Cooperative flight crossing method of unmanned aerial vehicle group in unknown environment based on adaptive target guiding strategy

Abstract

The invention relates to an unmanned aerial vehicle cluster cooperative flight crossing method based on a self-adaptive target guiding strategy, and solves the problem of conflict between obstacle avoidance and task target guiding of an unmanned aerial vehicle cluster in a complex environment. Through role division of an information agent and a non-information agent, the information agent is responsible for environment perception and key information transmission, and the non-information agent adjusts a flight path according to feedback. The task guiding and obstacle avoidance priorities are dynamically adjusted through the self-adaptive target guiding strategy, and the balance between the task target and safe obstacle avoidance is ensured. The bionic visual perception mechanism realizes real-time obstacle avoidance through local perception, and the willingness propagation mechanism ensures rapid sharing of information in the cluster, so that the cooperation efficiency is improved. The method effectively improves the task completion rate, robustness and cooperative capability of the unmanned aerial vehicle group in dynamic and unknown environments, and is widely applied to the fields of post-disaster rescue, environment monitoring and the like.

Description

Technical Field

[0001] The present invention relates to a drone cluster control technology, and in particular to a method for cooperative flight crossing of a drone cluster in an unknown environment based on an adaptive target-oriented strategy. Background Art

[0002] With the rapid development of artificial intelligence and drone technology, UAVs (Unmanned Aerial Vehicles, UAVs) are increasingly used in civil and military fields. However, a single UAV is limited by battery life, payload capacity and sensing range when performing tasks, making it difficult to meet the mission requirements in complex environments. To overcome these limitations, multi-UAV systems have emerged. Inspired by the clustering behavior of organisms in nature, multi-UAV systems improve mission execution efficiency and system robustness through the collaborative operation of multiple UAVs. Existing UAV cluster control methods mainly include centralized control and distributed control. Centralized control relies on the central node to make decisions. Once the central node fails, the entire system may be paralyzed; while distributed control adopts a decentralized strategy, allowing UAVs to make autonomous decisions and improve system adaptability and stability.

[0003] When a swarm of drones traverses a complex unknown environment, most existing methods rely on globally known environmental information. However, it is difficult for drones to obtain comprehensive information in an unknown environment, which limits their obstacle avoidance capabilities. Existing methods are inadequate in balancing obstacle avoidance requirements with mission objectives, causing drones to deviate from their mission directions during obstacle avoidance, affecting mission completion efficiency. In environments with high interference or limited communications, information transmission within the cluster is poor, resulting in reduced coordination of the drone swarm, affecting overall traversal capabilities. Summary of the invention

[0004] In order to solve the above technical problems, the purpose of the present invention is to provide a UAV swarm collaborative flight crossing method based on an adaptive goal-oriented strategy. This method introduces bionic visual perception mechanism, adaptive goal-oriented strategy and intention propagation mechanism to achieve collision-free crossing of UAV clusters in unknown environments, thereby improving the task completion rate and overall system robustness.

[0005] The method of the present invention relies on the swarm control of a multi-UAV system. To achieve efficient cooperative flight of the swarm, the system divides each UAV into two roles: Information Agents and Non-Info Agents. Among them, the Information Agents are responsible for environmental perception and transmitting key information. The Information Agents can obtain local information of the entire environment and make decisions based on this information. The Non-Info Agents rely on the environmental information provided by the Information Agents to make task decisions. The Non-Info Agents do not directly collect environmental data, but obtain task target information through local interaction with the Information Agents. This role division ensures that the swarm can achieve efficient task execution through distributed decision-making and local information exchange without central control.

[0006] Traditional swarm control methods often can only select a certain goal for priority processing when facing the conflict between task orientation and obstacle avoidance behavior, while ignoring the importance of the other goal. To solve this problem, the present invention proposes an adaptive goal-oriented strategy, which can adjust the priority between task orientation and obstacle avoidance behavior in real time to ensure that the swarm can respond flexibly in a complex environment.

[0007] During the task execution process, each UAV calculates the desired speed based on its target position and current position, and adjusts the flight path in combination with real-time environmental perception information.

[0008] When there is a conflict between task orientation (such as the flight direction pointing to the target area) and obstacle avoidance behavior (such as avoiding obstacles ahead), the system will dynamically adjust the task target direction based on an adaptive algorithm, so that the UAV can not only effectively avoid obstacles but also continue to move forward towards the target area. This adjustment is carried out in real time to ensure the balance between task orientation and obstacle avoidance behavior.

