Unmanned cluster motion control method based on Boids model and navigator follower model

Abstract

The invention discloses an unmanned cluster control method based on a Boids model and a navigator follower model. According to the method, a Boids cluster model and the navigator follower model are improved. The method is mainly used for modifying three motion principles in a Boids model. Firstly, on the basis that the alignment principle is removed, the separation principle is reserved to ensure the spatial distance between individuals. And secondly, aiming at the aggregation principle of the cluster, the center of the cluster is no longer taken as an aggregation target, but the aggregation directions of other followers are guided by selecting a navigator in the cluster, so that the dynamic behavior of the cluster has guidance through the adjustment. And finally, in order to enhance the autonomous obstacle avoidance capability of the individuals in the cluster, an obstacle avoidance principle is introduced to ensure that the individuals in the cluster can effectively avoid the obstacles when encountering the obstacles, so that the adaptability of the whole cluster to the environment is improved. Finally, the unmanned clustering method provided by the invention solves the problem that a traditional navigator follower model excessively depends on a navigator.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned cluster control algorithms, and in particular to an unmanned cluster control method based on a Boids model and a leader-follower model, and is mainly an invention for motion control of followers in a cluster. Background Art

[0002] As an important algorithm, cluster algorithm has been applied to all aspects of society. Cluster algorithm is widely used to solve various optimization problems, such as function optimization, combinatorial optimization, multi-objective optimization, etc. By simulating the behavior rules of biological groups, cluster algorithm can quickly search the solution space and find the optimal solution to the problem; in the fields of robot path planning and unmanned driving, cluster algorithm can be used to find the optimal path while avoiding collision and congestion; cluster algorithm can be used for multi-agent system control, that is, to control the behavior of multiple agents so that they can work together and complete tasks. For example, in the fields of drone formation flight and intelligent sensor networks, cluster algorithm can realize information sharing and collaboration between agents. In addition, in the field of digital media such as games and movies, the research of cluster algorithm is also a focus. The main characteristics of cluster algorithm are parallelism, self-organization and robustness. Parallelism enables individuals in the cluster to search different areas of the solution space at the same time, which improves the search efficiency; self-organization allows individuals to form complex global behaviors through local interactions and simple rules without central control; finally, robustness reduces the sensitivity of the algorithm to initial parameters and noise.

[0003] Swarm intelligence algorithm is a kind of biological heuristic algorithm, which originated from the study of the behavior of biological groups in nature. These biological groups, such as bird flocks, fish schools, ant colonies, etc., show a high degree of organization and coordination in activities such as foraging, migration, and social interaction. Cluster algorithms mainly simulate the collective movement behavior of biological groups, including particle swarm optimization (PSO), ant colony optimization (ACO), artificial fish swarm algorithm (AFSA), bird swarm algorithm (BSA), etc. These algorithms search the solution space and find the optimal solution to the problem by simulating the behavioral rules of biological groups, such as foraging, clustering, and tail-chasing. Among the cluster algorithms, the leader-follower algorithm is a very classic algorithm, but it has the disadvantage of lacking clear feedback control. Once the leader behaves incorrectly, the entire system will be paralyzed; each individual in the Boids model calculates its individual behavior alone, so the Boids model is used to calculate the behavior of the followers in the leader-follower model, which can improve the pressure of the leader in the leader-follower model and improve the stability of the system. Summary of the invention

[0004] The problem to be solved by the present invention is: the problem of over-reliance on the leader in the leader-follower model, and a motion control method for the follower. Specifically, the unmanned swarm is controlled by combining the Boids model and the leader-follower model, thereby avoiding the paralysis of the system when the leader makes an erroneous behavior and improving the operating efficiency of the system.

[0005] The present invention adopts the following technical solution: combining the Boids model and the leader-follower model to control the movement of the cluster, and performing the following steps for the individuals in the unmanned cluster to complete the speed, size and direction control of the individuals in the unmanned cluster:

[0006] Step S1: First, a navigator in the unmanned cluster is selected, and a path is set for the navigator. The path is a set of coordinates of multiple two-dimensional path points.

[0007] Step S2: For each follower in the unmanned cluster, a different offset value is set, and the coordinates of the point that the follower should follow are set using the offset value and the position of the leader itself.

