Unmanned aerial vehicle formation route automatic planning system and method

Through an automated drone fleet route planning system, multi-objective optimization and machine learning algorithms are used to dynamically adjust routes, solving the shortcomings of real-time and flexibility in the existing technology, and achieving efficient and safe route planning in complex environments and task execution.

CN118915817APending Publication Date: 2024-11-08上海辅雅智能技术有限公司
View PDF 0 Cites 4 Cited by

Patent Information

Application Number
CN202410949084.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-16
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The existing drone fleet route planning technology lacks real-time and flexibility, and it is difficult to meet the requirements of rapid response in complex environments.

Method used

Adopt an automated drone fleet route planning system, including data collection, environmental modeling, route generation, optimization, evaluation, adjustment and output modules, combining multi-objective optimization algorithms, genetic algorithms, particle swarm optimization, simulated annealing, machine learning and Strengthen learning to achieve dynamic adjustment and adaptive optimization of routes.

Benefits of technology

It improves the efficiency and safety of route planning, has the ability to learn and adapt to environmental changes, ensures the continuity and stability of task execution, and adapts to dynamic changes in complex environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118915817A_ABST
    Figure CN118915817A_ABST
Patent Text Reader

Abstract

The invention discloses an unmanned aerial vehicle formation route automatic planning system and method, and relates to the technical field of unmanned aerial vehicle formation route planning. Comprising a data collection module which is responsible for collecting data such as position information, speed information and performance parameters of each unmanned aerial vehicle in an unmanned aerial vehicle formation and environment information of a task area; the environment modeling module is used for constructing a three-dimensional space environment model according to the information provided by the data collection module, wherein the three-dimensional space environment model comprises factors such as terrain, obstacles and wind speed and direction; the route generation module is used for generating an initial route by using a formula and an algorithm; and the route optimization module is used for optimizing the initial route by adopting a multi-objective optimization algorithm, and balancing the route length, safety and energy consumption factors. According to the invention, through the automatic route generation and optimization module, the optimal or approximately optimal route can be quickly calculated, and compared with manual planning, the efficiency is greatly improved; various factors such as distance, obstacles and wind speed are accurately considered, and reliability and safety of the route are ensured.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of UAV formation route planning, and in particular to a UAV formation route automatic planning system and method. Background Art

[0002] With the rapid development of drone technology, it has been widely used in many fields such as military reconnaissance, disaster monitoring, and traffic monitoring. When drone formations perform missions, reasonable route planning is crucial to improving mission efficiency and ensuring flight safety. However, route planning in existing technologies often relies on manual experience, lacks flexibility and real-time performance, and cannot meet the requirements of rapid response in complex environments.

[0003] After searching, the application scheme of Chinese patent application number CN202410605119.5 discloses a method, system and electronic device for planning the route of a drone, wherein the method includes pre-allocating n drones for the survey area of ​​the orthophoto operation; planning the orthophoto route of the survey area, and calculating the entry point coordinates of each drone according to the preset drone flight distance, the location of the waypoint, the take-off point location, and the landing point location; wherein the entry point of the first drone is the first waypoint on the orthophoto route, and the entry point of the remaining drones is the return point of the previous drone; calculating the average value of the flight distance of n drones and the average difference of the flight distance of the drones; if the average difference is greater than or equal to the preset threshold, the drone flight distance is reset, and the entry point coordinates of each drone are calculated again according to the reset drone flight distance, until the average difference is less than the preset threshold, or the number of loop calculations of the drone entry point reaches the preset maximum number of loops, and the final entry point of each drone is obtained. The planning method in the above patent has the following deficiencies: the ability to adjust and respond to environmental changes in real time is still insufficient. Summary of the invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a system and method for automatically planning the route of a UAV formation.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] An automatic route planning system for a UAV formation, comprising:

[0007] Data collection module: responsible for collecting the location information, speed information, performance parameters of each drone in the drone formation and environmental information data of the mission area;

