

Abstract

This invention discloses a multi-rule-based method for allocating multi-target strike missions by a swarm of unmanned aerial vehicles (UAVs), comprising the following steps: UAVs calculate their respective payoff values ​​in real time based on range cost, attack threat cost, and target strategic value; UAVs adjust their positions and determine the swarm mission allocation scheme based on the real-time payoff values; as the swarm approaches the target, UAVs determine whether to update the swarm mission allocation scheme or reach a termination condition based on the real-time payoff values; and a lead UAV guides attacking and exploration UAVs to approach and strike the target. This invention, considering payoff values, constructs a multi-suicide UAV target strike model with the goal of maximizing payoff value, and proposes a multi-rule UAV algorithm that effectively balances global exploration and local utilization, improves UAV search efficiency, and ensures efficient and accurate mission allocation.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) control technology, and specifically relates to a method for allocating multi-target strike missions by UAV swarms based on multiple rules. Background Technology

[0002] In modern warfare, unmanned aerial vehicles (UAVs) have become key combat tools, significantly enhancing the precision and mobility of military operations due to their flexibility, efficiency, and low cost. Therefore, UAV swarm cooperative warfare has become a new trend. Through information sharing and coordinated mission execution, it greatly improves combat effectiveness, covering a wider battlefield and enhancing reconnaissance, jamming, and strike capabilities against enemy targets. Cooperative operations not only increase mission success rates but also effectively disperse enemy defensive firepower, enhancing combat survivability. This UAV swarm cooperative mechanism lays a solid foundation for future intelligent warfare, making research on UAV swarm target strike missions extremely valuable strategically. However, existing UAV swarm cooperative multi-target task allocation scenarios still face problems such as insufficient robustness and susceptibility to local optima, making it difficult to cope with complex and ever-changing battlefield environments. Summary of the Invention

[0003] To address the shortcomings of the prior art, this invention provides a multi-rule-based method for allocating multi-target strike missions among unmanned aerial vehicle (UAV) swarms.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] A method for allocating multi-target strike missions by a swarm of unmanned aerial vehicles (UAVs) based on multiple rules includes the following steps:

[0006] S1: The UAV calculates its independent benefits in real time based on the range cost, attack threat cost, and target strategic value for each multi-target strike mission, and calculates the UAV's real-time global benefits in the entire mission allocation based on the independent benefits.

[0007] S2: Adjust the position of the drones and determine the swarm task allocation scheme based on the independent and global benefits obtained in real time. The swarm task allocation scheme is that the drone with the highest benefit value is the leader drone, and the remaining drones are assigned to the attack drones and exploration drones according to their benefit values.

[0008] S3: As the drone swarm approaches the target, the drones determine whether to update the swarm task allocation scheme or reach the termination condition based on the real-time calculated benefit value. If the termination condition is reached, the target strike allocation scheme is output. If the termination condition is not reached, the swarm task allocation scheme is updated based on the benefit value.

[0009] S4: The lead drone guides the attack drones and reconnaissance drones to approach and strike the target according to the target strike allocation plan.

[0010] Furthermore, the range cost is

[0011]

[0012] Where i represents the set of unmanned aerial vehicles U = {U1, U2, ..., U...} N Let j represent any UAV in the set T = {T1, T2, ..., T}, and let j represent the set of enemy targets. M Any enemy target in}, d ij d represents the distance between drone i and enemy target j. min The minimum range for a swarm of drones to perform its various mission objectives, d max The maximum range for the drone swarm to perform various target missions.

[0013] Furthermore, the cost of the attack threat is

[0014]

[0015] Among them, C j For enemy target T j The attack threat level brought about by it, C min C is the minimum value of the threat to all targets. max This represents the maximum value for the threat level across all targets.

[0016] Furthermore, the strategic value of the target is

[0017]

[0018] Among them, v j For enemy target T j The strategic value of v min v is the minimum strategic value of all objectives. max This represents the maximum strategic value of all objectives.

[0019] Furthermore, the return value is the sum of the independent return and the global return, where the independent return is...

[0020] E ij =ω1·NV j -ω2·NC j -ω3·D ij

[0021] Wherein, ω1, ω2, and ω3 are the weighting factors for value cost, attack cost, and range cost, respectively, and ω1+ω2+ω3=1.