[0008] Step S3, calculate the convergence vector, separation vector and obstacle avoidance vector. The convergence vector is determined by the position of the follower and the follow point. This vector ensures that the followers in the unmanned cluster can follow the leader. The separation vector ensures that there will be no collision between individuals in the unmanned cluster. It is calculated by the position of the follower itself and other individuals. The obstacle avoidance vector can ensure that the follower avoids obstacles. It is calculated by the detection results of the ray sensor equipped on the follower. After the calculation is completed, these three vectors are synthesized in different ways, and the direction of the new vector is used as the direction of the follower's velocity vector.

[0009] Step S4, calculating the magnitude of the follower's speed. If both the separation vector and the obstacle avoidance vector calculated in step S3 exist or only one exists, the speed will be adjusted to the minimum to ensure that the follower can correctly avoid obstacles and other individuals.

[0010] If there is neither a separation vector nor an obstacle avoidance vector, the distance between the follower and the follow point is calculated.

[0011] If the distance is greater than the falling behind distance, the speed will be set to the maximum value to prevent the follower from falling behind;

[0012] If the distance is less than the lagging distance, the speed is calculated according to the interpolation function. The closer the distance between the follower and the following point, the slower the speed. However, it will not be less than the minimum speed, and vice versa. This speed adjustment ensures that the formation of the entire cluster team remains stable to the greatest extent.

[0013] Step S5: Set the velocity vector of the follower in combination with the velocity direction and velocity magnitude.

[0014] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0015] 1. The present invention provides a motion control method for unmanned clusters. To solve the problem of excessive burden on the leader when using a leader-follower model for unmanned clusters, the motion control of the follower in the model is handed over to the follower to control it independently. The only communication between the follower and the navigator is that the navigator informs the follower of its own position during the movement, which reduces the burden on the leader and enhances the stability of the cluster.

[0016] 2. The motion control method for unmanned clusters of the present invention makes certain modifications to the classic cluster model Boids model. First, on the basis of removing the alignment principle, the separation principle is retained to ensure the spatial distance between individuals. Secondly, for the aggregation principle of the cluster, the center of the cluster is no longer used as the aggregation target, but the aggregation direction of other followers is guided by selecting the leader in the cluster. This adjustment makes the dynamic behavior of the cluster more flexible and instructive. Finally, in order to enhance the autonomous obstacle avoidance capability of individuals in the cluster, the obstacle avoidance principle is introduced to ensure that individuals in the cluster can effectively avoid obstacles when encountering obstacles, thereby improving the adaptability of the entire cluster to the environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of the unmanned cluster control method based on the leader-follower model and the Boids model of the present invention;

[0018] Figure 2(a) is a schematic diagram of the aggregation vector calculation in the Boids model;

[0019] Figure 2(b)-Figure 2(d) It is the detection principle diagram of the sensor;

[0020] Figure 2(e) is a diagram showing the calculation principle of the separation vector in the Boids model;

[0021] Figure 3(a)-Figure 3(c) This is a simulation diagram of the normal vector calculation of unmanned swarm control based on the Boids model and the leader-follower model;

[0022] Figure 3(d) is the final experimental simulation diagram of the unmanned swarm control method based on the leader-follower model and the Boids model. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the application is further elaborated in detail below in conjunction with the accompanying drawings. The described embodiments are only a part of the embodiments involved in the present invention. All non-innovative embodiments of other researchers in the field on this embodiment belong to the protection scope of the present invention. At the same time, for the step numbering in the embodiment of the present invention, it is only set for the convenience of explanation, and the order between the steps is not limited in any way. The execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.

[0024] like Figure 1 As shown, an example of the present invention provides a control method for an unmanned cluster based on a Boids model and a leader and follower model, and performs the following steps S1 to S5 to complete the motion control of the leader and follower in the cluster:

[0025] Step S1: setting a path consisting of multiple two-dimensional path points for the navigator, which path is generally planned by a path planning algorithm.

[0026] Step S2: For each follower in the cluster, set a different offset value as the target point of the follower during the cluster movement:

[0027]

[0028] In the above formula, p l are the coordinates of the pilot, p f is the coordinate of the follower's follow point, where x e and e is the offset. The follow points should be distributed behind the leader, so y should be greater than 0. In addition, the coordinates of the follow points of different followers should not be the same.

