Unmanned aerial vehicle cluster control method based on artificial intelligence

Abstract

The invention relates to the technical field of unmanned aerial vehicle control, and discloses an unmanned aerial vehicle cluster control method based on artificial intelligence. According to the method, a preliminary flight path is generated, a laser radar, a camera and a temperature sensor are used for collecting surrounding environment information, collected data are cleaned and formatted, a Kalman filtering technology is used for integrating data from different unmanned aerial vehicles and sensors, and a deep learning algorithm is used for making a decision. Calculating an optimal flight path, calculating an obstacle avoidance path, a flight acceleration and a flight steering angle by applying an RRT planning algorithm according to environment data and unmanned aerial vehicle positions, designing a cluster control strategy based on collected and calculated data, ensuring cooperative work of the unmanned aerial vehicles, forming an effective cluster formation, and according to a decision-making result, determining the unmanned aerial vehicle position according to the decision-making result. And calculating speed, direction and height control instructions of each unmanned aerial vehicle, transmitting the control instructions to each unmanned aerial vehicle in real time, monitoring the execution condition of the unmanned aerial vehicle, and tracking the state of the unmanned aerial vehicle.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle control, and in particular to an unmanned aerial vehicle cluster control method based on artificial intelligence. Background Art

[0002] A drone is an unmanned aircraft that can fly through remote control or autonomous flight programs. Drones are usually equipped with sensors, cameras, GPS, etc., and can perform a variety of tasks without direct human control. According to their structure and purpose, drones can be divided into many types, including fixed-wing drones, rotary-wing drones, and hybrid-wing drones. Compared with traditional aircraft, drones usually have lower R&D, production, and operating costs, making them suitable for large-scale applications. They are small in size, can fly flexibly in complex or restricted environments, and perform various tasks. They can replace humans in dangerous environments, such as search and rescue at fire scenes, earthquake areas, and explosions. At the same time, drones can cover a wide area, be quickly deployed, and collect data and information in real time to improve decision-making efficiency.

[0003] As the number of drones increases, the calculation and optimization process may require a lot of time and resources, resulting in a significant increase in the complexity of collaborative control and path planning of drone clusters, thereby reducing the efficiency of drone operations. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the shortcomings of the prior art, the present invention provides a method for controlling a drone cluster based on artificial intelligence, which has the advantages of first generating a preliminary flight path, then using a laser radar, a camera and a temperature sensor to collect surrounding environment information, obtaining the status information of each drone in real time, and sharing information between drones, cleaning and formatting the collected data, using Kalman filtering technology to integrate data from different drones and sensors to obtain more comprehensive environmental and status information, applying a deep learning algorithm to make decisions, calculating the optimal flight path Fxlj, and using an RRT planning algorithm to calculate the obstacle avoidance path Bzlj, the flight acceleration Fxjd and the flight steering angle Fzxj according to the environmental data and the drone position. Based on the collected and calculated data, a cluster control strategy is designed to ensure that the UAVs work together to form an effective cluster formation. According to the decision results, the control instructions of each UAV are calculated, including the flight speed instruction Sdzl, the flight direction instruction Fxzl and the flight altitude instruction Gdzl. Dynamic adjustments are made for emergencies, and the control instructions are transmitted to each UAV in real time. The execution of the UAV is monitored and its status is tracked. By analyzing the relevant information of the UAV, the Kalman filter, deep learning algorithm and RRT planning algorithm are combined to calculate the optimal path of the UAV and generate the optimal control instructions, which reduces the complexity of collaborative control and path planning of the UAV cluster, thereby improving the efficiency of UAV operation and solving the above problems.

[0006] (II) Technical solution

[0007] To achieve the above object, the present invention provides the following technical solution: a drone cluster control method based on artificial intelligence, comprising the following steps:

[0008] S1. Clarify the safety, efficiency and mission requirements of the drone and generate a preliminary flight path;

[0009] S2. Collect data by using laser radar, camera and temperature sensor to collect information about the surrounding environment, including the location of obstacles, terrain data and climate conditions, obtain the status information of each drone in real time, including location, speed, altitude, battery power and attitude, and share information between drones;

[0010] S3, clean and format the collected data, and use Kalman filtering technology to integrate data from different drones and sensors to obtain more comprehensive environmental and status information;

[0011] S4. Apply deep learning algorithms to make decisions and calculate the optimal flight path Fxljj. According to environmental data and the position of the UAV, use the RRT planning algorithm to calculate the obstacle avoidance path Bzlj, flight acceleration Fxjd and flight steering angle Fzxj. At the same time, design a cluster control strategy based on the collected and calculated data to ensure that the UAVs work together to form an effective cluster formation.