[0009] To improve the obstacle avoidance ability, the present invention designs a bionic vision obstacle avoidance mechanism. Inspired by the visual perception mechanism of organisms in nature, it obtains environmental information through visual sensors (such as cameras) and dynamically adjusts the flight path.

[0010] Each UAV senses the obstacles in front of it through a visual sensor. The sensing range of the UAV is divided into multiple areas, and the sensor can capture the position and size of the obstacles.

[0011] Once the UAV detects an obstacle, its obstacle avoidance strategy will be activated. The UAV will adjust its path according to the relative position of the obstacle to avoid collision with the obstacle.

[0012] While sensing the obstacle, the UAV will also adjust its speed and direction to achieve the short-range repulsive effect of "the closer, the larger; the farther, the smaller", that is, when the obstacle approaches, the UAV will decelerate or turn to ensure safety.

[0013] The core advantage of this mechanism is that it does not rely on global information, but dynamically adjusts the path based on local perception, improving the flexibility and robustness of the cluster in complex environments.

[0014] The present invention introduces a willingness propagation mechanism, enabling each individual in the cluster to adjust its behavior according to the feedback information of other individuals. The willingness propagation mechanism ensures the rapid spread of key information (such as obstacle positions, target positions, etc.) within the cluster, thereby enhancing the response speed and collaboration ability of the cluster.

[0015] When an information agent in the cluster detects a key obstacle or change (such as a change in the task target area), it will immediately transmit this information to other individuals through the willingness propagation mechanism.

[0016] This information dissemination adopts a rapid response mechanism similar to that in biological groups, ensuring that information can quickly cover the entire cluster, thus avoiding mission failures caused by information lag or omission of individual drones.

[0017] When the drone cluster of the present invention executes a collaborative crossing mission, it follows the following basic process: Mission preparation stage: The operator presets the mission target area according to the mission requirements and transmits the mission target information to the information agents in the cluster.

[0018] Each information agent obtains local environmental data through environmental perception sensors and shares this data with other information agents and non-information agents through local interaction.

[0019] Flight stage: Each drone executes the mission according to its role (information agent or non-information agent). The information agent is responsible for collecting and processing environmental data, and the non-information agent adjusts its flight path relying on the feedback information of the information agent.

[0020] All drones adjust their flights according to the adaptive target-oriented strategy to ensure finding the best balance between mission orientation and obstacle avoidance requirements.

[0021] During the flight, when encountering an obstacle, the bionic vision obstacle avoidance mechanism is activated to ensure that the drone can avoid the obstacle and continue to execute the mission.

[0022] Collaboration stage: The drones in the cluster share key information (such as obstacles, path changes, etc.) through the willingness propagation mechanism to ensure that each individual makes decisions based on the latest information.

[0023] In complex mission scenarios, the drone cluster dynamically adjusts the flight state of the cluster according to the mission target and environmental information to ensure the successful completion of the mission.

[0024] Task Completion Phase: Once the cluster reaches the predetermined target area, the task ends. The operator can evaluate the task effect and optimize the execution strategy of subsequent tasks based on the task completion situation of the cluster.

[0025] Compared with the prior art, the beneficial effects of the method, system and device of the present invention are as follows: Through the role division of information agents and non-information agents, the present invention realizes the efficient cooperation of the UAV cluster in complex tasks, avoiding information redundancy and computational burden. The obstacle avoidance mechanism based on bionic vision enables the UAV to avoid obstacles in real time in a dynamic environment, ensuring flight safety. The adaptive target-oriented strategy enables the UAV to flexibly adjust the task direction, solve the conflict between task orientation and obstacle avoidance behavior, and ensure the smooth completion of the task. The willingness propagation mechanism ensures that information can be quickly spread to all individuals in the cluster, enhancing the response ability and cooperation effect of the cluster. Description of the Drawings

[0026] Figure 1 It is a framework diagram of the collaborative crossing method based on the adaptive target-oriented strategy of the present invention.

[0027] Figure 2 It is a schematic diagram of the synthesis of the expected speed of the UAV.

[0028] Figure 3 It is a schematic diagram of the SAC interaction rule.

[0029] Figure 4 It is a schematic diagram of the UAV using bionic vision to perceive the environment.

[0030] Figure 5 It is a schematic diagram of the influence of visual perception items on the speed of the UAV.

[0031] Figure 6 It is a schematic diagram of task information based on the adaptive target-oriented strategy.

[0032] Figure 7 It is a flowchart of the collaborative crossing method based on the adaptive target-oriented strategy.

[0033] Figure 8 It is a schematic diagram of the initial position and flight space of the UAV cluster in a typical task scenario.