[0022] The global benefit is

[0023]

[0024] The statement that all enemy targets were attacked by drones is indicated as follows: The statement that any UAV i can attack at most one enemy target is expressed as follows: x ij Let T be the decision variable, i.e., the objective T j Assigned to drone i.

[0025] Furthermore, the update rules for the swarm task allocation scheme are roaming rules and encirclement rules; the exploration drone updates its position according to the roaming rules, which are defined as follows:

[0026]

[0027] Among them, v i,d Let x be the position of the i-th exploration drone after a random walk in the d-th dimension. i,d The position of the i-th exploration drone in the d-th dimension, x k,d It is the random location of other exploration drones, x best,d For the optimal solution in the local neighborhood in dimension d, ε represents a very small positive number that controls the search neighborhood range, and λ is a probability factor that takes a random value according to the uniform distribution U[0,1].

[0028] The encirclement rules stipulate that the lead drone summons encircling drones to rapidly approach its location. During the approach, the encircling drones dynamically adjust according to the encirclement rules, specifically in the following form:

[0029]

[0030] Among them, step init Let α be the initial siege step size. d Let d represent the position of the leading drone in the d-th dimension, δ represent the decay rate of the step size, and t represent the number of iterations. Let be the position of the drone in the d-th dimension at the k-th iteration. Let d be the position of the drone in the d-th dimension during the (k+1)-th iteration.

[0031] Furthermore, the termination condition is whether the difference between the ideal optimal profit value and the current maximum profit value of the drone is less than a preset threshold or whether the maximum number of searches has been reached.

[0032] Furthermore, in the update rules of the swarm task allocation scheme, a certain number of drones with low revenue values ​​are eliminated in each round, and the same number of drones are randomly generated to replenish them.

[0033] Furthermore, after the siege ends, the lead drone in the drone swarm autonomously roams to further explore and search the surrounding area.

[0034] The advantages of this invention are as follows: the technical solution of this invention considers the cost of flight range, the cost of attack threat, and the strategic value of the target, and constructs a multi-suicide drone strike target model with the goal of maximizing the benefit value. It proposes a multi-rule drone algorithm, establishes a leader drone generation rule with an autonomous walking mechanism, a multi-strategy exploration drone walking rule, and an encirclement drone encirclement rule with an adaptive step size, effectively balancing global exploration and local utilization, improving drone search efficiency, and ensuring the efficiency and accuracy of task allocation. Attached Figure Description

[0035] Figure 1 This is a flowchart of a multi-rule-based method for allocating multi-target strike missions by a swarm of unmanned aerial vehicles (UAVs) according to the present invention.

[0036] Figure 2 This is a comparison chart of the errors of four different algorithms;

[0037] Figure 3 This is a comparison chart of the minimum fitness of four different algorithms under different sized clusters;

[0038] Figure 4 This is a comparison chart of the average fitness of four different algorithms under different sized clusters. Detailed Implementation

[0039] like Figure 1 As shown, a multi-rule-based method for allocating multi-target strike missions by a swarm of unmanned aerial vehicles (UAVs) includes the following steps:

[0040] S1: Each UAV calculates its independent benefit in real time based on range cost, attack threat cost, and target strategic value for a multi-target strike mission, and calculates its global benefit in real time throughout the mission allocation based on its independent benefit. During the search process of the UAV swarm, each UAV updates and calculates its own independent and global benefits in real time, and compares them with the independent and global benefits of other UAVs in the same communication network.

[0041] S2: Based on the independent and global benefits obtained through real-time calculation, the positions of the drones are adjusted, and a swarm task allocation scheme is determined. The drone with the highest benefit value (independent benefit + global benefit) is designated as the leader drone, and the remaining drones are divided into attack drones and exploration drones based on their benefit values. The swarm task allocation scheme primarily determines the status of the drones within the swarm. The leader drone guides the attack and exploration drones towards the target, adjusting their positions within the swarm continuously based on their real-time independent and global benefits during the approach process. The attack drones employ an adaptive step-size mechanism, dynamically adjusting their step size according to the current search state. This ensures both the flexibility of fine-grained searching and improved precision strike capability during the attack. The exploration drones enhance the overall swarm's search capability, expand the search space, improve global optimization performance, and ensure the discovery of potential global optimal solutions over a wider area. The status of the attack and exploration drones is determined by comparing their calculated benefit values ​​with those calculated by the leader drone. For example, the lead drone is calculated to have a profit value of 5.0, drones with a profit value between 4.9 and 4.0 are considered attack drones, drones with a profit value between 3.9 and 3.0 are considered exploration drones, and the remaining drones are considered to be not part of a swarm attacking a specific target.

