Computational method of UAV swarm intelligence model based on anti-predation characteristics of bird flocks

Abstract

The present invention provides an intelligent model and calculation method of drone clusters based on the anti-predation characteristics of bird flocks, including the following steps: Step 1, analyzing the source and mechanism of the anti-predation characteristics of bird flocks, and establishing the dynamic action characteristics relationship of multiple types of birds; Step 2, establishing an intelligent model of drone clusters based on the anti-predation characteristics of bird flocks; Step 3, designing an intelligent calculation method of drone clusters based on the anti-predation characteristics of bird flocks, and setting optimization rules and parameter adjustment methods. The present invention provides an effective way to efficiently and dynamically control drone clusters, ensuring that drone clusters can respond quickly to changes in the external environment. The model and calculation method can also be applied to other cluster intelligent control and optimization problems.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent computing models and algorithms based on biological swarm intelligence, and in particular to a computing method for an unmanned aerial vehicle (UAV) swarm intelligence model based on the anti-predation characteristics of bird flocks. Background Art

[0002] In the current study, PSO (Particle Swarm Optimization) has been shown to be effective in optimizing static environments, i.e., solutions to most static problems usually converge to the optimal solution. However, PSO performs poorly in environments where the optimal value is dynamic and constantly changing. The PSO algorithm is limited in capturing characteristic optimal deviations in dynamic environments, which limits its speed and convergence to the optimal solution. Summary of the invention

[0003] Purpose of the invention: The technical problem to be solved by the present invention is to provide a calculation method of a drone swarm intelligence model based on the anti-predation characteristics of bird flocks, in view of the shortcomings of the prior art, including:

[0004] Step 1: Analyze the source and mechanism of anti-predation characteristics of bird flocks and establish the relationship between dynamic action characteristics of multiple types of birds.

[0005] Step 2: Digitally describe various behaviors and establish a drone swarm intelligence model based on the anti-predation characteristics of bird flocks.

[0006] Step 3: Design a cluster intelligence calculation method based on the anti-predation characteristics of bird flocks, set optimization rules and parameter adjustment methods, and design optimization rules and parameter adjustment methods for explorers and followers according to the different roles in the drone cluster, and establish a complete algorithm process and judgment mechanism.

[0007] The anti-predation characteristics of the bird flock in step 1 are manifested in that the bird flock will gather regularly near the habitat to form a spherical structure, which will continue to appear, collapse and evolve over time to prevent attacks by birds of prey such as peregrine falcons.

[0008] Step 1 includes:

[0009] Step 1-1, build a reconnaissance characteristic relationship: when one member of the team senses a threat, the other individuals in the team are immediately informed of the danger;

[0010] Step 1-2: Build a clustering relationship: Continuously prepare for any changes in the environment. When predators appear in the environment, the flock will form clustering behaviors to avoid or respond to danger;

[0011] Steps 1-3. Build collaborative relationships: In a flock of birds, the direction of movement of each bird depends not only on the individual’s knowledge, but also on the knowledge of the rest of the birds in the flock, which then adjust their routes based on the directions of their neighbors.

[0012] In step 2, the drone cluster intelligence model based on the anti-predation characteristics of bird flocks includes a reconnaissance characteristic model, a clustering model and a collaborative relationship model.

[0013] In step 2, the reconnaissance characteristic model is: the probability P of at least one of the N drones being able to detect the predator GD The calculation formula is as follows:

[0014] P GD =1-e -VTN

[0015] Where V is the scanning frequency of each drone, T is the time interval after the predator is discovered and survived, and e is the natural logarithm; the scanning frequency V of the drone is:

[0016]

[0017] In step 2, the clustering model is as follows: each drone moves in an aerodynamic form and transmits information within the drone cluster. The information sharing model within the cluster is constructed as follows:

[0018]

[0019] Among them, C int is the average value of the correlation matrix of drones within the neighborhood size, n i c is the neighborhood size of the i-th UAV, n c is the neighborhood size of the N drones forming a cluster, is the correlation matrix between the normalized velocities of the field data, represents the normalized speed of UAV i and j.

[0020] In step 2, the collaborative relationship model is as follows: the collaborative relationship C between individuals i and j in the cluster ij The expression is as follows:

[0021]

[0022] in, Representing collaborative relationships The probability distribution of Z({J ij}) is the normalization factor, J ij represents the interaction strength; C ij It is the collaborative relationship between individuals i and j.