[0034] Figure 9 It is a movement trajectory diagram of the UAV cluster based on a non-adaptive target-oriented strategy.

[0035] Figure 10 It is a movement trajectory diagram of the UAV cluster based on the adaptive target-oriented strategy.

[0036] Figure 11It is a comparison graph of the task completion time of the collaborative crossing method based on the adaptive target-oriented strategy.

[0037] Figure 12 They are the task completion times of the UAV swarm under two strategies.

[0038] Figure 13 They are the 15 random test results of the non-adaptive target-oriented strategy and the adaptive target-oriented strategy.

[0039] Figure 14 They are the motion trajectories and quantization metrics of the UAV swarm in complex scenarios.

[0040] Figure 15 It is the construction of the verification scenario for the collaborative crossing method based on the adaptive target-oriented strategy.

[0041] Figure 16 It is the actual flight trajectory of the UAV swarm based on the non-adaptive target-oriented strategy.

[0042] Figure 17 It is the actual flight trajectory of the UAV swarm based on the adaptive target-oriented strategy.

[0043] Figure 18 They are the task completion times of the actual flight of the UAV swarm under two strategies.

[0044] Figure 19 It is the construction of a complex experimental scenario.

[0045] Figure 20 It is the collaborative crossing test based on the adaptive target-oriented strategy in complex scenarios. Specific implementation manner

[0046] The following further elaborates the present invention in detail in conjunction with the accompanying drawings and specific embodiments. For the step numbers in the following embodiments, they are only set for the convenience of elaboration and explanation, and no limitation is imposed on the order between the steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0047] Embodiment 1: UAV swarm collaborative crossing based on the adaptive target-oriented strategy

[0048] In this embodiment, it is assumed that the UAV swarm consists of 6 DJI Tello quadrotor UAVs, and the goal is to cross a complex obstacle area and reach a predetermined target area. The specific process is as follows: S1. Transmission of task target information and role division The operator inputs mission objective information through the ground control station, including the location of the target area and the flight path requirements. Six unmanned aerial vehicles (UAVs) in the UAV cluster are divided into two information agents and four non-information agents. The information agents are responsible for collecting environmental data and transmitting the mission objective information, while the non-information agents make flight decisions based on the feedback information provided by the information agents. The information agents use optical sensors (such as cameras, lidar, etc.) to obtain the position information of obstacles and targets in real time and transmit it to other UAVs through a wireless link. This step corresponds to Figure 1 , which illustrates the framework of the cooperative crossing method based on the adaptive target-oriented strategy.

[0049] S2. Environmental perception and information exchange

[0050] The environmental data collected by the information agents includes the positions of obstacles, potential dangerous areas in the flight path, and real-time changes in the mission target area. These data are quickly transmitted to other UAVs in the cluster to ensure that each UAV can share the latest environmental information. The non-information agents obtain the mission objectives and path adjustment information through communication with the information agents and perform optimized adjustments to the flight path. This step corresponds to Figure 2 and Figure 3 , Figure 2 showing the schematic diagram of the synthesis of the desired speed of the UAV, Figure 3 showing the schematic diagram of the SAC interaction rules.

[0051] The information agents also adjust the flight plan in real time according to the mission objectives and obstacle avoidance requirements. If there is a conflict between the mission objectives and the obstacle avoidance behavior, the information agents will adjust the direction of the mission objectives according to the environmental data to ensure a balance between the two. This adjustment is dynamic to ensure flexibility and safety during the flight process. This step corresponds to Figure 4 , which illustrates the UAV using bionic vision to perceive the environment.

[0052] S3. Activation of the bionic vision obstacle avoidance mechanism

[0053] Each UAV is equipped with a vision sensor (such as a high-definition camera) and a depth sensor to detect obstacles ahead. The UAV uses the bionic vision mechanism to identify obstacles and estimate the size and distance of the obstacles. When an obstacle approaches, the system will activate the obstacle avoidance algorithm to adjust the flight path to avoid collisions. At this time, the system will adjust the flight speed and direction according to the relative position and size of the obstacle to ensure safe obstacle avoidance. This step corresponds to Figure 5 , showing the schematic diagram of the influence of the visual perception item on the speed of the UAV.

[0054] S4. Real-time feedback and willingness propagation mechanism

[0055] During flight, once the information agents detect obstacles ahead or changes in the flight path, they transmit the new environmental information to other drones through the willingness propagation mechanism. Each non-information agent adjusts its flight path according to the latest mission objectives and environmental information, thus ensuring the coordination of the cluster in a complex environment. This mechanism ensures that each drone in the cluster can synchronously adjust its flight state and avoid mission failure caused by information lag. This step corresponds to Figure 6 and illustrates the schematic diagram of mission information based on the adaptive goal-oriented strategy.