[0029] Step S3: According to the Boids model, first calculate the aggregation vector, as shown in Figure 2(a). The coordinates of the follower point are obtained from S2, and the coordinates of the follower are set to p = (x p ,y p ), then the aggregation vector The calculation of is as follows:

[0030]

[0031] In order to avoid collision between the follower and the obstacle, as shown in Figure 2(b), four ray detection sensors are installed on the follower. The length sensors are located in the front half of the follower, where sensors A and B form a group, and sensors C and D form a group. Sensors in the same group will emit rays at the same position, where the ray directions B and C are along the follower's forward direction. The ray emission direction of sensor A is the follower's forward direction offset to the left by θ degrees, and the ray emission direction of sensor D is the follower's forward direction offset to the right by θ degrees. In order to prevent interference between sensors, only one sensor's detection result will be valid during sensor detection. As shown in Figure 2(c), when an obstacle is detected by both sensors A and B, only the detection result of sensor B is taken; as shown in Figure 2(d), when an obstacle is detected by four sensors at the same time, only the detection result of sensor B is taken. If a sensor detects an obstacle, assume that the coordinates of the hit point are p h =(x h ,y h ), the follower coordinate is p=(x p ,y p ), the position coordinate of the sensor is s=(x s ,y s ). First, we can obtain the following variables:

[0032]

[0033] in, It represents the vector formed by the coordinates of the hit point and the follower position. It represents the vector formed by the sensor position and the follower position. is the direction vector of the follower's forward direction, and y represents and The angle between and The angle between the obstacle avoidance vector The calculation is mainly done by rotating The result is as follows:

[0034]

[0035] in The multiplication is a rotation matrix. When the detection results from different sensors are judged to be valid, The vectors are multiplied by different rotation matrices to achieve specific rotation transformations. When the detection results of sensors B and C are valid, the obstacle is close to the follower. At this time, the degree of rotation of the vector is large, which is 2γ+ω. This ensures that the follower can move away from the obstacle as quickly as possible. The direction of rotation is also different. For sensor B, the direction of rotation is clockwise, and for sensor D, the direction of rotation is counterclockwise. Conversely, when the detection results of sensor A are valid, It should be rotated clockwise by 2γ-ω. When the detection result of sensor D is valid, It should rotate 2γ-ω counterclockwise. When the detection of sensors A and D is effective, the follower has enough space to avoid obstacles. In this way, avoiding obstacles at a small angle can ensure the stability of the cluster formation as much as possible.

[0036] In order to avoid collisions between individuals in the cluster, the separation vector is calculated according to the separation principle in the Boids model. As shown in Figure 2(e), the position of the follower is p = (x p ,y p ) is the center of the circle and r is the radius for range detection. If no other cluster individuals are detected in the range, then the separation vector does not exist; if the remaining cluster individuals are detected and the positions of these individuals are stored in a set, then the separation vector The calculation of is as follows:

[0037]

[0038] In the above formula, U represents the two-dimensional position coordinate set of all other individuals in the detection range. The final separation vector can be obtained by synthesizing the vectors composed of each coordinate and the follower coordinates and then inverting them.

[0039] For the above aggregation vector Separation vector and obstacle avoidance vector After normalization and synthesis, the velocity direction vector of the final follower can be obtained. This process uses different calculation methods to achieve flexible and accurate control according to different scenarios and needs, as follows:

[0040] When there is no need to avoid obstacles and other cluster individuals, the final velocity direction vector is the direction of the normalized aggregation vector, as shown below:

[0041]

[0042] If you need to avoid obstacles or other individuals in the cluster individually, the priority of the aggregation will be reduced to ensure that individuals in the cluster do not hit obstacles or internal collisions. The calculation method is as follows

[0043]

[0044] In the above formula, α and β represent the aggregation vectors and obstacle avoidance vector where α is less than 1 and β is greater than 1.

[0045] When you only need to avoid other individuals in the group, the velocity direction vector The calculation method is as follows

[0046]

[0047] In the above formula, α and γ represent the aggregation vectors and separation vector where α is less than 1 and γ is greater than 1.