[0012] S5. Calculate the control instructions for each drone based on the decision results, including the flight speed instruction Sdzl, the flight direction instruction Fxzl, and the flight altitude instruction Gdzl, and make dynamic adjustments for emergencies;

[0013] S6, transmits control instructions to each UAV in real time, monitors the execution of the UAV, and tracks its status.

[0014] Preferably, S3 cleans and formats the collected data, wherein the position coordinates are unified into a WGS84 coordinate system.

[0015] Preferably, S3 cleans and formats the collected data, and integrates the data from different drones and sensors using Kalman filtering technology. The Kalman filtering formula is expressed as follows:

[0016] x k|k-1 =F k *x k-1|k-1 +B k *u k +w k

[0017] In the formula, x k|k-1 represents the state estimation at time k based on time k-1, F k represents the state transition matrix, describing how the system migrates from time k-1 to time k, x k-1|k-1 represents the state estimate at time k-1, B k represents the control input matrix, u k Indicates the control instruction of external input, W k represents process noise.

[0018] Preferably, the S4 applies a deep learning algorithm to make a decision and calculate the optimal flight path Fxlj, and the calculation formula is as follows:

[0019] Fxlj(s t , a t )=r t +γ*max Fxlj(s t+1 , a′)

[0020] In the formula, Fxlj(s t , at ) means in state s t Take action a t The expected return, r t represents the immediate reward obtained at time t, γ represents the discount factor, which ranges from 0 to 1 and represents the importance of future rewards, max Fxlj(s t+1 , a′) represents the maximum expected return of the next state, and continuously optimizes Fxlj(s t , a t ) function, calculates the best action under a given state to guide the flight path selection of the drone.

[0021] Preferably, the S4 uses a deep learning algorithm to make decisions, and uses an RRT planning algorithm to calculate the obstacle avoidance path Bzlj according to the environmental data and the position of the drone. The calculation formula is as follows:

[0022]

[0023] In the formula, Bzlj represents the obstacle avoidance path, Zjjd represents the nearest node in the RRT algorithm tree, Sjcy represents a target point randomly sampled in the RRT algorithm tree, Δt represents the step size, that is, the distance extended to the target point in the tree, and ||Sjcy-Zjjd|| represents the norm of the vector.

[0024] Preferably, the S4 uses a deep learning algorithm to make decisions, and uses an RRT planning algorithm to calculate the flight acceleration Fxjd according to environmental data and the position of the drone. The calculation formula is as follows:

[0025]

[0026] In the formula, Fxjd represents the flight acceleration, F net It represents the net external force on the drone, m represents the mass of the drone, and g represents the acceleration due to gravity, which is 9.81m / s 2 .

[0027] Preferably, the S4 applies a deep learning algorithm to make a decision, and uses an RRT planning algorithm to calculate the flight steering angle Fzxjj according to the environmental data and the position of the drone. The calculation formula is as follows:

[0028] Fzxj=k*atan2(sin(θ target -θ current ), cos(θ target -θ current ))

[0029] In the formula, Fzxj represents the flight steering angle, k represents the gain k, which is used to smooth the steering, and θ target represents the target heading, θcurrent Indicates the current heading.

[0030] Preferably, the S5 calculates the control instructions for each UAV according to the decision result, wherein the calculation formula of the flight speed instruction Sdzl is as follows:

[0031] Sdzl=V0+k v *(Mbsd-V0)+f v

[0032] In the formula, Sdzl represents the flight speed instruction, V0 represents the current flight speed, k v Represents the speed gain coefficient, between 0 and 1, used for smooth transition, f v It represents the impact factor of environmental factors on speed, and Mbsd represents the target speed.

[0033] Preferably, the S5 calculates the control instructions for each UAV according to the decision result, wherein the calculation formula of the flight direction instruction Fxzl is as follows:

[0034] Fxzl=θ0+k θ *Δθ

[0035] In the formula, Fxzl represents the flight direction instruction, θ0 represents the current heading angle, kθ represents the directional gain coefficient, controls the adjustment speed, and Δθ represents the angle difference between the target heading and the current heading.