[0056] S5. Target Tracking and Mission Completion

[0057] The drones in the cluster continuously adjust their flight paths during flight to ensure that they can avoid obstacles and fly towards the mission target area. When the cluster successfully crosses all obstacle areas and reaches the predetermined target area, the mission is completed. At this time, the ground control station will receive the feedback data of mission completion, and the system will generate a mission report to evaluate the execution effect of the mission. This step corresponds to Figure 7 and illustrates the flow chart of the cooperative crossing method based on the adaptive goal-oriented strategy.

[0058] S6. Mission Evaluation and Feedback Optimization

[0059] After the mission is completed, the cluster analyzes the key data during flight, including flight time, flight path, mission completion efficiency, obstacle avoidance success rate, etc. The system automatically optimizes the execution strategy of subsequent missions according to the evaluation results of mission execution. If a drone shows performance degradation or deviates from the target during flight, the system will record this data and make adjustments and optimizations in subsequent missions. This step corresponds to Figure 8 and illustrates the schematic diagram of the initial positions and flight spaces of the drone cluster in a typical mission scenario.

[0060] Embodiment 2: Application of Drone Cluster in Complex Environment

[0061] To further verify the effectiveness of the method of the present invention, this embodiment sets a complex urban environment mission, requiring the drone cluster to cross an area with dense buildings. This environment has multiple obstacles and signal interference in some areas. The specific steps are as follows: S1. Environmental Perception and Target Position Confirmation Before the mission starts, the operator inputs the approximate location of the mission target area through the ground control station. After the information broker in the cluster receives the target area information via the wireless link, it begins autonomous search. The information broker scans the surrounding environment through visual sensors and other sensors (such as lidar, infrared sensors, etc.), identifies obstacles, and confirms the precise location of the mission target. In an urban environment, due to the occlusion between buildings, the wireless signal is weak in some areas. The information broker will locate the target area based on the acquired environmental data and prior knowledge, and ensure that the UAV cluster can effectively search for the target. This step corresponds to Figure 9 and Figure 10 corresponding to Figure 9 showing the movement trajectories of the UAV cluster based on the non - adaptive target - oriented strategy, Figure 10 showing the movement trajectory diagram of the UAV cluster based on the adaptive target - oriented strategy.

[0062] S2. Resolution of conflicts between mission objectives and obstacle avoidance behaviors

[0063] During the flight, the UAV cluster may encounter multiple obstacles, resulting in conflicts between mission objectives and obstacle avoidance behaviors. At this time, the information broker dynamically adjusts the flight path according to the real - time obstacle information and target location. When the flight path conflicts with an obstacle, the information broker will slightly adjust the direction of the mission objective and notify the non - information brokers to adjust the flight path. Through the adaptive target - oriented strategy, the cluster can balance the mission orientation and obstacle avoidance requirements and achieve the successful completion of the mission. This step corresponds to Figure 11 corresponding to the diagram showing the comparison of mission completion times of the cooperative crossing method based on the adaptive target - oriented strategy.

[0064] S3. Target tracking and path adjustment

[0065] Each UAV in the cluster continuously monitors the surrounding environment through visual sensors, radar, and other sensors and adjusts the flight path in real - time. When a UAV approaches an obstacle, the system will automatically decelerate or change direction to ensure the flight safety of the cluster. In addition, the cluster ensures the synchronization of mission objectives through local information interaction and cooperative control, avoiding individual UAVs deviating from the mission objectives. This step corresponds to Figure 12 corresponding to the diagram showing the mission completion times of the UAV cluster under two strategies.

[0066] S4. Mission evaluation and flight data analysis

[0067] After the task is completed, all drones will feed back flight data to the ground control station to analyze key data such as flight paths, task completion times, and obstacle avoidance effects. The system will conduct task evaluations based on this data and optimize subsequent tasks. The system will generate task reports, analyze key issues during flight, and provide references for optimizing subsequent tasks. This step corresponds to Figure 13 and illustrates the results of 15 random tests of non - adaptive and adaptive goal - oriented strategies.

[0068] The method for collaborative crossing of a drone swarm based on an adaptive goal - oriented strategy provided by the present invention combines the role division of information agents and non - information agents, real - time environmental perception, task goal adjustment, bionic vision obstacle avoidance mechanism, and willingness propagation mechanism, and can effectively improve the collaborative efficiency and task completion ability of the drone swarm in complex environments. Through the method of the present invention, the swarm can flexibly adjust task goals in a dynamically changing environment to ensure flight safety and successful task execution.

