Unmanned aerial vehicle cluster cooperation system based on ad hoc network

Abstract

The invention relates to an unmanned aerial vehicle cluster cooperative system based on an ad hoc network, which adopts a star topology structure, and the star topology structure takes a leader unmanned aerial vehicle as a center node and performs instruction issuing and data collection with a subordinate unmanned aerial vehicle through direct communication; all the subordinate unmanned aerial vehicles perform task execution around the leader unmanned aerial vehicle, so that centralized management and efficient coordination of tasks are ensured; the leader unmanned aerial vehicle serves as a core node of communication and task command and is responsible for distributing tasks, planning paths and monitoring task progress; under the star topology, part of subordinate unmanned aerial vehicles take other nodes as relays to ensure that the remote unmanned aerial vehicle can receive a task instruction and feed back data to the leader unmanned aerial vehicle, connectivity in a cluster is kept through multi-hop communication, and the remote unmanned aerial vehicle can receive the task instruction. A multi-hop communication mechanism is used to ensure that communication connection and task instruction transmission can still be kept even if a task area is relatively large or the flight distance of the unmanned aerial vehicle is relatively long. And the efficient collaborative operation of the unmanned aerial vehicle cluster in a complex environment is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle cluster control, and in particular to a unmanned aerial vehicle cluster coordination system based on a self-organizing network. Background Art

[0002] With the rapid development of drone technology, drone swarms have been widely used in logistics distribution, urban monitoring, emergency rescue and other fields. However, multi-drone collaborative flight faces several key problems in practical applications. First, the real-time communication problem is more prominent. Traditional communication modes have time delays in large areas or complex environments, which may lead to delays and incoordination in task execution. Secondly, the efficiency of path planning faces challenges. Drone swarms need to dynamically adjust paths according to mission requirements and environmental changes. Traditional distributed control modes are difficult to adapt to complex path planning requirements. Finally, obstacle avoidance in dynamic environments is also a major difficulty. Drones may encounter dynamic obstacles or sudden environmental changes during flight. Existing systems often cannot respond quickly and effectively, affecting the smooth execution of drone missions.

[0003] In the existing technology, the distributed control mode is prone to delays in issuing task instructions, increased complexity in path planning, and uncoordinated task execution due to the lack of unified global management. Therefore, a UAV cluster management system based on centralized control is proposed, which, combined with ad hoc network technology, can achieve efficient and reliable collaborative work of UAVs in dynamic and complex environments, and has important research value and practical application potential. Summary of the invention

[0004] The purpose of the present invention is to provide a drone cluster collaboration system based on a self-organizing network. The system centrally controls subordinate drones through a leader drone to ensure that the drone cluster can work together efficiently and reliably when performing tasks, and has dynamic obstacle avoidance, real-time task adjustment and multi-hop communication capabilities.

[0005] To achieve the above object, the present invention proposes the following technical solutions: A drone cluster coordination system based on a self-organizing network, the system adopts a star topology structure, the star topology structure has a leader drone as a central node, and issues instructions and collects data with subordinate drones through direct communication; all subordinate drones perform tasks around the leader drone to ensure centralized management and efficient coordination of tasks; the leader drone serves as a core node for communication and task command, and is responsible for assigning tasks, planning paths, and monitoring task progress; In the star topology, some slave drones use other nodes as relays to ensure that remote drones can receive mission instructions and feed data back to the leader drone, maintaining connectivity within the cluster through multi-hop communication. The multi-hop communication mechanism ensures that communication connections and mission instructions can be maintained even when the mission area is large or the drones fly far away.

[0006] Furthermore, CBS combined with the improved RRT* algorithm is used to generate a conflict-free path when planning the path: first, the improved RRT* algorithm performs random sampling on a plane at a specific altitude layer to generate a preliminary path based on the starting point, end point and flight altitude mission requirements of the UAV, ensuring that the path planning meets the flight altitude restrictions; on this basis, the CBS algorithm performs dynamic conflict detection by detecting the overlap points of the nodes of each path with other UAV paths in real time. If a potential conflict is detected, the system automatically calls the improved RRT* algorithm to replan the conflicting path segment to ensure conflict-free execution of the path.