[0036] Preferably, the S5 calculates the control instructions for each UAV according to the decision result, wherein the calculation formula of the flight altitude instruction Gdzl is as follows:

[0037] Gdzl=h0+k h *(h target -h0)+f h

[0038] In the formula, Gdzl represents the flight altitude instruction, h0 represents the current flight altitude, and h target represents the target flight altitude, k h represents the height gain coefficient, f h Represents the impact of environmental factors on height control.

[0039] Compared with the prior art, the present invention provides a drone cluster control method based on artificial intelligence, which has the following beneficial effects:

[0040] The present invention first generates a preliminary flight path, then uses laser radar, camera and temperature sensor to collect surrounding environment information, obtains the status information of each drone in real time, and shares information between drones, cleans and formats the collected data, uses Kalman filtering technology to integrate data from different drones and sensors, obtains more comprehensive environment and status information, applies deep learning algorithm to make decisions, calculates the optimal flight path Fxljj, and uses RRT planning algorithm to calculate obstacle avoidance path Bzlj, flight acceleration Fxjd and flight steering angle Fzxj according to environmental data and drone position. At the same time, based on the collected and calculated data, the proposed method designs a flight path based on the collected and calculated data. The cluster control strategy ensures that the UAVs work together to form an effective cluster formation. According to the decision results, the control instructions for each UAV are calculated, including the flight speed instruction Sdzl, the flight direction instruction Fxzl, and the flight altitude instruction Gdzl. Dynamic adjustments are made to emergencies, and the control instructions are transmitted to each UAV in real time. The execution of the UAV is monitored and its status is tracked. By analyzing the relevant information of the UAV, the Kalman filter, deep learning algorithm and RRT planning algorithm are combined to calculate the optimal path of the UAV and generate the optimal control instructions, which reduces the complexity of the coordinated control and path planning of the UAV cluster, thereby improving the efficiency of the UAV operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a schematic diagram of the steps of the method of the present invention. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] With the increase in the number of drones, the calculation and optimization process may take a lot of time and resources, resulting in a significant increase in the complexity of coordinated control and path planning of drone clusters, thereby reducing the efficiency of drone operations. Therefore, a drone cluster control method based on artificial intelligence is proposed. Figure 1 , the method comprises the following steps:

[0044] S1. Clarify the safety, efficiency and mission requirements of the drone and generate a preliminary flight path;

[0045] S2. Collect data by using laser radar, camera and temperature sensor to collect information about the surrounding environment, including the location of obstacles, terrain data and climate conditions, obtain the status information of each drone in real time, including location, speed, altitude, battery power and attitude, and share information between drones;

[0046] Sensors will efficiently identify and capture the location of obstacles, terrain features, and climate conditions, such as wind speed, temperature, humidity, and other data, to ensure the safe flight of drones in complex environments. In addition, the status information of each drone needs to be monitored in real time, including its location, speed, altitude, battery power, and attitude. This status information is crucial for path adjustment, fault warning, and safety assurance. At the same time, in order to enhance the intelligence and response capabilities of the cluster, an information sharing mechanism is established between different drones to ensure that each drone can access each other's real-time data. Through this mutual exchange and sharing of information, the collaborative ability of the cluster can be effectively improved, autonomous decision-making adjustments can be achieved in a dynamic environment, and unnecessary risks and waste of resources can be reduced;

[0047] S3, clean and format the collected data, and use Kalman filtering technology to integrate data from different drones and sensors to obtain more comprehensive environmental and status information;

[0048] During the data collection process, different types of sensors may be affected by noise, interference or errors, resulting in reduced data accuracy and reliability. Therefore, the original data needs to be cleaned to remove invalid or erroneous information. This process uses technologies such as data deduplication, outlier detection and missing value filling to ensure that the final data set is clean and consistent. The cleaned data is then formatted and the location coordinates are unified into the WGS84 coordinate system to meet the requirements of subsequent analysis and processing. At the same time, the timestamp and data type of the data are unified so that data from different sources can be compatible with each other and lay the foundation for subsequent fusion and analysis.