[0069] The content in the above - mentioned method embodiments is applicable to the system embodiments of the present invention. The functions specifically implemented by the system embodiments of the present invention are the same as those of the above - mentioned method embodiments, and the beneficial effects achieved are also the same as those of the above - mentioned method embodiments.

[0070] The content in the above - mentioned method embodiments is applicable to the device embodiments of the present invention. The functions specifically implemented by the device embodiments of the present invention are the same as those of the above - mentioned method embodiments, and the beneficial effects achieved are also the same as those of the above - mentioned method embodiments.

[0071] The above has specifically described the preferred embodiments of the present invention, but the present invention is not limited to the described embodiments. Those skilled in the art can also make various equivalent deformations or substitutions without departing from the spirit of the present invention, and these equivalent deformations or substitutions are all included within the scope defined by the claims of this application.

Claims

1. A method for cooperative flight crossing of a swarm of unmanned aerial vehicles in an unknown environment based on an adaptive goal-oriented strategy, characterized in that: The method comprises the following steps: Drone swarm role division: Drones are divided into information agents and non-information agents. Information agents are responsible for perceiving the environment and guiding the swarm to traverse, while non-information agents rely on local interaction and perception for navigation. SAC interaction rules: Based on separation, alignment, and cohesion rules, collaborative movement between drones is achieved to ensure safe obstacle avoidance and cluster stability. Bionic visual obstacle avoidance mechanism: Use the visual projection field to perceive obstacle information, and adjust the flight trajectory of the drone by changing speed and steering to avoid collision. Dynamic adjustment of mission information: During target-oriented flight, the mission direction is adaptively adjusted to balance obstacle avoidance requirements and target mission requirements, thereby improving the mission completion rate of the drone cluster. Intention propagation mechanism: By sharing information between neighboring drones, it ensures that obstacle avoidance and target information can be quickly transmitted to the entire drone cluster, enhancing the dynamic adaptability of the cluster. Mission execution and evaluation: The UAV performs the crossing mission according to the optimized trajectory. After the mission is completed, the flight data is evaluated to optimize the mission parameters.

2. The method according to claim 1, characterized in that: The SAC interaction rules specifically include: Separation rule: Set a minimum safety distance. When the distance between drones is less than the threshold, increase the separation force to avoid collision. Alignment rule: adjust its own speed according to the speed of neighboring drones to keep the entire cluster in a consistent flight state and reduce the impact of speed fluctuations. Aggregation rules: Within a safe distance, by calculating the center position of the cluster, guide drones to maintain cluster compactness and avoid excessive dispersion of drones.

3. The method according to claim 1, characterized in that: The bionic visual obstacle avoidance mechanism comprises: Visual projection field calculation: The drone obtains surrounding environment information through its own sensors, constructs a visual projection field, and identifies the occluded area of ​​obstacles ahead. Sight Angle Adjustment: By calculating the angle of the traversable direction, the flight direction of the drone is adjusted when encountering obstacles to avoid collision. Dynamic speed adjustment: When an obstacle is in front of the drone, the speed will be automatically reduced; when an obstacle is on the side, the heading will be adjusted to avoid the obstacle.

4. The method according to claim 1, characterized in that: The task information dynamic adjustment method comprises: Calculate preferred mission direction: Calculate the preferred flight direction of the drone based on the target position. Adjust direction based on visual perception: When the target direction conflicts with the obstacle avoidance path, the flight direction is adaptively adjusted based on the UAV's visual projection field to make the mission information items compatible with the obstacle avoidance requirements. Mission direction correction: After obstacle avoidance is completed, the direction is gradually restored to the target direction to improve the mission execution efficiency of the drone cluster.

5. The method according to claim 1, characterized in that: The intention dissemination mechanism includes: Mission information dissemination: The information agent disseminates obstacle avoidance or target information to neighboring drones to improve the overall response speed of the cluster. Willingness identification model: By calculating the willingness factor, the degree of response of the drone to external information is determined, thereby enhancing the dynamic adaptability of the cluster. Optimizing information dissemination in restricted environments: When drones are in communication-restricted or high-interference environments, local dissemination strategies are used to ensure that the cluster can still collaborate efficiently.

6. The method according to claim 1, characterized in that This method is suitable for complex UAV collaborative flight missions such as environmental monitoring, search and rescue, and logistics transportation. It can achieve safe and efficient crossing of multiple UAVs in unknown environments without relying on global environmental information.

Images