[0042] S3: As the drone swarm approaches the target, the drones determine whether to update the swarm task allocation scheme or reach a termination condition based on real-time calculated gains. If the termination condition is met, the target attack allocation scheme is output; otherwise, the swarm task allocation scheme is updated based on the gains. Updating the swarm task allocation scheme involves continuously changing the drones' positions within the swarm based on their calculated gains, until the optimal solution for drones to approach the target is obtained or the number of iterations reaches its limit. The termination condition is whether the difference between the ideal optimal gains and the drone's current maximum gains is less than a preset threshold or the maximum number of searches is reached. Once the termination condition is met, the swarm task allocation scheme is no longer updated until the leading drone is attacked and destroyed. In the update rules of the swarm task allocation scheme, a certain number of drones with lower gains are eliminated each round, and the same number of drones are randomly generated to replenish them, ensuring that the number of drones attacking the target does not decrease.

[0043] S4: The lead drone guides the attacking and exploration drones to approach and engage the target according to the target engagement allocation plan. After the target attack is completed, the lead drone will autonomously roam to find new targets, that is, after the siege ends, the lead drone will autonomously roam to further explore and search the surrounding area.

[0044] The target attack allocation scheme is that the lead drone guides the attacking drones. If the lead drone is destroyed, the swarm will re-determine a new lead drone based on the benefit value and continue to attack the target until the target is destroyed.

[0045] The range cost is that each drone carries a fixed power source and has a limited range. The longer the drone flies, the longer it remains exposed to the enemy's view, increasing the probability of detection. Therefore, a shorter range yields higher benefits for the drone. Thus, the range cost is constructed as follows:

[0046]

[0047] Where i represents the set of unmanned aerial vehicles U = {U1, U2, ..., U...} N Let j represent any UAV in the set T = {T1, T2, ..., T}, and let j represent the set of enemy targets. M Any enemy target in}, d ij d represents the distance between drone i and enemy target j. min The minimum range for a swarm of drones to perform its various mission objectives, d max The maximum range for the drone swarm to perform various target missions.

[0048] The attack threat cost refers to the potential for attacks from different equipment carried by the enemy target when the drone attacks, which will affect the gains to varying degrees. Therefore, the attack threat cost is expressed as...

[0049]

[0050] Among them, C j For enemy target T j The attack threat level brought about by it, C min C is the minimum value of the threat to all targets. max This represents the maximum value for the threat level across all targets.

[0051] The strategic value of the target is the strategic value of the target itself, therefore the strategic value of the target is expressed as:

[0052]

[0053] Among them, v j For enemy target T j The strategic value of v j Normalization is performed, v min v is the minimum strategic value of all objectives. max This represents the maximum strategic value of all objectives.

[0054] Based on the range cost, attack threat cost, and target strategic value, the benefit E obtained by any UAV i after performing target mission j can be derived. ij The independent revenue is

[0055] E ij =ω1·NVj -ω2·NC j -ω3·D ij

[0056] Wherein, ω1, ω2, and ω3 are the weighting factors for value cost, attack cost, and range cost, respectively, and ω1+ω2+ω3=1.

[0057] Assuming M = N, and each suicide drone can only attack one target, the drone swarm cooperative task allocation model can be described as maximizing the total benefit, where the global benefit is...

[0058]

[0059] The statement that all enemy targets were attacked by drones is indicated as follows: The statement that any UAV i can attack at most one enemy target is expressed as follows: x ij Let T be the decision variable, i.e., the objective T j Assigned to drone i.

[0060] The revenue value is the sum of the independent revenue and the global revenue, and the position of the drone in the swarm is determined by the revenue value.

[0061] The update rules for the swarm task allocation scheme are roaming rules and encirclement rules. Based on these rules, targets are optimized to achieve the best task allocation. The exploratory drones update their positions according to the roaming rules, which are defined as follows:

[0062]

[0063] Among them, v i,d Let x be the position of the i-th exploration drone after a random walk in the d-th dimension. i,d The position of the i-th exploration drone in the d-th dimension, x k,d It is the random location of other exploration drones, x best,d For the optimal solution in the local neighborhood in dimension d, ε represents a very small positive number that controls the range of the search neighborhood, and λ is a probability factor that takes a random value according to the uniform distribution U[0,1].