[0007] Furthermore, the process of generating conflict-free paths by combining CBS with the improved RRT* algorithm is: S1: After setting the mission requirements (including starting point, end point, flight altitude and other parameters), the leader UAV first uses the improved RRT* algorithm to generate a preliminary path. The improved RRT* algorithm performs layered sampling in three-dimensional space, generates path points independently on the plane of each height layer and combines them to generate a complete path. The specific process is: 1) Define the sampling space and layered altitude. The altitude range of the flight area is divided into several independent altitude layers Z1, Z2, ..., Zk. The path generation of each altitude layer is performed independently. For the kth altitude layer, the path points are sampled only in the X and Y planes, while Zk remains constant. The path generation formula is expressed as:

[0008] Among them, Pathk represents the set of path points generated at the kth altitude layer, It represents the path points generated at the altitude layer Zk, where Zk is the flight altitude of the layer and n is the number of path points; 2) After the sampling point set in each altitude layer is generated, the complete three-dimensional path is formed by merging the path points of multiple altitude layers. The final path is defined as:

[0009] Among them, K is the total number of height layers, Path final represents the final path; S2: Use the CBS algorithm to perform dynamic conflict detection on the paths generated by the improved RRT* algorithm: The CBS algorithm decomposes the path planning task of multiple drones into a constraint tree structure (conflict when finding paths). In the conflict detection stage, the system detects the overlap of the path points of each drone with the path points of other drones in real time. When the node positions of two paths overlap at time t, a path conflict occurs, and the following conditions are met:

[0010] Among them, Path p (t) and Path q (t) represents the path point set of the two UAVs, If there is a conflict, the CBS algorithm will create a new node in the constraint tree structure and impose constraints to ensure that the conflicting paths no longer overlap; S3: When CBS detects a path conflict, the system automatically calls the improved RRT* algorithm to regenerate the conflicting path segment in the constrained area; the improved RRT* algorithm gives priority to avoiding the conflicting area when sampling, and searches for a feasible path under the constraints. The path adjusted by the improved RRT* algorithm is defined as:

[0011] Among them, Path adjusted Represents the regenerated path set, represents the adjusted path points, n′ is the number of adjusted path points; S4: The conflict-free path generated by combining the CBS algorithm with the improved RRT* algorithm will be used as the final path and assigned by the leader UAV to each subordinate UAV for execution. Each UAV will perform flight according to the mission requirements and the assigned path to ensure that there is no conflict in the paths within the cluster and the flight altitude is controlled.

[0012] Furthermore, the UAV cluster collaborative system realizes efficient task planning and execution through a self-organizing network, and the specific process is: S1: When the mission is initialized, the leader drone receives external mission requirements, performs mission analysis and path planning. The leader drone builds the three-dimensional map required for the mission based on the environmental information of the mission area (such as terrain, obstacles, weather conditions, etc.) and plans the optimal flight route; S2: After the path planning is completed, the leader UAV distributes the mission instructions to each subordinate UAV through the self-organizing network. The mission instructions include the flight route, mission area and mission execution sequence information. The leader UAV ensures that the tasks of all subordinate UAVs are reasonably allocated to avoid task overlap or conflict; S3: After the mission instructions are distributed, each slave UAV starts to execute the mission according to the received mission instructions. The slave UAV uses built-in sensors (such as lidar, infrared sensors, etc.) to detect the environment in real time, and flies according to the flight path planned by the mission, while collecting data; S4: During the mission execution, each slave UAV maintains real-time communication with the leader UAV through the ad hoc network. The leader UAV conducts a comprehensive analysis based on the received mission progress data. If the environment changes dynamically (such as new obstacles appearing or the path is blocked), the leader UAV will dynamically adjust the mission plan based on the new environmental information and issue new mission instructions in real time; S5: During the mission execution, if a subordinate drone is unable to continue to perform the mission (such as communication failure, flight path obstruction, etc.), the system will automatically trigger the backup drone election mechanism, and another drone will take over its mission to ensure the continuity of the mission. The backup election mechanism is automatically completed through the self-organizing network and does not affect the overall mission execution; S6: After the mission is completed, all subordinate drones will transmit the collected data back to the leader drone through the self-organizing network. The leader drone will integrate and process the data, generate a mission report, and confirm the mission completion status. All drones will return to the initial position or the preset recovery point according to the instructions, and the mission will be successfully completed.