[0048] If obstacle avoidance and separation are required at the same time, avoiding obstacles and avoiding other individuals in the cluster have the same priority, and the velocity direction vector The calculation method is as follows

[0049]

[0050] In the above formula, α, β and γ represent the aggregation vectors Obstacle Avoidance Vector and separation vector The weights are α less than 1, β greater than 1, and γ greater than 1.

[0051] Step S4: If one or both of the follower's separation vector and obstacle avoidance vector exist, the follower will give priority to avoiding obstacles or other individuals in the cluster. At this time, the speed of the follower will be set to the minimum speed to ensure the safety of the cluster. If the follower does not need to avoid obstacles or other individuals in the cluster, the speed of the follower is related to the distance from the corresponding follow point, as shown in the following formula:

[0052]

[0053] In the above formula (x f ,y f ) is the coordinate of the follower’s follow point, (x p ,y p ) is the coordinate of the follower, then the velocity is calculated as follows:

[0054]

[0055] As shown in the above formula, v represents the actual speed of the follower. When d is less than r, it means that there are other followers within the safety range of the follower. At this time, according to step S51, the system first allows the follower to avoid collision. At this time, v takes the minimum value v min ; If d is greater than r and less than d max When , it indicates that the follower is at a certain distance from the follow point. At this time, the interpolation function is used according to d and d max The relationship between d and r is used to interpolate to obtain a reasonable speed; if d is greater than d max , the follower is too far away from the follow point and should move at the maximum speed v max The vehicle is chasing the follower to prevent it from falling behind, and the final speed is calculated accordingly.

[0056] Step S5: Finally, the velocity vector of the follower is set in combination with the required velocity direction and velocity magnitude.

[0057] As shown in Figures 3(a), 3(b), 3(c), and 3(d), they are simulation experiment diagrams of the unmanned cluster control method of the present invention combining the Boids model and the leader-follower model. Figure 3(a) is the calculation of the aggregation vector of the follower individual in the unmanned cluster, the blue car in the figure is the follower, the white one is the leader, the white ball is the follow point followed by the follower, and the red line is the aggregation vector; Figure 3(b) is the calculation of the obstacle avoidance vector of the follower, the brown object in the figure is the obstacle, the blue car is the follower, the white line is the ray emitted by the sensor, and the red line is the obstacle avoidance vector of the follower; Figure 3(c) is the calculation result of the obstacle avoidance vector of the follower, the two cars in the figure are followers, and the red line represents the obstacle avoidance vector; Figure 3(d) is the final cluster motion result according to the method of the present invention, the white car in the figure is the leader, the two cars behind it are followers, the three brown objects are obstacles, the red line is the motion trajectory of the leader, and the yellow line and the green line are the motion trajectories of the followers.

[0058] It can be seen from the simulation diagram that the present invention enables the entire unmanned cluster to reach the end point in a relatively stable formation.

Claims

1. An unmanned swarm control method based on the Boids model and the leader-follower model, characterized in that The three key motion vectors in Boids are adjusted in combination with the leader-follower model. By giving the followers greater autonomy, they can move on their own according to the actual situation. The following steps S1 to S5 are executed to complete the motion control of the followers in the cluster. Step S1: Select a navigator in the cluster and set a path for the navigator, which path consists of multiple two-dimensional path points. Step S2: For each follower in the cluster, a following point is set behind the leader, and different followers have different following points. Step S3, calculate the aggregation vector, separation vector and obstacle avoidance vector. The calculation of the aggregation vector is different from the traditional Boids model calculation method. Here, the aggregation vector is calculated by the position of each follower and the position of its following point. Since the following points of different followers are inconsistent, this calculation method enables the followers to aggregate around the leader more flexibly; the separation vector is calculated by the positions of other individuals in the cluster and the position of the follower itself. This vector prevents the follower from colliding with other individuals in the cluster; in order to calculate the obstacle avoidance vector, each follower is equipped with a ray sensor. These sensors detect obstacles in the surrounding environment through a ray detection mechanism, and then calculate the obstacle avoidance vector based on the detection results. This measure effectively improves the follower's autonomous navigation and obstacle avoidance capabilities in complex environments, ensuring that it can safely avoid obstacles during movement. Step S4, calculating the magnitude of the follower's speed. The present invention designs a strategy that prioritizes obstacle avoidance and collision avoidance, thereby avoiding the problem of over-reliance on parameter selection in the traditional Boids model. Specifically, if both the separation vector and the obstacle avoidance vector calculated in step S3 exist or only one exists, the magnitude of the speed will be set to the minimum value to ensure that the follower can avoid obstacles and other individuals. If neither the separation vector nor the obstacle avoidance vector exists, first specify a parameter, the falling behind distance, which refers to the maximum distance allowed for the follower to be away from the following point in the cluster system, and then calculate the distance between the follower and the following point. If the distance is greater than the falling behind distance, the speed will be set to the maximum value to prevent the follower from falling behind; If the distance is less than the lagging distance, the speed is calculated according to the interpolation function. The closer the distance between the follower and the following point, the slower the speed, but it will not be less than the minimum speed, and vice versa. This speed adjustment strategy ensures that the formation of the entire cluster team remains stable. Step S5: Set the movement state of the follower in combination with the speed and direction calculated in the previous steps.