[0008] Environmental modeling module: constructs a three-dimensional spatial environment model based on the information provided by the data collection module;

[0009] Route generation module: Generates the initial route using formulas and algorithms; factors considered include:

[0010] Avoid obstacles: O is the cumulative shortest distance between all waypoints and obstacles in the route; m is the number of obstacles; P j is the jth waypoint in the route; O j is the position of the jth obstacle; d(·) is the function for calculating the distance between two points;

[0011] Route optimization module: uses a multi-objective optimization algorithm to optimize the initial route, balancing route length, safety and energy consumption factors;

[0012] Route evaluation module: evaluates the feasibility and safety of routes through flight simulation and risk assessment algorithms;

[0013] Route adjustment module: dynamically adjusts the route based on the evaluation results to ensure real-time response to environmental changes and mission requirements;

[0014] Route output module: outputs the final optimized route to the drone control system to achieve automatic flight.

[0015] Preferably, the automatic planning method of the UAV formation route automatic planning system comprises the following steps:

[0016] S1: Obtain necessary information through the data collection module;

[0017] S2: The environment modeling module builds an environment model based on this information;

[0018] S3: The route generation module calculates the initial route;

[0019] S4: The route optimization module uses genetic algorithm and particle swarm optimization algorithm to optimize the route;

[0020] S5: The route evaluation module evaluates the optimized route. If it does not meet the requirements, the route adjustment module will adjust it;

[0021] S6: The route output module transmits the final route to the UAV control system for execution.

[0022] Preferably: the route optimization module includes:

[0023] Optimization algorithm selection unit: automatically selects appropriate optimization algorithms, including genetic algorithms, particle swarm optimization or simulated annealing, according to task requirements and environment complexity;

[0024] Multi-objective weighing unit: sets and weighs the weights of multiple optimization objectives to ensure that the optimization results meet multiple requirements;

[0025] Constraint processing unit: handles various constraints, including flight altitude restrictions, no-fly zone avoidance, and drone performance parameter restrictions.

[0026] Preferably: the route optimization module solves the optimization problem by a genetic algorithm, wherein each point of the route is regarded as a gene, and the total length and safety of the route are fitness. The specific formula is as follows:

[0027]

[0028] Where f(x) represents the route length, x i and i are the coordinates of the i-th point on the route.

[0029] Preferably: the route optimization module particle swarm optimization algorithm solves the optimization problem, each particle represents a potential solution, and updates its position and speed according to the optimal experience of individuals and groups. The specific formula is as follows:

[0030]

[0031] Among them, v id and x id are the velocity and position of particle i in the dth dimension, w is the inertia weight, c1 is the c2 acceleration constant, pBest id is the individual best, gBest d It is the global optimum.

[0032] Preferably: the route evaluation module includes:

[0033] Risk Analysis Unit: Evaluate possible risk factors in the route;

[0034] Simulated flight unit: simulates the flight of the UAV along the planned route through simulation technology to verify the feasibility of the route;

[0035] Performance evaluation unit: evaluates the flight time and energy consumption performance indicators of the route and determines the optimization direction of the route;

[0036] The route adjustment module includes:

[0037] Real-time feedback unit: receives real-time flight data and environmental information from the drone and detects whether route adjustment is needed;

[0038] Dynamic correction unit: dynamically corrects the route based on real-time data and evaluation results to ensure real-time optimization of the route;

[0039] The route output module includes:

[0040] Route visualization unit: displays the optimized route in a graphical way to facilitate operators’ understanding and decision-making;

[0041] Data conversion unit: converts the data format of the planned route into instructions that can be recognized by the drone control system;

[0042] Interface communication unit: establish a communication interface with the UAV control system to achieve seamless data transmission.