[0013] Furthermore, the real-time Dijkstra variant algorithm is used to enhance the dynamic optimization capability of the communication path and the adaptability of the system. The real-time Dijkstra variant algorithm monitors the status information of the drone nodes in real time, including distance, power, and signal strength, and dynamically adjusts the communication path weight to ensure the communication efficiency and stability of each drone node in a complex environment. The specific process is: In each communication cycle, the leader UAV dynamically calculates the communication path weight based on the status information of each slave UAV. The communication path weight W ij represents the communication cost from node i to node j, and the calculation formula is:

[0014] Where: d ij is the distance from node i to node j. The shorter the distance, the lower the path cost. j is the signal strength of node j. The higher the signal strength, the lower the path cost. j is the power of node j. The higher the power, the lower the path cost. α, β, and γ are dynamic weight coefficients, which reflect the relative importance of distance, signal strength, and power, respectively. Select the relay node with the lowest communication cost as the relay node. During the flight of the drone cluster, set the communication path cost threshold T and state threshold E. If the state of a relay node is lower than the preset state threshold E, the system will automatically trigger the communication path update; or when W ij >When the communication path cost threshold T is reached, the real-time Dijkstra variant algorithm will recalculate the optimal communication path; after the communication path is updated, a new relay node will be selected so that the drone cluster can still maintain a stable communication link when the node status changes frequently; To avoid overload caused by frequent use of some relay nodes, a load balancing mechanism is integrated into the real-time Dijkstra variant algorithm. The load factor L of each relay node is j Calculated as:

[0015] Among them: F j is the current data transmission frequency of node j, C j is the communication capability of node j, which is proportional to the signal strength and power; When the node load factor L j When the preset load threshold Lth is exceeded, the system will exclude the node from the candidate list of relay nodes and give priority to nodes with lower load for data transmission.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The system of the present invention adopts a star topology structure, and centrally controls the subordinate drones through the leader drone, ensuring the efficient collaborative operation of the drone cluster in a complex environment, and has self-organizing network functions, dynamic obstacle avoidance and real-time communication capabilities.

[0017] The system of the present invention can dynamically adjust the communication path according to the position of the drone during flight, the quality of the communication link and the mission requirements. When the drone is flying in a complex environment or a long-distance mission area, the adaptive routing can intelligently select the best communication path to ensure that data and instructions can be transmitted with the lowest delay. This function is particularly suitable for mission scenarios with large areas and complex terrain. By dynamically optimizing the communication path, it ensures that the mission instructions of the drone cluster can be transmitted in real time, improving the stability of mission execution and the overall communication efficiency.

[0018] The system of the present invention supports multi-hop communication and relay functions. Through self-organizing network technology, the slave drones in the cluster can serve as communication relay nodes to help transmit data and instructions of other remote drones, ensuring that even if they are beyond the direct communication range of the leader drone, the drones in the cluster can still maintain a stable communication connection, ensuring that mission instructions are efficiently transmitted within a larger mission area, and adapting to a variety of complex terrains and environments, expanding the communication coverage of the drone cluster. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a functional structure diagram of each module of the leader drone of the drone cluster cooperative system of the present invention.

[0020] Figure 2 It is a functional structure diagram of each module of the slave UAV of the UAV cluster cooperative system of the present invention.

[0021] Figure 3 It is a schematic diagram of the star topology structure of the drone cluster of the drone cluster collaborative system of the present invention.

[0022] Figure 4 It is a flowchart of the planning and execution of UAV cluster tasks based on a self-organizing network in the present invention. DETAILED DESCRIPTION

[0023] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments, but this is not intended to limit the scope of protection of the present application.

[0024] The UAV cluster coordination system based on the self-organizing network of the present invention adopts a star topology structure. The subordinate UAVs are nodes for executing specific tasks, and the leader UAV is the core node of the cluster control system, which undertakes the functions of global task planning, environmental perception, obstacle avoidance, communication coordination, etc. In the cluster, the leader UAV is relatively fixed and directly controls the subordinate UAVs without the need to communicate with the ground. When the leader UAV fails, a new leader UAV can be elected.