[0023] Step 3 includes:

[0024] Step 3-1, initializing individual parameters within the drone cluster, wherein the individual parameters include location information;

[0025] Step 3-2: Use the Rastrigin function to judge the quality of each individual extreme value and select the optimal solution. The parameter boundary range of the Rastrigin function is [-5,5]. The formula of the Rastrigin function f(x) is as follows:

[0026]

[0027] Among them, x i is the location information of individual i; by optimizing x i Group, find the minimum solution of the function f(x);

[0028] Step 3-3, scan and determine the external predators. If the current environment has not changed or the optimal solution has not changed, go to step 3-4. If the current environment has changed or the optimal solution has changed, go to step 3-2.

[0029] Step 3-4: Update and adjust the collective response position, find the 8 nearest individuals around the optimal solution individual, and update the position of each individual based on the topological relationship between them and the optimal individual. as follows;

[0030]

[0031] in, is the individual speed, Δt is the time interval;

[0032] Step 3-5: Update and adjust the cluster reaction speed. According to the topological relationship between the eight nearest individuals around the optimal solution individual and the optimal solution individual, construct the following speed update equation:

[0033]

[0034] in, are the speed of individual k at time t+1 and time t, respectively; c1 and c2 are cognitive coefficients; r1 and r2 are social coefficients; ω is the inertia weight; The best solution in history. For the current optimal solution, each iteration updates the next position data.

[0035] Step 3-6: Optimize and adjust the local optimization trap problem, introduce a random jump-out mechanism for the position and speed update equations, and construct the following anti-local trap position update equation:

[0036]

[0037] Among them, x k is the optimized individual k position information, To optimize the position information of the previous individual k, r x is a random number between [0,1], N k is the index of the eight adjacent individuals of individual k, x n is the location information of the adjacent individual n of individual k.

[0038] The present invention simulates the response of bird groups to dynamic external threats, establishes an intelligent model of drone swarms and an optimization calculation method, and solves and optimizes numerical problems of the interaction between the dynamic environment and the individual positions of drones.

[0039] Beneficial effects: The present invention designs a calculation method for a drone swarm intelligent model based on the anti-predation characteristics of bird flocks. The model introduces the dynamic stress behavior of bird flocks under predator attacks, so that PSO adapts to dynamic environments and can converge to the optimal solution simply and quickly. The present invention provides an effective way to efficiently and dynamically control drone swarms, ensuring that drone swarms can respond quickly to changes in the external environment. The model and calculation method can also be applied to other swarm intelligent control and optimization problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments, and the above and / or other advantages of the present invention will become more clear.

[0041] Figure 1 It is a flow chart of the method of the present invention.

[0042] Figure 2 This is a schematic diagram of the anti-predation characteristics of bird flocks.

[0043] Figure 3 This is a schematic diagram of the Rastrigin function.

[0044] Figure 4 This is the initial drone cluster distribution map.

[0045] Figure 5 It is a distribution map of drone clusters based on the anti-predation characteristics of bird flocks under the control of an intelligent model.

[0046] Figure 6 This is the drone cluster distribution map calculated by this model for dynamic stress behavior. DETAILED DESCRIPTION

[0047] like Figure 1 As shown, the present invention provides a method for calculating a drone cluster intelligence model based on the anti-predation characteristics of a flock of birds, including:

[0048] Step 1: Analyze the source and mechanism of anti-predation characteristics of bird flocks and establish the relationship between dynamic action characteristics of multiple types of birds.

[0049] Birds are a highly gregarious species, especially in the fall and winter. Although flock sizes are extremely diverse and large, birds gather in regular clusters near their roosts, called cacophony. This concentration is thought to be a preventative measure against attacks by birds of prey such as peregrine falcons. Figure 2 Schematic diagram of the anti-predation characteristics of bird flocks.

[0050] Birds in a flock will form spherical structures that will appear, collapse, and evolve over time. For this reason, it is necessary to establish characteristic relationships to provide theoretical support for the drone swarm intelligence model.

[0051] Step 1-1, build the reconnaissance characteristic relationship: The reconnaissance characteristic of birds means that when one member of the group senses a threat, such as a predator, the other individuals in the group will be informed of the danger almost immediately. Generally speaking, there are three parameter variables involved in the predator detection process: (a) scanning speed, (b) vigilance and nearest neighbors, and (c) previous dangerous experiences in its surrounding environment, which will lead to the activation of defense mechanisms within the bird group.

[0052] Step 1-2: Build a flocking relationship: Continuous preparation for any changes in the environment. When the environment changes dramatically, the flock will form a flocking behavior to avoid or respond to danger. The flocking behavior of birds follows three simple rules, namely: cohesion, alignment and separation.

[0053] Step 1-3, build collaborative relationships: In a flock of birds, the direction of movement of each bird depends not only on the individual's knowledge, but also on the knowledge of the rest of the birds in the flock. The flock communicates with their neighbors and then adjusts their routes based on the directions of their neighbors. This feature involves the topological relationship of the flock and needs to be built based on the topological structure of the cluster.