[0043] Preferably, it also includes:

[0044] Machine Learning Module: Use machine learning algorithms to analyze historical flight data and predict future environmental changes and potential risks;

[0045] Adaptive adjustment module: uses Q-learning algorithm to adjust the route to adapt to dynamically changing flight conditions and mission requirements;

[0046] Multi-UAV collaborative planning module: introduces multi-agent reinforcement learning algorithm to coordinate the route planning of multiple UAVs;

[0047] Real-time re-planning module: When encountering emergencies or drastic changes in the environment, this module re-plans the route; using model predictive control algorithms, it ensures that the drone can respond quickly and adjust its route in complex environments.

[0048] Preferably: the machine learning module includes:

[0049] Environmental feature learning unit: uses deep learning algorithms to learn and extract environmental features from historical flight data;

[0050] Pattern recognition unit: identifies common environmental patterns and flight conditions to provide reference for route planning;

[0051] Predictive Analysis Unit: Predict future environmental changes and flight risks based on learned features and patterns.

[0052] Preferably: the machine learning module uses convolutional neural networks and recurrent neural networks to extract environmental features related to flight safety and efficiency;

[0053] Among them, terrain classification or obstacle recognition uses convolutional neural network, and the specific formula is:

[0054] X i =σ(W conv *X i-1 +B conv )

[0055] Among them, X i is the feature map of the i-th layer, W conv and B conv are the weight and bias of the convolutional layer, and σ is the activation function;

[0056] Among them, for time series data, a recurrent neural network is applied, and the specific formula is:

[0057] h t =σ(W rnn *[h t-1 , X t ]+B rnn )

[0058] Among them, h t is the hidden state at time t, W rnn and B rnn are the weights and biases of the RNN layer, X t is the input at time t.

[0059] Preferably, the adaptive adjustment module continuously optimizes the route planning strategy through reinforcement learning to improve the adaptability and optimization effect of the route. The specific implementation method is:

[0060] State definition: define the flight state of the drone, including position, speed, and environmental factors;

[0061] Action design: design possible flight actions;

[0062] Reward mechanism: Design a reward function to give positive rewards for favorable behaviors such as reaching the goal and avoiding obstacles, and negative rewards for unfavorable behaviors such as collision and deviation from the route;

[0063] Strategy optimization: By continuously interacting with the environment, Q-learning is used to iteratively update the Q value and optimize the route planning strategy;

[0064] Strategy application: Apply optimized strategies to real-time route planning to improve route adaptability and overall performance;

[0065] Among them, the specific formula for using Q-learning to iteratively update the Q value and optimize the route planning strategy is:

[0066]

[0067] Where Q(s, a) is the state-action value function, α is the learning rate, R is the reward, γ is the discount factor, and s' and a' are the next state and action, respectively.

[0068] The beneficial effects of the present invention are:

[0069] 1. The present invention can quickly calculate the optimal or approximately optimal route through an automated route generation and optimization module, greatly improving efficiency compared to manual planning; accurately consider multiple factors, such as distance, obstacles, wind speed, etc., to ensure the reliability and safety of the route.

[0070] 2. The present invention introduces machine learning and reinforcement learning, so that the system has the ability to self-learn and adapt to environmental changes, thereby improving the level of intelligent decision-making; the real-time replanning module allows the system to quickly respond to environmental changes or emergencies, make dynamic adjustments, and ensure the continuity and stability of task execution.

[0071] 3. The multi-UAV collaborative planning module of the present invention ensures that the UAV formation can effectively cooperate when performing tasks, thereby improving overall operational efficiency and effectiveness; the system not only takes static environmental factors into consideration, but also adapts to dynamic environmental factors such as weather changes through advanced algorithms, thereby improving its applicability in different environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 This is a flow chart of a planning method for an automatic route planning system for a UAV formation proposed by the present invention. DETAILED DESCRIPTION

[0073] The technical solution of the present invention is further described in detail below in conjunction with specific implementation methods.