[0025] The path planning part of the task planning and management module in the present invention provides three-dimensional map support for path planning through the environmental perception and mapping module. The leading UAV first perceives the environmental information of the task area in real time based on multi-source sensor data (such as lidar, visual sensing, etc.), builds a three-dimensional map containing terrain, buildings and obstacles, and transmits the map data to the task planning and management module. Based on the map constructed by environmental perception, the path planning part uses CBS (conflict base search) combined with the improved RRT* algorithm to generate a conflict-free path. First, the improved RRT* algorithm randomly samples on a plane at a specific altitude layer according to the starting point, end point and flight altitude mission requirements of the UAV to generate a preliminary path to ensure that the path planning meets the flight altitude limit; on this basis, the CBS algorithm performs dynamic conflict detection, and detects the overlap points of the nodes of each path with the paths of other UAVs in real time. If a potential conflict is detected, the system automatically calls the improved RRT* algorithm to re-plan the conflicting path segment to ensure the conflict-free execution of the path.

[0026] The star topology structure uses the leader drone as the central node, which issues instructions and collects data with the subordinate drones through direct communication. All subordinate drones perform tasks around the leader drone to ensure centralized management and efficient coordination of tasks. As the core node of communication and task command, the leader drone is responsible for assigning tasks, planning routes and monitoring task progress. Through the star topology structure, the leader drone can maintain direct communication with each subordinate drone to ensure accurate transmission of task instructions. In the multi-hop communication and communication relay, under the star topology, some subordinate drones can use other nodes as relays to ensure that remote drones can receive task instructions and feed data back to the leader drone. The multi-hop communication mechanism effectively expands the communication coverage of the system, ensuring that communication connections and task instruction transmission can be maintained even when the mission area is large or the drone flies a long distance.

[0027] Specifically, Leader drones, such as Figure 1 As shown, the leader UAV is divided into mission planning and management module, environment perception and map building module, dynamic obstacle avoidance and flight control module, and communication and coordination module.

[0028] The mission planning and management module performs path planning according to mission requirements, sets mission objectives, flight routes and execution sequences for subordinate UAVs, and monitors the mission execution progress of subordinate UAVs and dynamic changes during mission execution in real time. It reallocates tasks or optimizes paths according to dynamic changes (such as environmental changes or subordinate UAV failures), and stores mission-related data for subsequent analysis and optimization of mission execution.

[0029] The environmental perception and map construction module uses multi-source sensors such as lidar, visual cameras, infrared sensors, etc. to perceive environmental information such as obstacles and terrain changes in the mission area, collect and integrate environmental data to build a three-dimensional dynamic map, and update map information in real time to adapt to complex mission environments.

[0030] The dynamic obstacle avoidance and flight control module detects dynamic obstacles in the mission area in real time, adjusts the UAV's flight path based on environmental perception data, and dynamically adjusts the UAV's flight altitude based on mission requirements and environmental changes to ensure the UAV's flight safety and mission execution stability.

[0031] The communication and coordination module manages the communication routes of the drone cluster, uses multi-hop communication to ensure smooth communication among all drones in the cluster, and uses the self-organizing network function to ensure that mission instructions and data can be smoothly transmitted between drones. Especially when it is beyond the direct communication range, other subordinate drones can act as relay nodes to maintain the connectivity of the communication link.

[0032] Slave drones, such as Figure 2 As shown, the slave UAV is divided into a mission execution module, a dynamic obstacle avoidance and flight control module, and a communication and relay module.

[0033] The mission execution module receives and parses the mission instructions sent by the leader drone, and executes the mission according to the planned path, including the flight path, target point, mission parameters, etc. It will fly along the set route according to the mission instructions and provide status feedback during the mission.

[0034] The dynamic obstacle avoidance and flight control module detects local obstacles and performs obstacle avoidance operations during flight to ensure the safe flight of the UAV in complex environments, and timely adjusts the flight altitude, speed and direction to ensure the stable flight of the UAV and follow the path planning of the leader UAV.

[0035] The communication and relay module is responsible for receiving mission instructions from the leader UAV, and providing feedback on flight status, mission progress, and collected environmental data. It transmits data between the leader UAV and other subordinate UAVs, acts as a communication relay, and ensures smooth communication links within the system.