[0064] when The proposed rule is a neighbor-based exploratory drone walking rule. This rule retains the basic search capabilities of the original swarm algorithm and improves the algorithm's exploration capabilities by using the positions of other drones to guide the exploratory drones' walking. This effectively enhances the algorithm's initial exploration capabilities and helps find better local solutions in the solution space.

[0065] when In this context, the exploration drone walking rules based on local optimum information allow the exploration drone to more fully explore the current area, which helps to conduct a more detailed search near known better solutions.

[0066] when In this context, the exploration drone walking rule based on neighborhood search allows the exploration drone to perform a fine search in the vicinity of the current solution, which helps to further explore the vicinity of the potential good solutions that have already been found, thereby improving the accuracy of the solution and the local optimization capability.

[0067] The exploration drone's current reward value is calculated according to the roaming rules. If it is greater than the leader drone's reward value, it becomes the new leader drone; otherwise, it roams according to the multi-strategy roaming rules, i.e., swapping element positions, until its reward value exceeds the leader drone's reward value or the maximum number of roams is reached. Therefore, by implementing multi-strategy exploration drone roaming rules, the algorithm's search capability and global optimization performance can be enhanced.

[0068] The encirclement rule is that the lead drone summons encircling drones to rapidly approach its location. During this approach, the encircling drones dynamically adjust according to the encirclement rule. Specifically, by introducing an adaptive step size mechanism, the step size of the encircling drone gradually decreases with each iteration as it approaches the lead drone, or it dynamically adjusts based on the distance between the encircling and lead drones. The specific form is as follows:

[0069]

[0070] Among them, step init Let α be the initial siege step size. d Let d represent the position of the leading drone in the d-th dimension, δ represent the decay rate of the step size, and t represent the number of iterations. Let be the position of the drone in the d-th dimension at the k-th iteration. Let d be the position of the drone in the d-th dimension during the (k+1)-th iteration.

[0071] In the siege rule, the leader drone summons attacking drones to rapidly approach its location. The attacking drones adaptively adjust according to the siege rule. If their reward value is greater than that of the current leader drone, then the attacking drone becomes the new leader drone; otherwise, the siege continues until the prey is captured. The adaptive step size (siege rule) not only improves the algorithm's convergence speed but also effectively avoids convergence difficulties caused by excessively large step sizes.

[0072] To verify the feasibility of the allocation method (denoted as MR-WPA) of this invention, a comparative analysis was conducted with Genetic Algorithm (GA), Particle Swarm Optimization (PSO), and Wolf Pack Algorithm (WPA). By subtracting the reward values ​​obtained by each algorithm from the optimal reward value obtained based on linear programming, error variation graphs for different algorithms at different iteration numbers can be obtained, as shown below. Figure 2 As shown in the figure, the MR-WPA algorithm has significant advantages in both speed and accuracy. The improved MR-WPA algorithm can reduce the error from the optimal return value to less than 0.1 at a faster speed, and under the same number of iterations, the final error is smaller than that of the other three algorithms.

[0073] For small-scale swarm problems, due to the small solution space, the four algorithms have similar performance and yield similar rewards. However, as the swarm size increases, the performance differences between the different algorithms gradually widen, especially the gap between the MR-WPA and WPA algorithms and the genetic algorithm and particle swarm optimization algorithm. Figure 3 and Figure 4 As shown.

[0074] Although the specific embodiments of the present invention have been described and illustrated in detail above, it should be noted that various changes and modifications can be made to the above embodiments without departing from the spirit of the present invention and the scope set forth in the appended claims.

Claims

1. A method for allocating multi-target strike missions by a swarm of unmanned aerial vehicles (UAVs) based on multiple rules, characterized in that, Includes the following steps: S1: The UAV calculates its independent benefits in real time based on the range cost, attack threat cost, and target strategic value for each multi-target strike mission, and calculates the UAV's real-time global benefits in the entire mission allocation based on the independent benefits. S2: Adjust the position of the drones and determine the swarm task allocation scheme based on the independent and global benefits obtained in real time. The swarm task allocation scheme is that the drone with the highest benefit value is the leader drone, and the remaining drones are divided into attack drones and exploration drones according to their benefit values. S3: As the drone swarm approaches the target, the drones determine whether to update the swarm task allocation scheme or reach the termination condition based on the real-time calculated benefit value. If the termination condition is reached, the target strike allocation scheme is output. If the termination condition is not reached, the swarm task allocation scheme is updated based on the benefit value. S4: The lead drone guides the attack drones and reconnaissance drones to approach and strike the target according to the target strike allocation plan.