[0074] Embodiment 1:

[0075] An automatic route planning system for a UAV formation, comprising:

[0076] Data collection module: responsible for collecting data such as the location information, speed information, performance parameters of each drone in the drone formation, and environmental information of the mission area;

[0077] Environmental modeling module: constructs a three-dimensional spatial environment model based on the information provided by the data collection module, including factors such as terrain, obstacles, wind speed and direction;

[0078] Route generation module: Generates the initial route using formulas and algorithms; factors considered include:

[0079] Avoid obstacles: O is the cumulative shortest distance between all waypoints and obstacles in the route; m is the number of obstacles; P j is the jth waypoint in the route; O j is the position of the jth obstacle; d(·) is the function for calculating the distance between two points;

[0080] Route optimization module: uses a multi-objective optimization algorithm to optimize the initial route, balancing factors such as route length, safety, and energy consumption;

[0081] Route evaluation module: evaluates the feasibility and safety of routes through flight simulation and risk assessment algorithms;

[0082] Route adjustment module: dynamically adjusts the route based on the evaluation results to ensure real-time response to environmental changes and mission requirements;

[0083] Route output module: outputs the final optimized route to the drone control system to achieve automatic flight.

[0084] The automatic planning method of the UAV formation route automatic planning system comprises the following steps:

[0085] S1: Obtain necessary information through the data collection module;

[0086] S2: The environment modeling module builds an environment model based on this information;

[0087] S3: The route generation module calculates the initial route;

[0088] S4: The route optimization module uses genetic algorithms, particle swarm optimization and other algorithms to optimize the route;

[0089] S5: The route evaluation module evaluates the optimized route. If it does not meet the requirements, the route adjustment module will adjust it;

[0090] S6: The route output module transmits the final route to the UAV control system for execution.

[0091] Wherein, the route optimization module includes:

[0092] Optimization algorithm selection unit: automatically selects appropriate optimization algorithms, such as genetic algorithm, particle swarm optimization or simulated annealing, according to task requirements and environment complexity;

[0093] Multi-objective weighing unit: sets and weighs the weights of multiple optimization objectives, such as route length, safety, energy consumption, etc., to ensure that the optimization results meet multiple requirements;

[0094] Constraint processing unit: handles various constraints, such as flight altitude restrictions, no-fly zone avoidance, drone performance parameter restrictions, etc.

[0095] The route optimization module solves the optimization problem through a genetic algorithm, where each point of the route is regarded as a gene, and the total length and safety of the route are fitness. The specific formula is as follows:

[0096]

[0097] Where f(x) represents the route length, x i and i are the coordinates of the i-th point on the route.

[0098] The particle swarm optimization algorithm in the route optimization module solves the optimization problem. Each particle represents a potential solution, and its position and speed are updated according to the optimal experience of individuals and groups. The specific formula is as follows:

[0099]

[0100] Among them, v id and x id are the velocity and position of particle i in the dth dimension, w is the inertia weight, c1 is the c2 acceleration constant, pBEst id is the individual best, gBest d It is the global optimum.

[0101] Wherein, the route evaluation module includes:

[0102] Risk analysis unit: evaluates possible risk factors in the route, such as weather conditions, potential obstacle collision risks, etc.;

[0103] Simulated flight unit: simulates the flight of the UAV along the planned route through simulation technology to verify the feasibility of the route;

[0104] Performance evaluation unit: evaluates the flight time, energy consumption and other performance indicators of the route, and determines the optimization direction of the route.

[0105] Wherein, the route adjustment module includes:

[0106] Real-time feedback unit: receives real-time flight data and environmental information from the drone and detects whether route adjustment is needed;

[0107] Dynamic correction unit: dynamically corrects the route based on real-time data and evaluation results to ensure real-time optimization of the route;

[0108] Safety Priority Unit: When adjusting routes, safety is given priority and any route design that may lead to safety accidents is avoided.

[0109] Wherein, the route output module includes:

[0110] Route visualization unit: displays the optimized route in a graphical way to facilitate operators’ understanding and decision-making;

[0111] Data conversion unit: converts the data format of the planned route into instructions that can be recognized by the drone control system;

[0112] Interface communication unit: establish a communication interface with the UAV control system to achieve seamless data transmission.