2. The unmanned swarm control method based on the Boids model and the leader-follower model according to claim 1 is characterized in that: In step S1, a leader is selected in the cluster, and the leader will move along a pre-set path. The path consists of multiple path points, each of which is a two-dimensional coordinate point. In addition, the path point should have a certain distance from the edge of the obstacle, so as to give the follower enough space to avoid the obstacle.

3. The unmanned swarm control method based on the Boids model and the leader-follower model according to claim 1, characterized in that: In step S2, in a cluster, there are multiple followers, and the following points of these followers are not the same. As the navigator moves, the coordinates of the following point will also change. Compared with the traditional navigator-follower model, the only communication between the navigator and the followers in the present invention is that the navigator sends the coordinates of the following point to the followers, thereby reducing the communication burden of the navigator.

4. The unmanned swarm control method based on the Boids model and the leader-follower model according to claim 1, characterized in that: In step S3, according to the aggregation principle in the Boids model, each individual in the cluster wants to get close to the cluster center, but in the present invention, the follower individual in the cluster wants to get close to the follow point. Since there is a certain distance between different follow points, this measure improves the collision avoidance ability of the cluster. Take a follower individual in the cluster as an example, its position is p = (x p ,y p ), the coordinates of the following point f=(x f ,y f ), aggregation vector The expression is as follows According to the separation principle in the Boids model, in order to avoid collisions between individuals in the cluster, a safe distance is set for the follower, with this distance as the radius and the follower's position as the center for range detection. If there are no other cluster individuals in the range, it is considered that the follower does not need to be separated; otherwise, the positions of the remaining individuals are recorded into a set, and then the separation vector is calculated in combination with its own position, as shown in the following formula: In the above formula, U represents the set of all cluster individuals that meet the conditions. Each cluster individual in the set is represented in the form of coordinates, that is, the i-th individual is represented as (x i ,y i ), subtract the coordinates of the follower from this coordinate, then add up all the calculation results, and finally negate the result to get the separation vector In order to realize the autonomous obstacle avoidance function of the follower during movement, the follower is equipped with four sensors: A, B, C and D. These sensors are located in different positions to provide all-round obstacle detection capabilities. Specifically, sensors A and B are located in the upper left corner of the follower. The detection direction of sensor A is consistent with the forward direction of the follower, while the detection direction of sensor B is offset θ degrees to the left from the direction of A. Correspondingly, sensors C and D are located in the upper right corner of the follower. The detection direction of C is the same as the forward direction, and the detection direction of D is offset θ degrees to the right from the forward direction. When dealing with the situation where multiple sensors detect obstacles at the same time, a priority system is adopted. If two sensors in the same group (such as A and B, or C and D) detect an obstacle at the same time, we will give priority to the detection result of the sensor with a higher priority in the group (for example, A is preferred between A and B) and discard the other one. If the obstacle is detected by three or four sensors at the same time, we will filter according to the priority: among the four sensors, A and B have higher priority than C and D. If A and B are detected at the same time, the result of A is preferred. This design strategy is to avoid mutual interference between sensors. Taking the sensor group A and B as an example, if a sensor detects an obstacle, the ray emitted by the sensor hits the obstacle, and the position of the sensor is s = (x s ,y s ), the coordinates of the impact point are h=(x h ,y h ), the position of the follower is p = (x p ,y p ), the follower’s forward direction vector is First, we can find: in, represents the vector formed by the coordinates of the point where the ray sensor hits the obstacle and the position of the follower, and γ represents The angle between the follower's forward vector and represents the vector formed by the sensor position and the follower position, and ω represents The angle between the follower's forward vector and the obstacle avoidance vector is calculated mainly by rotating It is found that the calculation of obstacle avoidance vectors is inconsistent when different sensors detect obstacles. When sensor A or sensor C detects an obstacle, it indicates that the distance between the obstacle and the follower is too small. At this time, it is necessary to increase On the contrary, when sensor B or sensor D detects an obstacle, the obstacle is still some distance away from the follower. The rotation angle can be smaller than when sensor A or sensor C detects an obstacle. In addition, the rotation direction is not consistent. When sensor A and sensor B detect an obstacle, It should rotate clockwise, otherwise counterclockwise. In summary, the calculation is as follows: As shown in the above formula, when different sensors detect obstacles, they first pass Multiply it by a rotation matrix to get the final obstacle avoidance vector For the above aggregation vector Separation vector and obstacle avoidance vector After synthesis, the velocity direction vector of the final follower can be obtained This process uses different calculation methods to achieve flexible and accurate control according to different scenarios and needs, as follows: When there is no need to avoid obstacles and other cluster individuals, the final velocity direction vector is the direction of the normalized aggregation vector, as shown below: If you need to avoid obstacles or other cluster individuals individually, the priority of the aggregation will be reduced to ensure that the individuals in the cluster will not hit obstacles or have internal collisions. You only need to avoid obstacles when the speed direction vector The calculation method is as follows In the above formula, α and β represent the aggregation vectors and obstacle avoidance vector where α is less than 1 and β is greater than 1. When you only need to avoid other individuals in the group, the velocity direction vector The calculation method is as follows In the above formula, α and γ represent the aggregation vectors and separation vector where α is less than 1 and γ is greater than 1. If obstacle avoidance and separation are required at the same time, avoiding obstacles and avoiding other individuals in the cluster have the same priority, and the velocity direction vector The calculation method is as follows In the above formula, α, β and γ represent the aggregation vectors Obstacle Avoidance Vector and separation vector The weights are α less than 1, β greater than 1, and γ greater than 1. The velocity direction vector is calculated differently in different situations This allows followers to avoid obstacles and collisions with other individuals in the cluster, while also maintaining the stability of the entire cluster formation.