[0036] CBS combines the improved RRT* algorithm to generate multi-UAV paths without conflict paths, ensuring efficient and safe flight of UAV clusters in complex environments. The specific process is: S1: After setting the mission requirements (including starting point, end point, flight altitude and other parameters), the leader UAV first uses the improved RRT* algorithm to generate a preliminary path. The improved RRT* algorithm performs layered sampling in three-dimensional space, generates path points independently on the plane of each height layer and combines them to generate a complete path. The specific process is: 1) Define the sampling space and layered altitude. The altitude range of the flight area is divided into several independent altitude layers Z1, Z2, ..., Zk. The path generation of each altitude layer is performed independently. For the kth altitude layer, the path points are sampled only in the X and Y planes, while Zk remains constant. The path generation formula is expressed as:

[0037] Among them, Pathk represents the set of path points generated at the kth altitude layer, It represents the path points generated at the altitude layer Zk, where Zk is the flight altitude of the layer and n is the number of path points; 2) After the sampling point set in each altitude layer is generated, the complete three-dimensional path is formed by merging the path points of multiple altitude layers. The final path is defined as:

[0038] Among them, K is the total number of height layers, Path final represents the final path; S2: Use the CBS algorithm to perform dynamic conflict detection on the paths generated by the improved RRT* algorithm: The CBS algorithm decomposes the path planning task of multiple drones into a constraint tree structure (conflict when finding paths). In the conflict detection stage, the system detects the overlap of the path points of each drone with the path points of other drones in real time. When the node positions of two paths overlap at time t, a path conflict occurs, and the following conditions are met:

[0039] Among them, Path p (t) and Path q (t) represents the path point set of the two UAVs, If there is a conflict, the CBS algorithm will create a new node in the constraint tree structure and impose constraints to ensure that the conflicting paths no longer overlap; S3: When CBS detects a path conflict, the system automatically calls the improved RRT* algorithm to regenerate the conflicting path segment in the constrained area; the improved RRT* algorithm gives priority to avoiding the conflicting area when sampling, and searches for a feasible path under the constraints. The path adjusted by the improved RRT* algorithm is defined as:

[0040] Among them, Path adjusted Represents the regenerated path set, represents the adjusted path points, n′ is the number of adjusted path points; S4: The conflict-free path generated by combining the CBS algorithm with the improved RRT* algorithm will be used as the final path and assigned by the leader UAV to each subordinate UAV for execution. Each UAV will perform flight according to the mission requirements and the assigned path to ensure that there is no conflict in the paths within the cluster and the flight altitude is controlled.

[0041] The CBS of the present invention combines the improved RRT* algorithm to generate a conflict-free path, which can dynamically adjust to ensure the real-time resolution of path conflicts, improve the flexibility of path planning and the safety of UAVs. Each UAV performs flight according to mission requirements and assigned paths, ensuring that there are no conflicts in the paths within the cluster and the flight altitude is controlled.

[0042] The improved RRT* algorithm in the present invention ensures efficient path planning for drones in complex environments, and performs random sampling based on different altitude layers, thereby optimizing path generation and obstacle avoidance capabilities. The method performs layered sampling in three-dimensional space, independently generates path points on the plane of each altitude layer and combines them to generate a complete path. The final path contains path points at different altitude layers, forming a conflict-free, smooth three-dimensional path to meet the path requirements of drones in complex environments.

[0043] Star topology, such as Figure 3 As shown in the figure, the leader drone acts as a central node to maintain direct communication with all subordinate drones to ensure centralized management and efficient coordination of tasks. Under the star topology, all subordinate drones perform tasks around the leader drone to ensure accurate communication of task instructions and real-time feedback of data. When some subordinate drones exceed the direct communication range of the leader drone, the system expands the communication coverage through a multi-hop communication mechanism to ensure that the remote drones can maintain communication connection. In order to further improve the communication efficiency and adaptability of the system in a complex dynamic environment, the system preferably introduces a real-time Dijkstra variant algorithm. On the basis of the star topology and multi-hop communication, more efficient communication management is achieved through dynamic path optimization and load balancing. The real-time Dijkstra variant algorithm optimizes the communication path in real time according to the status information of each node (such as distance, signal strength and power), and gives priority to nodes with high signal strength and sufficient power as relays. When the state of a relay node changes (such as insufficient power or weakened signal), the system will automatically select a new relay node to ensure the continuity and stability of data transmission. At the same time, to prevent a single relay node from being overloaded, the system dynamically allocates data traffic through a load balancing mechanism, giving priority to nodes with lower loads as relays, effectively balancing the communication load, extending the flight time of the drone, and improving the overall efficiency of cluster communications.