[0113] Embodiment 2:

[0114] A UAV formation route automatic planning system, based on Example 1, this embodiment also includes:

[0115] Machine Learning Module: Use machine learning algorithms to analyze historical flight data to predict future environmental changes and potential risks. This module can use deep learning networks, such as convolutional neural networks (CNNs) to process terrain image data, or recurrent neural networks (RNNs) to analyze time series data, thereby improving the accuracy and reliability of route planning.

[0116] Adaptive adjustment module: Combined with reinforcement learning algorithms, the system can autonomously adjust the route planning strategy based on real-time feedback; the system obtains rewards or penalties through interaction with the environment and continuously optimizes the decision-making process of route planning; for example, the Q-learning algorithm is used to adjust the route to adapt to dynamically changing flight conditions and mission requirements;

[0117] Multi-UAV collaborative planning module: Introduces multi-agent collaborative algorithms, such as multi-agent reinforcement learning (MARL), to coordinate the route planning of multiple UAVs; this module ensures that the UAV formation can effectively avoid mutual interference while maintaining the formation, thus optimizing the overall mission execution efficiency;

[0118] Real-time re-planning module: When encountering emergencies or drastic changes in the environment, this module can quickly re-plan the route; using real-time data processing and online optimization algorithms, such as model predictive control (MPC), to ensure that the UAV can quickly respond and adjust the route in complex environments.

[0119] Wherein, the machine learning module includes:

[0120] Environmental feature learning unit: uses deep learning algorithms to learn and extract environmental features from historical flight data;

[0121] Pattern recognition unit: identifies common environmental patterns and flight conditions to provide reference for route planning;

[0122] Predictive Analysis Unit: Predict future environmental changes and flight risks based on learned features and patterns.

[0123] The machine learning module uses convolutional neural networks (CNN) and recurrent neural networks (RNN) to extract environmental features related to flight safety and efficiency;

[0124] Among them, terrain classification or obstacle recognition uses convolutional neural network, and the specific formula is:

[0125] X i =σ(W conv *X i-1 +B conv )

[0126] Among them, X i is the feature map of the i-th layer, W convand B conv are the weight and bias of the convolutional layer, and σ is the activation function;

[0127] Among them, for time series data, such as changes in wind speed and wind direction, a recurrent neural network is applied, and the specific formula is:

[0128] h t =σ(W rnn *[h t-1 , X t ]+B rnn )

[0129] Among them, h t is the hidden state at time t, W rnn and B rnn are the weights and biases of the RNN layer, X t is the input at time t.

[0130] Wherein, the adaptive adjustment module includes:

[0131] Strategy Optimization Unit: Through reinforcement learning, continuously optimize route planning strategies to improve route adaptability and optimization effects;

[0132] Autonomous decision-making unit: Based on the results of reinforcement learning, the system can make route adjustment decisions autonomously;

[0133] Environmental adaptation unit: Adaptively adjust route planning parameters and strategies in response to different environmental changes.

[0134] The adaptive adjustment module continuously optimizes the route planning strategy through reinforcement learning to improve the adaptability and optimization effect of the route. The specific implementation method is as follows:

[0135] State definition: define the flight state of the drone, including position, speed, environmental factors, etc.;

[0136] Action design: design possible flight actions, such as heading adjustment, speed adjustment, etc.;

[0137] Reward mechanism: Design a reward function to give positive rewards for favorable behaviors such as reaching the goal and avoiding obstacles, and negative rewards for unfavorable behaviors such as collision and deviation from the route;

[0138] Strategy optimization: By continuously interacting with the environment, Q-learning is used to iteratively update the Q value and optimize the route planning strategy;

[0139] Strategy application: Apply optimized strategies to real-time route planning to improve route adaptability and overall performance;

[0140] Among them, the specific formula for using Q-learning to iteratively update the Q value and optimize the route planning strategy is:

[0141]

[0142] Where Q(s,a) is the state-action value function, α is the learning rate, R is the reward, γ is the discount factor, and s' and a' are the next state and action, respectively.