[0049] After data cleaning and formatting, Kalman filtering technology is used to integrate data from different drones and their sensors. This technology is particularly effective in dynamic systems and can estimate and predict system states through mathematical models. The Kalman filtering formula is expressed as follows:

[0050] x k|k-1 =F k *x k-1|k-1 +B k *u k +w k

[0051] In the formula, x k|k-1 represents the state estimation at time k based on time k-1, Fk represents the state transition matrix, describing how the system migrates from time k-1 to time k, x k-1|k-1 represents the state estimate at time k-1, B k represents the control input matrix, u k Indicates the control instruction of external input, W k Represents process noise. Kalman filtering can not only smooth state information such as position and speed, but also effectively fuse measurement data from various sensors, such as the environment-terrain information provided by the lidar, the visual data captured by the camera, and the climate state of the temperature sensor. All of these can be weighted and synthesized through Kalman filtering to generate a comprehensive state estimate.

[0052] S4. Apply deep learning algorithms to make decisions and calculate the optimal flight path Fxlj. According to environmental data and the position of the UAV, use the RRT planning algorithm to calculate the obstacle avoidance path Bzlj, flight acceleration Fxjd and flight steering angle Fzxj. At the same time, design a cluster control strategy based on the collected and calculated data to ensure that the UAVs work together to form an effective cluster formation.

[0053] Among them, the optimal flight path calculation formula is as follows:

[0054] Fxlj(s t , a t )=r t +γ*max Fxlj(s t+1 , a′)

[0055] By calculating the best flight path, the drone can reach the target location in a shorter time. In the formula, Fxlj(s t , a t ) means in state s t Take action a t The expected return, r t represents the immediate reward obtained at time t, γ represents the discount factor, which ranges from 0 to 1 and represents the importance of future rewards, max Fxlj(s t+1 , a′) represents the maximum expected return of the next state, and continuously optimizes Fxlj(s t , a t ) function, calculates the best action under a given state to guide the drone's flight path selection. The AI ​​algorithm can analyze environmental data and perform dynamic path optimization based on real-time climate and obstacle information, thereby significantly improving flight efficiency, which not only reduces flight time but also reduces energy consumption in the air;

[0056] The obstacle avoidance path calculation formula is as follows:

[0057]

[0058] In complex environments, drones face the risk of various obstacles such as buildings, wires or other aircraft. By calculating the obstacle avoidance path, AI can identify these obstacles in real time and quickly adjust the flight path to avoid collisions. In the formula, Bzlj represents the obstacle avoidance path, Zjjd represents the nearest node in the RRT algorithm tree, Sjcy represents a randomly sampled target point in the RRT algorithm tree, Δt represents the step size, that is, the distance extended to the target point in the tree, and ||Sjcy-Zjjd|| represents the norm of the vector. This autonomous obstacle avoidance capability enhances overall flight safety, especially in urban environments or dynamically changing scenes;

[0059] The flight acceleration calculation formula is as follows:

[0060]

[0061] In a drone cluster, each drone can work together through AI algorithms and share location information and environmental data with each other. In the formula, Fxjd represents flight acceleration, F net It represents the net external force on the drone, m represents the mass of the drone, and g represents the acceleration due to gravity, which is 9.81m / s 2 ,Calculating the flight acceleration can ensure that the UAVs in the cluster maintain a reasonable distance and formation during the flight to prevent mutual interference;

[0062] The flight steering angle calculation formula is as follows:

[0063] Fzxj=k*atan2(sin(θ target -θ current ), cos(θ target -θ current ))

[0064] Calculating the flight steering angle helps the drone to fly safely in complex environments. Real-time adjustment of the steering angle can effectively avoid obstacles, thereby reducing the risk of collision and ensuring the safety of the drone during the mission. In the formula, Fzxj represents the flight steering angle, k represents the gain k, which is used to smooth the steering, and θ target represents the target heading, θ current Indicates the current heading. By accurately calculating the steering angle, the drone can more flexibly respond to environmental changes such as wind speed, airflow or sudden obstacles, and continuously optimize the flight path. This flexibility enables the drone to perform tasks more reliably in dynamic environments.