[0044] Preferably, the real-time Dijkstra variant algorithm is used to enhance the dynamic optimization capability of the communication path and the adaptability of the system. The real-time Dijkstra variant algorithm dynamically adjusts the communication path weight by real-time monitoring of the status of the drone nodes (including distance, power, signal strength, etc.) to ensure the communication efficiency and stability of each drone node in a complex environment; the specific process is: In each communication cycle, the leader UAV dynamically calculates the communication path weight based on the status information of each slave UAV (including power, signal strength and distance). The communication path weight W ij represents the communication cost from node i to node j, and the calculation formula is:

[0045] Where: d ij is the distance from node i to node j. The shorter the distance, the lower the path cost. j is the signal strength of node j. The higher the signal strength, the lower the path cost. j is the power of node j. The higher the power, the lower the path cost. α, β, and γ are dynamic weight coefficients, which reflect the relative importance of distance, signal strength, and power, respectively. Select the relay node with the lowest communication cost as the relay node. During the flight of the drone cluster, set the communication path cost threshold T and state threshold E. If the state of a relay node (such as power or signal strength) is lower than the preset state threshold E, the system will automatically trigger the communication path update and recalculate the communication cost; or when W ij >When the communication path cost threshold T is reached, the real-time Dijkstra variant algorithm will recalculate the optimal communication path and select new relay nodes, so that the drone cluster can still maintain a stable communication link when the node status changes frequently; To avoid overload caused by frequent use of some relay nodes, a load balancing mechanism is integrated into the real-time Dijkstra variant algorithm. The load factor L of each relay node is j Calculated as:

[0046] Among them: F j is the current data transmission frequency of node j, C j is the communication capability of node j, which is proportional to the signal strength and power; When the node load factor L j When the preset load threshold Lth is exceeded, the system will exclude the node from the candidate list of relay nodes and give priority to nodes with lower load for data transmission.

[0047] The communication method of the present invention helps to balance the communication load, extend the flight time of the UAV and improve the overall operating efficiency of the system.

[0048] Preferably, during the execution of a mission by a drone cluster, if the leader drone cannot continue to work due to power exhaustion, equipment failure or other unforeseen reasons, the system has the function of automatically electing a backup leader. When the leader drone fails, the system will automatically elect a slave drone with good performance as the new leader to take over the command task through the status feedback of the slave drones (including parameters such as power, communication quality, and mission progress). The backup leader election mechanism ensures the continuity of the mission, prevents the drone cluster from stagnating when the leader fails, and effectively improves the system's fault tolerance and the reliability of mission completion.

[0049] The drone cluster system of the present invention realizes efficient task planning and execution through self-organizing network. The specific process is: S1: When the mission is initialized, the leader drone receives external mission requirements and performs mission analysis and path planning through its mission planning and management module. The leader drone builds the three-dimensional map required for the mission based on the environmental information of the mission area (such as terrain, obstacles, weather conditions, etc.) and plans the optimal flight route.

[0050] S2: After the path planning is completed, the leader UAV distributes the mission instructions to the subordinate UAVs through the self-organizing network. The mission instructions include information such as the flight route, mission area, and mission execution order. The leader UAV ensures that the tasks of all subordinate UAVs are reasonably allocated to avoid task overlap or conflict.

[0051] S3: After the mission instructions are distributed, each slave UAV starts to execute the mission according to the received mission instructions. They use built-in sensors (such as lidar, infrared sensors, etc.) to detect the environment in real time, fly according to the flight path planned by the mission, and collect data at the same time.