[0143] Wherein, the multi-UAV collaborative planning module includes:

[0144] Formation management unit: ensures that the UAV formation maintains the prescribed formation and distance during the mission;

[0145] Collaborative optimization unit: Collaboratively optimize the flight routes based on the flight characteristics of each UAV in the formation;

[0146] Communication coordination unit: manages communications between drones to ensure timely sharing and coordination of information.

[0147] Wherein, the real-time re-planning module includes:

[0148] Event trigger unit: monitors events that require real-time re-planning, such as sudden weather changes, drone failures, etc.

[0149] Rapid response unit: Once an event requiring re-planning is detected, the re-planning procedure is initiated quickly to reduce mission delays;

[0150] Plan generation unit: quickly generate new feasible route plans based on the current environment and status.

[0151] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. An automatic route planning system for a UAV formation, characterized in that: include: Data collection module: responsible for collecting the location information, speed information, performance parameters of each drone in the drone formation and environmental information data of the mission area; Environmental modeling module: constructs a three-dimensional spatial environment model based on the information provided by the data collection module; Route generation module: Generates initial routes using formulas and algorithms; Considerations include: Avoid obstacles: O is the cumulative shortest distance between all waypoints and obstacles in the route; m is the number of obstacles; P j is the jth waypoint in the route; O j is the position of the jth obstacle; d(·) is the function for calculating the distance between two points; Route optimization module: uses a multi-objective optimization algorithm to optimize the initial route, balancing route length, safety and energy consumption factors; Route evaluation module: evaluates the feasibility and safety of routes through flight simulation and risk assessment algorithms; Route adjustment module: dynamically adjusts the route based on the evaluation results to ensure real-time response to environmental changes and mission requirements; Route output module: outputs the final optimized route to the drone control system to achieve automatic flight.

2. The automatic route planning system for a UAV formation according to claim 1 is characterized in that: The automatic planning method of the UAV formation route automatic planning system comprises the following steps: S1: Obtain necessary information through the data collection module; S2: The environment modeling module builds an environment model based on this information; S3: The route generation module calculates the initial route; S4: The route optimization module uses genetic algorithm and particle swarm optimization algorithm to optimize the route; S5: The route evaluation module evaluates the optimized route. If it does not meet the requirements, the route adjustment module will adjust it; S6: The route output module transmits the final route to the UAV control system for execution.

3. The automatic route planning system for a UAV formation according to claim 1 is characterized in that: The route optimization module includes: Optimization algorithm selection unit: automatically selects appropriate optimization algorithms, including genetic algorithms, particle swarm optimization or simulated annealing, according to task requirements and environment complexity; Multi-objective weighing unit: sets and weighs the weights of multiple optimization objectives to ensure that the optimization results meet multiple requirements; Constraint processing unit: handles various constraints, including flight altitude restrictions, no-fly zone avoidance, and drone performance parameter restrictions.

4. The automatic route planning system for a UAV formation according to claim 3 is characterized in that: The route optimization module solves the optimization problem through a genetic algorithm, where each point of the route is regarded as a gene, and the total length and safety of the route are fitness. The specific formula is as follows: Where f(x) represents the route length, x i and i are the coordinates of the i-th point on the route.

5. The automatic route planning system for a UAV formation according to claim 3 is characterized in that: The particle swarm optimization algorithm in the route optimization module solves the optimization problem. Each particle represents a potential solution, and its position and speed are updated according to the optimal experience of individuals and groups. The specific formula is as follows: Among them, v id and x id are the velocity and position of particle i in the dth dimension, w is the inertia weight, c1 is the c2 acceleration constant, pBest id is the individual best, gBest d It is the global optimum.