5. The unmanned swarm control method based on the Boids model and the leader-follower model according to claim 1, characterized in that: In step S4, the speed of the followers in the unmanned cluster is controlled by an interpolation function. Different from the traditional Boids model method for controlling the speed, the present invention uses an interpolation function to control the speed, which allows the speed of the followers in the cluster to be adjusted according to the distance between their own position and the follow point. Specifically, the following parameters are set for each follower: Maximum speed v max and minimum speed v min , the radius r of the follower's collision avoidance range detection and the follower's maximum acceleration distance d max , d max Indicates the maximum distance allowed between the follower and the follow point. First, it is necessary to determine whether the follower needs to perform obstacle avoidance or separation. If necessary, the follower's speed should be immediately adjusted to the minimum speed v min ; If neither is needed, then perform the following steps: Find the distance d between the follower and the follow point, and calculate the distance d according to d and r and d max The speed of the follower is calculated based on the relationship as follows: Then the velocity v is calculated as follows: As shown in the above formula, v represents the actual speed of the follower, (x f ,y f ) is the coordinate of the follow point, (x p ,y p ) is the coordinate of the follower. When d is less than r, it means that other followers appear within the safety range of the follower. At this time, the system will first let the follower perform a separation operation, and v will be immediately adjusted to the minimum value; if d is greater than r and less than d max When , it indicates that the follower is at a certain distance from the follow point. At this time, the interpolation function is used according to d and d max The relationship between d and r is used to interpolate to obtain a reasonable speed; if d is greater than d max , the distance between the follower and the follow point is too large, and the maximum speed v max Chase to prevent the follower from separating from the entire team.

6. The unmanned swarm control method based on the Boids model and the leader-follower model according to claim 1, characterized in that: The direction and magnitude of the final velocity vector of the follower are determined in step S4.

Images