[0065] S5. Calculate the control instructions for each drone based on the decision results, including the flight speed instruction Sdzl, the flight direction instruction Fxzl, and the flight altitude instruction Gdzl, and make dynamic adjustments for emergencies;

[0066] The flight speed command calculation formula is as follows:

[0067] Sdzl=V0+k v *(Mbsd-V0)+f v

[0068] In cluster flight, calculating the flight speed command helps ensure the speed consistency of the drones, making the formation flight more stable, avoiding collisions, and improving the overall execution efficiency. In the formula, Sdzl represents the flight speed command, V0 represents the current flight speed, and k v Represents the speed gain coefficient, between 0 and 1, used for smooth transition, f v It represents the impact factor of environmental factors on speed, Mbsd represents the target speed. By optimizing the speed command, the UAV can complete the task with higher energy efficiency, thereby extending the flight time and reducing battery consumption;

[0069] The flight direction command calculation formula is as follows:

[0070] Fxzl=θ0+k θ *Δθ

[0071] In the UAV cluster, by calculating the flight direction command, the UAVs can effectively adjust their relative positions and ensure the coordination and consistency of team activities. In the formula, Fxzl represents the flight direction command, θ0 represents the current heading angle, and k θ It represents the directional gain coefficient, which controls the adjustment speed. Δθ represents the angle difference between the target heading and the current heading. The autonomous direction command calculation capability enables the drone to fly in a more precise path, thus effectively avoiding obstacles and accurately reaching the target location.

[0072] The flight altitude command calculation formula is as follows:

[0073] Gdzl=h0+k h *(h target -h0)+f h

[0074] By calculating the reasonable altitude command, ground obstacles such as trees and buildings can be effectively avoided to ensure flight safety. In the formula, Gdzv represents the flight altitude command, h0 represents the current flight altitude, and h target represents the target flight altitude, k h represents the height gain coefficient, f hIndicates the impact of environmental factors on altitude control. While accurately controlling the flight altitude, it maximizes the use of airflow and wind power, thereby reducing energy consumption and improving flight efficiency.

[0075] S6, transmit control instructions to each drone in real time, monitor the execution of the drone, and track its status;

[0076] In the control system of a drone cluster, it is crucial to transmit control instructions to each drone in real time. Through efficient cellular network communication, instructions can be quickly and reliably distributed to each drone in the cluster, ensuring that they respond to mission requirements and environmental changes synchronously. During the execution process, the system continuously monitors the status of the drone, including flight speed, direction, altitude, attitude and sensor data to ensure that it flies according to the predetermined trajectory, while identifying and handling potential abnormal situations in real time. In addition, using data analysis and machine learning technology, the system can effectively track the execution of the drone, evaluate its performance and predict its status, and automatically adjust flight instructions to optimize the overall mission effect. This closed-loop feedback mechanism not only improves the coordination ability and mission completion of the drone cluster, but also significantly enhances the flexibility and safety of working in complex environments.

[0077] By analyzing the relevant information of the drone and combining Kalman filtering, deep learning algorithm and RRT planning algorithm to calculate the optimal path of the drone and generate the optimal control instructions, the complexity of the coordinated control and path planning of the drone cluster is reduced, thereby improving the efficiency of the drone operation.

[0078] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A drone cluster control method based on artificial intelligence, characterized in that: The following steps are involved: S1. Clarify the safety, efficiency and mission requirements of the drone and generate a preliminary flight path; S2. Collect data by using laser radar, camera and temperature sensor to collect information about the surrounding environment, including the location of obstacles, terrain data and climate conditions, obtain the status information of each drone in real time, including location, speed, altitude, battery power and attitude, and share information between drones; S3, clean and format the collected data, and use Kalman filtering technology to integrate data from different drones and sensors to obtain more comprehensive environmental and status information; S4. Apply deep learning algorithms to make decisions and calculate the optimal flight path Fxlj. According to environmental data and the position of the UAV, use the RRT planning algorithm to calculate the obstacle avoidance path Bzlj, flight acceleration Fxjd and flight steering angle Fzxj. At the same time, design a cluster control strategy based on the collected and calculated data to ensure that the UAVs work together to form an effective cluster formation. S5. Calculate the control instructions for each drone based on the decision results, including the flight speed instruction Sdzl, the flight direction instruction Fxzl, and the flight altitude instruction Gdzl, and make dynamic adjustments for emergencies; S6, transmits control instructions to each UAV in real time, monitors the execution of the UAV, and tracks its status.

2. The method for controlling a drone cluster based on artificial intelligence according to claim 1, characterized in that: The S3 cleans and formats the collected data, wherein the position coordinates are unified into the WGS84 coordinate system.