6. The automatic route planning system for a UAV formation according to claim 1 is characterized in that: The route evaluation module includes: Risk Analysis Unit: Evaluate possible risk factors in the route; Simulated flight unit: simulates the flight of the UAV along the planned route through simulation technology to verify the feasibility of the route; Performance evaluation unit: evaluates the flight time and energy consumption performance indicators of the route and determines the optimization direction of the route; The route adjustment module includes: Real-time feedback unit: receives real-time flight data and environmental information from the drone and detects whether route adjustment is needed; Dynamic correction unit: dynamically corrects the route based on real-time data and evaluation results to ensure real-time optimization of the route; The route output module includes: Route visualization unit: displays the optimized route in a graphical way to facilitate operators’ understanding and decision-making; Data conversion unit: converts the data format of the planned route into instructions that can be recognized by the drone control system; Interface communication unit: establish a communication interface with the UAV control system to achieve seamless data transmission.

7. The automatic route planning system for a UAV formation according to claim 1 is characterized in that: Also includes: Machine Learning Module: Use machine learning algorithms to analyze historical flight data and predict future environmental changes and potential risks; Adaptive adjustment module: uses Q-learning algorithm to adjust the route to adapt to dynamically changing flight conditions and mission requirements; Multi-UAV collaborative planning module: introduces multi-agent reinforcement learning algorithm to coordinate the route planning of multiple UAVs; Real-time re-planning module: When encountering emergencies or drastic changes in the environment, this module re-plans the route; using model predictive control algorithms, it ensures that the drone can respond quickly and adjust its route in complex environments.

8. The automatic route planning system for a UAV formation according to claim 7 is characterized in that: The machine learning module includes: Environmental feature learning unit: uses deep learning algorithms to learn and extract environmental features from historical flight data; Pattern recognition unit: identifies common environmental patterns and flight conditions to provide reference for route planning; Predictive Analysis Unit: Predict future environmental changes and flight risks based on learned features and patterns.

9. The automatic route planning system for a UAV formation according to claim 8, characterized in that: The machine learning module uses convolutional neural networks and recurrent neural networks to extract environmental features related to flight safety and efficiency; Among them, terrain classification or obstacle recognition uses convolutional neural network, and the specific formula is: X i =σ(W conv *X i-1 +B conv ) Among them, X i is the feature map of the i-th layer, W conv and B conv are the weight and bias of the convolutional layer, and σ is the activation function; Among them, for time series data, a recurrent neural network is applied, and the specific formula is: h t =σ(W rnn *[h t-1 ,X t ]+B rnn ) Among them, h t is the hidden state at time t, W rnn and B rnn are the weights and biases of the RNN layer, X t is the input at time t.

10. The automatic route planning system for a UAV formation according to claim 9, characterized in that: The adaptive adjustment module continuously optimizes the route planning strategy through reinforcement learning to improve the adaptability and optimization effect of the route. The specific implementation method is as follows: State definition: define the flight state of the drone, including position, speed, and environmental factors; Action design: design possible flight actions; Reward mechanism: Design a reward function to give positive rewards for favorable behaviors such as reaching the goal and avoiding obstacles, and negative rewards for unfavorable behaviors such as collision and deviation from the route; Strategy optimization: By continuously interacting with the environment, Q-learning is used to iteratively update the Q value and optimize the route planning strategy; Strategy application: Apply optimized strategies to real-time route planning to improve route adaptability and overall performance; Among them, the specific formula for using Q-learning to iteratively update the Q value and optimize the route planning strategy is: Where Q(s,a) is the state-action value function, α is the learning rate, R is the reward, γ is the discount factor, and s' and a' are the next state and action, respectively.

Citation Information

Cited By

  • Omnibearing supervision system for low-altitude aircraft

    CN119379009A

  • Wind power plant topographic surveying and mapping method and system based on unmanned aerial vehicle

    CN119413149A

  • Deep sea deduction scene air route path intelligent planning system based on optimal control theory

    CN119806167A

  • Unmanned aerial vehicle real-time communication and control method based on 5.5 G network

    CN119854749A