[0052] S4: During the mission execution, each slave drone maintains real-time communication with the leader drone through the self-organizing network, and the leader drone conducts a comprehensive analysis based on the received mission progress data. If the environment changes dynamically (such as new obstacles appearing or the path is blocked), the leader drone will dynamically adjust the mission plan based on the new environmental information and issue new mission instructions in real time.

[0053] S5: During the mission execution, if a subordinate drone is unable to continue the mission (such as communication failure, flight path obstruction, etc.), the system will automatically trigger the backup drone election mechanism, and another drone will take over its mission to ensure the continuity of the mission. The backup election mechanism is automatically completed through the self-organizing network and does not affect the overall mission execution.

[0054] S6: After the mission is completed, all subordinate drones transmit the collected data back to the leader drone through the self-organizing network. The leader drone integrates and processes the data, generates a mission report, and confirms the mission completion status. All drones return to the initial position or the preset recovery point according to the instructions, and the mission is successfully completed.

[0055] Any matters not described in the present invention are applicable to the prior art.

Claims

1. A drone cluster collaboration system based on a self-organizing network, characterized in that: The system adopts a star topology structure, in which the leader drone serves as the central node, and issues commands and collects data with the subordinate drones through direct communication; All subordinate drones execute tasks around the leader drone, ensuring centralized management and efficient coordination of tasks; The leader drone serves as the core node for communication and mission command, responsible for assigning tasks, planning routes, and monitoring mission progress; In the star topology, some slave drones use other nodes as relays to ensure that remote drones can receive mission instructions and feed data back to the leader drone, maintaining connectivity within the cluster through multi-hop communication. The multi-hop communication mechanism ensures that communication connections and mission instructions can be maintained even when the mission area is large or the drones fly far away.

2. The unmanned aerial vehicle cluster collaborative system based on ad hoc network according to claim 1 is characterized in that: When planning the path, CBS combined with the improved RRT* algorithm is used to generate a conflict-free path: First, the improved RRT* algorithm performs random sampling on a plane at a specific altitude layer to generate a preliminary path based on the starting point, end point and flight altitude mission requirements of the UAV, ensuring that the path planning meets the flight altitude restrictions; on this basis, the CBS algorithm performs dynamic conflict detection, by real-time detection of the nodes of each path and the overlap points of other UAV paths. If a potential conflict is detected, the system automatically calls the improved RRT* algorithm to re-plan the conflicting path segment to ensure conflict-free execution of the path.

3. The unmanned aerial vehicle cluster coordination system based on ad hoc network according to claim 2 is characterized in that: The process of generating conflict-free paths by combining CBS with the improved RRT* algorithm is: S1: After setting the mission requirements, the leader UAV first uses the improved RRT* algorithm to generate a preliminary path. The improved RRT* algorithm performs layered sampling in three-dimensional space, generates path points independently on the plane of each height layer and combines them to generate a complete path. The specific process is: 1) Define the sampling space and layered altitude. The altitude range of the flight area is divided into several independent altitude layers Z1, Z2, ..., Zk. The path generation of each altitude layer is performed independently. For the kth altitude layer, the path points are sampled only in the X and Y planes, while Zk remains constant. The path generation formula is expressed as: Among them, Pathk represents the set of path points generated at the kth altitude layer, It represents the path points generated at the altitude layer Zk, where Zk is the flight altitude of the layer and n is the number of path points; 2) After the sampling point set in each altitude layer is generated, the complete three-dimensional path is formed by merging the path points of multiple altitude layers. The final path is defined as: Among them, K is the total number of height layers, Path final represents the final path; S2: Use the CBS algorithm to perform dynamic conflict detection on the paths generated by the improved RRT* algorithm: The CBS algorithm decomposes the path planning task of multiple drones into a constraint tree structure. In the conflict detection stage, the system detects the overlap of each drone’s path points with other drones’ path points in real time. When the node positions of two paths overlap at time t, a path conflict occurs, and the following conditions are met: Among them, Path p (t) and Path q (t) represents the path point set of the two UAVs, If there is a conflict, the CBS algorithm will create a new node in the constraint tree structure and impose constraints to ensure that the conflicting paths no longer overlap; S3: When CBS detects a path conflict, the system automatically calls the improved RRT* algorithm to regenerate the conflicting path segment in the constrained area; the improved RRT* algorithm gives priority to avoiding the conflicting area when sampling, and searches for a feasible path under the constraints. The path adjusted by the improved RRT* algorithm is defined as: Among them, Path adjusted Represents the regenerated path set, represents the adjusted path points, n′ is the number of adjusted path points; S4: The conflict-free path generated by combining the CBS algorithm with the improved RRT* algorithm will be used as the final path and assigned by the leader UAV to each subordinate UAV for execution. Each UAV will perform flight according to the mission requirements and the assigned path to ensure that there is no conflict in the paths within the cluster and the flight altitude is controlled.