[0054] Step 2: Parameterize and model the reconnaissance characteristics, aggregation characteristics, and collaborative relationships, and establish a drone swarm intelligence model based on the anti-predation characteristics of bird flocks.

[0055] Step 2-1, reconnaissance characteristic model: the probability P of at least one of the N drones being able to detect the predator GD The calculation formula is as follows:

[0056] P GD =1-e -VTN

[0057] Where V is the scanning frequency of each drone, T is the time interval after the predator is discovered and survived, and e is the natural logarithm.

[0058] The scanning frequency of the drone is

[0059]

[0060] Step 2-2, Cluster Model: Each drone moves in thousands of forms in certain aerodynamic forms and transmits information very quickly within the cluster. The information sharing model within the cluster is constructed as follows:

[0061]

[0062] In the formula, C int is the average value of the correlation matrix of drones within the neighborhood size, n i c is the area size of the i-th UAV, n c is the neighborhood size of the N drones forming a cluster, is the correlation matrix between the normalized velocities of the field data, represents the normalized speed of UAV i and j.

[0063] Step 2-3, synergistic relationship model: synergistic relationship C between individuals i and j in the population ij The expression is as follows:

[0064]

[0065] In the formula, Representing collaborative relationships The probability distribution of Z({J ij}) is the normalization factor, represents the normalized speed of i and j, J ij Indicates the interaction strength.

[0066] Step 3: Design an intelligent computing method for drone swarms based on the anti-predation characteristics of bird flocks, set optimization rules and parameter adjustment methods, and establish a complete algorithm process and judgment mechanism.

[0067] Step 3-1: Initialize individual parameters within the drone cluster, i.e., location information;

[0068] Step 3-2, select the Rastrigin function (such as Figure 3 As shown in the figure, the value of each individual is judged and the optimal solution is selected. Figure 3 The horizontal and vertical coordinates are the corresponding horizontal positions of the individuals.

[0069] The parameter boundary range of the function is [-5,5], and the function formula is as follows:

[0070]

[0071] Among them, x i is the location information of individual i. By optimizing x i Group, find the minimum solution of the function f(x).

[0072] Step 3-3: Scan and determine external predators. If the current environment has not changed or the optimal solution has not changed, go to step 3-4. If the current environment has changed or the optimal solution has changed, go to step 3-2.

[0073] Step 3-4: Update and adjust the collective response position, find the 8 nearest individuals around the optimal solution individual, and update the position of each individual based on the topological relationship between them and the optimal individual. as follows;

[0074]

[0075] in, is the individual speed, Δt is the time interval; Step 3-5, update and adjust the cluster reaction speed, and construct the speed update equation as follows based on the position topological relationship between the 8 nearest individuals around the optimal solution individual and the optimal solution individual:

[0076]

[0077] in, are the speeds of individual k at time t+1 and t respectively, c1 and c2 are cognitive coefficients, r1 and r2 are social coefficients, ω is the inertia weight, The best solution in history. For the current optimal solution, each iteration updates the next position data.

[0078] Step 3-6: Optimize and adjust the local optimization trap problem, introduce a random jump-out mechanism for the position and speed update equations, and construct the following anti-local trap position and speed update equations:

[0079]

[0080] Among them, x k is the optimized individual k position information, To optimize the position information of the previous individual k, r x is a random number between [0,1], N k is the index of the eight adjacent individuals of individual k, x n is the location information of the adjacent individual n of individual k.

[0081] The calculation model and calculation method of the present invention are used to analyze the movement of drone clusters. The distribution of drone clusters under initial conditions is as follows: Figure 4 The calculation model of the present invention can be used to calculate the movement trajectory of the drone cluster through multiple iterations, and the distribution map of the drone cluster can be quickly obtained (such as Figure 5For emergencies, the distribution map of drone clusters after dynamic stress behavior can be obtained based on the stress response of the bird flock (as shown in Figure 6 shown). Figure 4 , Figure 5 , Figure 6 The horizontal and vertical coordinates are the horizontal positions of the individual drones. From the above results, it can be seen that the calculation model and calculation method of the present invention can quickly realize the control of the drone cluster formation, and can quickly respond to various emergencies and adjust the cluster formation.

[0082] In a specific implementation, the present application provides a computer storage medium and a corresponding data processing unit, wherein the computer storage medium can store a computer program, and when the computer program is executed by the data processing unit, the invention content of the calculation method of the drone cluster intelligent model based on the anti-predation characteristics of the bird flock provided by the present invention and some or all of the steps in each embodiment can be run. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0083] Those skilled in the art can clearly understand that the technical solutions in the embodiments of the present invention can be implemented by means of computer programs and their corresponding general hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention can be essentially or partly contributed to the prior art in the form of a computer program, i.e., a software product, which can be stored in a storage medium and includes several instructions for enabling a device including a data processing unit (which can be a personal computer, a server, a single-chip microcomputer, a MUU or a network device, etc.) to execute the methods described in various embodiments of the present invention or certain parts of the embodiments.