3. The method for controlling a drone cluster based on artificial intelligence according to claim 2, characterized in that: S3 cleans and formats the collected data and integrates the data from different drones and sensors using Kalman filtering technology. The Kalman filtering formula is as follows: x k|k-1 =F k *x k-1|k-1 +B k *u k +w k In the formula, x k|k-1 represents the state estimation at time k based on time k-1, F k represents the state transition matrix, describing how the system migrates from time k-1 to time k, x k-1|k-1 represents the state estimate at time k-1, B k represents the control input matrix, u k Indicates the control instruction of external input, w k represents process noise.

4. The method for controlling a drone cluster based on artificial intelligence according to claim 3, characterized in that: The S4 applies a deep learning algorithm to make decisions and calculate the optimal flight path Fxlj. The calculation formula is as follows: Fxlj(s t ,a t )=r t +γ*max Fxlj(s t+1 ,a′) In the formula, Fxlj(s t ,a t ) means in state s t Take action a t The expected return, r t represents the immediate reward obtained at time t, γ represents the discount factor, which ranges from 0 to 1 and represents the importance of future rewards, max Fxlj(s t+1 ,a′) represents the maximum expected return of the next state, and continuously optimizes Fxlj(s t ,a t ) function, calculates the best action under a given state to guide the flight path selection of the drone.

5. The method for controlling a drone cluster based on artificial intelligence according to claim 4, characterized in that: The S4 uses a deep learning algorithm to make decisions, and uses the RRT planning algorithm to calculate the obstacle avoidance path Bzlj according to the environmental data and the position of the drone. The calculation formula is as follows: In the formula, Bzlj represents the obstacle avoidance path, Zjjd represents the nearest node in the RRT algorithm tree, Sjcy represents a target point randomly sampled in the RRT algorithm tree, Δt represents the step size, that is, the distance extended to the target point in the tree, and ||Sjcy-Zjjd|| represents the norm of the vector.

6. The method for controlling a drone cluster based on artificial intelligence according to claim 5, characterized in that: The S4 uses a deep learning algorithm to make decisions and uses the RRT planning algorithm to calculate the flight acceleration Fxjd based on environmental data and the position of the drone. The calculation formula is as follows: In the formula, Fxjd represents the flight acceleration, F net It represents the net external force on the drone, m represents the mass of the drone, and h represents the acceleration due to gravity, which is 9.81 m / s 2 .

7. The method for controlling a drone cluster based on artificial intelligence according to claim 6, characterized in that: The S4 uses a deep learning algorithm to make decisions, and uses the RRT planning algorithm to calculate the flight steering angle Fzxj according to the environmental data and the position of the drone. The calculation formula is as follows: Fzxj=k*atan2(sin(θ target -θ current ),cos(θ target -θ current )) In the formula, Fzxj represents the flight steering angle, k represents the gain k, which is used to smooth the steering, and θ target represents the target heading, θ current Indicates the current heading.

8. The method for controlling a drone cluster based on artificial intelligence according to claim 7, characterized in that: The S5 calculates the control instructions for each UAV according to the decision result, wherein the calculation formula of the flight speed instruction Sdzl is as follows: Sdzl=V0+k v *(Mbsd-V0)+f v In the formula, Sdzl represents the flight speed instruction, V0 represents the current flight speed, k v Represents the speed gain coefficient, between 0 and 1, used for smooth transition, f v It represents the impact factor of environmental factors on speed, and Mbsd represents the target speed.

9. The method for controlling a drone cluster based on artificial intelligence according to claim 8, characterized in that: The S5 calculates the control instructions for each UAV according to the decision result, wherein the calculation formula of the flight direction instruction Fxzl is as follows: Fxzl=θ0+k θ *Dth In the formula, Fxzl represents the flight direction instruction, θ0 represents the current heading angle, and k θ It represents the directional gain coefficient, which controls the adjustment speed. Δθ represents the angular difference between the target heading and the current heading.

10. The method for controlling a drone cluster based on artificial intelligence according to claim 9, characterized in that: The S5 calculates the control instructions for each UAV according to the decision result, wherein the calculation formula of the flight altitude instruction Gdzl is as follows: Gdzl=h0+k h *(h target -h0)+f h In the formula, Gdzl represents the flight altitude instruction, h0 represents the current flight altitude, and h target represents the target flight altitude, k h represents the height gain coefficient, f h Represents the impact of environmental factors on height control.

Images