4. The unmanned aerial vehicle cluster coordination system based on ad hoc network according to claim 1, characterized in that: The UAV swarm coordination system realizes efficient task planning and execution through self-organizing network. The specific process is: S1: When the mission is initialized, the leader drone receives external mission requirements, performs mission analysis and path planning. The leader drone builds the three-dimensional map required for the mission based on the environmental information of the mission area and plans the optimal flight route. S2: After the path planning is completed, the leader UAV distributes the mission instructions to each subordinate UAV through the self-organizing network. The mission instructions include the flight route, mission area and mission execution sequence information. The leader UAV ensures that the tasks of all subordinate UAVs are reasonably allocated to avoid task overlap or conflict; S3: After the mission instructions are distributed, each slave UAV starts to execute the mission according to the received mission instructions. The slave UAV uses built-in sensors to detect the environment in real time, flies according to the flight path planned by the mission, and collects data at the same time; S4: During the mission execution, each slave drone maintains real-time communication with the leader drone through the ad hoc network. The leader drone conducts a comprehensive analysis based on the received mission progress data. If the environment changes dynamically, the leader drone will dynamically adjust the mission plan based on the new environmental information and issue new mission instructions in real time; S5: During the mission execution, if a subordinate drone is unable to continue to perform the mission, the system will automatically trigger the backup drone election mechanism, and another drone will take over its mission to ensure the continuity of the mission. The backup election mechanism is automatically completed through the self-organizing network and does not affect the overall mission execution; S6: After the mission is completed, all subordinate drones will transmit the collected data back to the leader drone through the self-organizing network. The leader drone will integrate and process the data, generate a mission report, and confirm the mission completion status. All drones will return to the initial position or the preset recovery point according to the instructions, and the mission will be successfully completed.

5. The unmanned aerial vehicle cluster coordination system based on ad hoc network according to claim 1, characterized in that: The real-time Dijkstra variant algorithm is used to enhance the dynamic optimization capability of the communication path and the adaptability of the system. The real-time Dijkstra variant algorithm monitors the status information of the drone nodes in real time, including distance, power, and signal strength, and dynamically adjusts the communication path weight to ensure the communication efficiency and stability of each drone node in a complex environment. The specific process is: In each communication cycle, the leader UAV dynamically calculates the communication path weight based on the status information of each slave UAV. The communication path weight W ij represents the communication cost from node i to node j, and the calculation formula is: Where: d ij is the distance from node i to node j. The shorter the distance, the lower the path cost. j is the signal strength of node j. The higher the signal strength, the lower the path cost. j is the power of node j. The higher the power, the lower the path cost. α, β, and γ are dynamic weight coefficients, which reflect the relative importance of distance, signal strength, and power, respectively. Select the relay node with the lowest communication cost as the relay node. During the flight of the drone cluster, set the communication path cost threshold T and state threshold E. If the state of a relay node is lower than the preset state threshold E, the system will automatically trigger the communication path update; or when W ij >When the communication path cost threshold T is reached, the real-time Dijkstra variant algorithm will recalculate the optimal communication path; after the communication path is updated, a new relay node will be selected so that the drone cluster can still maintain a stable communication link when the node status changes frequently; To avoid overload caused by frequent use of some relay nodes, a load balancing mechanism is integrated into the real-time Dijkstra variant algorithm. The load factor L of each relay node is j Calculated as: Among them: F j is the current data transmission frequency of node j, C j is the communication capability of node j, which is proportional to the signal strength and power; When the node load factor L j When the preset load threshold Lth is exceeded, the system will exclude the node from the candidate list of relay nodes and give priority to nodes with lower load for data transmission.

Images