[0084] The present invention provides a calculation method for a drone swarm intelligence model based on the anti-predation characteristics of bird flocks. There are many methods and ways to implement the technical solution. The above is only a preferred implementation of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the protection scope of the present invention. All components not specified in this embodiment can be implemented using existing technologies.

Claims

1. The calculation method of the UAV swarm intelligence model based on the anti-predation characteristics of bird flocks is characterized by: The following steps are involved: Step 1: Analyze the source and mechanism of anti-predation characteristics of bird flocks and establish the dynamic action characteristics relationship of multiple types of birds; Step 2: Establish a UAV swarm intelligence model based on the anti-predation characteristics of bird flocks; Step 3: Design an intelligent computing method for drone swarms based on the anti-predation characteristics of bird flocks, and set optimization rules and parameter adjustment methods; In step 2, the drone swarm intelligence model based on the anti-predation characteristics of bird flocks includes a reconnaissance characteristic model, a clustering model and a collaborative relationship model; In step 2, the reconnaissance characteristic model is: the probability P of at least one of the N drones being able to detect the predator GD The calculation formula is as follows: P GD =1-e -VTN Where V is the scanning frequency of each drone, T is the time interval after the predator is discovered and survived, and e is the natural logarithm; the scanning frequency V of the drone is: In step 2, the clustering model is as follows: each drone moves in an aerodynamic form and transmits information within the drone cluster. The information sharing model within the cluster is constructed as follows: Among them, C int is the average value of the correlation matrix of drones within the neighborhood size, n i c is the neighborhood size of the i-th drone, n c is the neighborhood size of the N drones forming a cluster, is the correlation matrix between the normalized velocities of the field data, represents the normalized speed of UAV i and j; In step 2, the collaborative relationship model is as follows: the collaborative relationship C between individuals i and j in the cluster ij The expression is as follows: in, Representing collaborative relationships The probability distribution of Z({J ij }) is the normalization factor, J ij represents the interaction strength; C ij It is the collaborative relationship between individuals i and j.

2. The method according to claim 1, characterized in that Step 1 includes: Step 1-1, build a reconnaissance characteristic relationship: when one member of the team senses a threat, the other individuals in the team are immediately informed of the danger; Step 1-2: Build aggregative relationships: When predators appear in the environment, the flock will form aggregative behaviors to avoid or deal with danger; Steps 1-3. Build collaborative relationships: In a flock of birds, the direction of movement of each bird depends not only on the individual’s knowledge, but also on the knowledge of the rest of the birds in the flock, which then adjust their routes based on the directions of their neighbors.

3. The method according to claim 2, characterized in that Step 3 includes: Step 3-1, initializing individual parameters within the drone cluster, wherein the individual parameters include location information; Step 3-2: Use the Rastrigin function to judge the quality of each individual extreme value and select the optimal solution. The parameter boundary range of the Rastrigin function is [-5,5]. The formula of the Rastrigin function f(x) is as follows: Among them, x i is the location information of individual i; by optimizing x i Group, find the minimum solution of the function f(x); Step 3-3, scan and determine the external predators. If the current environment has not changed or the optimal solution has not changed, go to step 3-4. If the current environment has changed or the optimal solution has changed, go to step 3-2. Step 3-4: Update and adjust the collective response position, find the 8 nearest individuals around the optimal solution individual, and update the position of each individual based on the topological relationship between them and the optimal individual. as follows; in, is the individual speed, Δt is the time interval; Step 3-5: Update and adjust the cluster response speed; Step 3-6: Optimize and adjust the local optimization problems.

4. The method according to claim 3, characterized in that In step 3-5, the following speed update equation is constructed based on the position topological relationship between the 8 nearest individuals around the optimal solution individual and the optimal solution individual: in, are the speed of individual k at time t+1 and time t, respectively; c1 and c2 are cognitive coefficients; r1 and r2 are social coefficients; ω is the inertia weight; The best solution in history. For the current optimal solution, each iteration updates the next position data.

5. The method according to claim 4, characterized in that In steps 3-6, a random jump-out mechanism is introduced for the position and velocity update equations to construct the following position update equation to prevent local entrapment: Among them, x k is the optimized individual k position information, To optimize the position information of the previous individual k, r x is a random number between [0,1], N k is the index of the eight adjacent individuals of individual k, x n is the location information of the adjacent individual n of individual k.

Images