2. The method for allocating multi-target strike missions by a swarm of unmanned aerial vehicles (UAVs) based on multiple rules, as described in claim 1, is characterized in that... The cost of the journey is Where i represents the set of unmanned aerial vehicles U = {U1, U2, ..., U...} N Let j represent any UAV in the set T = {T1, T2, ..., T}, and let j represent the set of enemy targets. M Any enemy target in}, d ij d represents the distance between drone i and enemy target j. min The minimum range for a swarm of drones to perform its various mission objectives, d max The maximum range for the drone swarm to perform various target missions.

3. The method for allocating multi-target strike missions by a swarm of unmanned aerial vehicles (UAVs) based on multiple rules, as described in claim 2, is characterized in that... The cost of the attack threat is Among them, C j For enemy target T j The attack threat level brought about by it, C min C is the minimum value of the threat to all targets. max This represents the maximum value for the threat level across all targets.

4. The method for allocating multi-target strike missions by a swarm of unmanned aerial vehicles (UAVs) based on multiple rules, as described in claim 3, is characterized in that... The strategic value of the target is Among them, v j For enemy target T j The strategic value of v min v is the minimum strategic value of all objectives. max This represents the maximum strategic value of all objectives.

5. The method for allocating multi-target strike missions by a swarm of unmanned aerial vehicles (UAVs) based on multiple rules, as described in claim 4, is characterized in that... The return value is the sum of the independent return and the global return, where the independent return is... E ij =ω1·NV j -ω2·NC j -ω3·D ij Where ω1, ω2, and ω3 are the weighting factors for value cost, attack cost, and range cost, respectively, and ω1 + ω2 + ω3 = 1. The global benefit is The statement that all enemy targets were attacked by drones is indicated as follows: The statement that any UAV i can attack at most one enemy target is expressed as follows: x ij Let T be the decision variable, i.e., the objective T j Assigned to drone i.

6. The method for allocating multi-target strike missions by a swarm of unmanned aerial vehicles (UAVs) based on multiple rules, as described in claim 5, is characterized in that... The update rules for the cluster task allocation scheme are the roaming rule and the siege rule; The exploration drone updates its position according to a roaming rule, which is defined as follows: Among them, v i,d Let x be the position of the i-th exploration drone after a random walk in the d-th dimension. i,d The position of the i-th exploration drone in the d-th dimension, x k,d It is the random location of other exploration drones, x best,d For the optimal solution in the local neighborhood in dimension d, ε represents a very small positive number that controls the search neighborhood range, and λ is a probability factor that takes a random value according to the uniform distribution U[0,1]. The encirclement rules stipulate that the lead drone summons encircling drones to rapidly approach its location. During the approach, the encircling drones dynamically adjust according to the encirclement rules, specifically in the following form: Among them, step init Let α be the initial siege step size. d Let d represent the position of the leading drone in the d-th dimension, δ represent the decay rate of the step size, and t represent the number of iterations. Let be the position of the drone in the d-th dimension at the k-th iteration. Let d be the position of the drone in the d-th dimension during the (k+1)-th iteration.

7. The method for allocating multi-target strike missions by a swarm of unmanned aerial vehicles (UAVs) based on multiple rules, as described in claim 6, is characterized in that... The termination condition is whether the difference between the ideal optimal profit value and the current maximum profit value of the drone is less than a preset threshold or whether the maximum number of searches has been reached.

8. The method for allocating multi-target strike missions by a swarm of unmanned aerial vehicles (UAVs) based on multiple rules, as described in claim 7, is characterized in that... In the update rules of the swarm task allocation scheme, a certain number of drones with low revenue values ​​are eliminated in each round, and the same number of drones are randomly generated to replenish them.

9. A method for allocating multi-target strike missions by a swarm of unmanned aerial vehicles (UAVs) based on multiple rules, as described in claim 8, is characterized in that... After the siege ended, the lead drone in the drone swarm autonomously roamed around to further explore and search the surrounding area.

Images