Distributed dynamic coordination method and device for unmanned aerial vehicle group

Abstract

The application provides a kind of unmanned aerial vehicle group distributed dynamic cooperation method and device, it is related to unmanned aerial vehicle scheduling technical field, the information state vector and energy state vector of each unmanned aerial vehicle are collected in the application, the remaining endurance time is calculated, neighbor list is constructed based on maximum information sending radius and state beacon is exchanged;According to the remaining endurance time, temporarily select the center of collection and divide unmanned aerial vehicle collection, the longest endurance unmanned aerial vehicle in unmanned aerial vehicle collection is set as center unmanned aerial vehicle;Set monitoring period, based on energy-load balancing factor dynamically updates the center of collection, and the subtask is distributed by center unmanned aerial vehicle;Task correlation weight and remaining endurance time are comprehensively calculated, and information consensus item and energy consensus item are calculated respectively, and then the actual position of each member unmanned aerial vehicle in unmanned aerial vehicle collection is determined, the task cooperation and position optimization of unmanned aerial vehicle group under energy constraint are realized, and the overall endurance capability and task execution robustness of cluster are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned aerial vehicle scheduling, in particular to a distributed dynamic coordination method and device for a group of unmanned aerial vehicles. BACKGROUND

[0002] With the wide application of unmanned aerial vehicle technology in fields such as cooperative monitoring and disaster rescue, the dynamic coordination capability of the group of unmanned aerial vehicles has become a key factor restricting the system efficiency. In the prior art, the coordination of the group of unmanned aerial vehicles is mostly based on centralized control or fixed clustering strategy, which has the problems of single-point failure risk, concentrated communication pressure, and inability to adapt to dynamic energy decay and task changes. Especially in complex scenarios where the endurance capabilities are uneven and the task distribution is dynamically changing, the traditional methods are difficult to achieve load balancing and efficient cooperation of the cluster, resulting in premature energy depletion of some unmanned aerial vehicles, unreasonable task allocation, and decreased overall sustained operation capability of the cluster. Therefore, there is an urgent need for a group of unmanned aerial vehicles coordination method that can adapt to dynamic changes in energy state and achieve distributed autonomous coordination, in order to improve the overall task execution efficiency and robustness of the cluster.

[0003] In the prior art, the publication number CN119882823A discloses a distributed dynamic coordination method and system for a group of unmanned aerial vehicles, which includes establishing a Markov game model for the group of unmanned aerial vehicles, defining the state of each unmanned aerial vehicle, the environment state, the task state, and the comprehensive state matrix of each unmanned aerial vehicle. Then, real unmanned aerial vehicle task data is collected, resource demand is counted, the environment of the group of unmanned aerial vehicles is initialized, and a simulation environment is constructed. The action strategy neural network is used to generate the action and probability of the unmanned aerial vehicle, the task priority and action are generated by combining the task and environment state matrix, the joint action is formed, and the reward value is calculated. The value evaluation neural network is used to evaluate the value of the current action to generate a trajectory; the loss function of the action strategy and value evaluation network is constructed for parameter updating; after the network parameter is updated, the policy similarity is detected, and the training strategy can be adjusted.

[0004] The main problem of the above-mentioned scheme is that it relies on deep neural networks to generate action strategies and evaluate values, which has large computational overhead in the training and reasoning processes, especially in large-scale groups of unmanned aerial vehicles or high-dynamic task scenarios, making it difficult to meet the real-time coordination decision-making requirements; the trained strategy is often specific to a particular task distribution and environment configuration, and if the task type, number of unmanned aerial vehicles, or environment features change significantly, the model may need to be retrained; the endurance capability of the unmanned aerial vehicle itself is not considered, and the trained strategy may pursue immediate task completion rate, ignoring the maintenance of the overall persistent operation capability of the cluster, which easily leads to the frequent selection of high-energy consumption actions, accelerating the energy depletion of some unmanned aerial vehicles, and ultimately leading to the premature failure of the cluster.

[0005] The above information disclosed in the background section is only intended to strengthen the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a distributed dynamic coordination method and apparatus for unmanned aerial vehicle (UAV) swarms to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A distributed dynamic coordination method for unmanned aerial vehicle (UAV) swarms, comprising the following steps:

[0009] Step 1: Collect the information state vector and energy state vector of each UAV in real time, and calculate the remaining flight time of the UAV in real time based on the energy state vector;

[0010] Step 2: Determine the neighboring drones of each drone based on the maximum information transmission radius to build a neighbor list for each drone, and enable each drone to obtain the status beacon of each neighboring drone in its neighbor list;

[0011] Step 3: Select a temporary assembly center from the list of drones and their neighbors based on the remaining flight time, and combine the drones that are not selected as temporary assembly centers with the temporary assembly centers to construct several drone sets. Set the drone with the longest remaining flight time in the drone set as the center drone, and the other drones as member drones.

[0012] Step 4: Set the monitoring period. In each monitoring period, update the central drone of the drone collection based on the remaining battery life and load of each drone in the drone collection, and assign sub-tasks to the member drones through the central drone.

[0013] Step 5: Determine the expected position of the member drone in the drone ensemble and the task association weight with other member drones based on the sub-tasks assigned to the member drones. Then calculate the information consensus term, calculate the energy consensus term based on the remaining flight time of the drones, and determine the actual position of the drones in the drone ensemble based on the information consensus term and the energy consensus term.

[0014] Furthermore, the information state vector represents the UAV's position coordinates in three-dimensional space, and the energy state vector includes the UAV's remaining available energy and instantaneous power; the UAV's information state vector is: ,in, Indicates the first The information state vector of the drone, Indicates the index of the drone. They represent the first The horizontal, vertical, and coordinates of the drone;

[0015] The energy state vector is: ,in represents the energy state vector of the i-th UAV, represents the energy state vector of the i-th UAV, represents the remaining available energy of the i-th UAV, represents the remaining available energy of the i-th UAV, represents the instantaneous power of the i-th UAV, composed of flight power, computing power and communication power of the UAV; The formula based on which the remaining endurance time of the UAV is calculated is:

[0016]

[0017]

[0018] wherein, represents the remaining endurance time of the i-th UAV. represents the remaining endurance time of the i-th UAV.

[0019] Further, the state beacon comprises the information state vector of the UAV and the remaining endurance time;

[0020] The logic for determining the neighbor list of any UAV is as follows: taking the UAV as a target UAV and other UAVs as to-be-selected UAVs, calculating the spatial distance between the target UAV and the to-be-selected UAVs, if the spatial distance between a to-be-selected UAV and the target UAV is less than the maximum information sending radius of both, then the to-be-selected UAV is taken as a neighbor UAV of the target UAV, and all neighbor UAVs of the target UAV are summarized to construct the neighbor list of the target UAV, wherein the maximum information sending radius represents the farthest distance that the UAV can send its own information.

[0021] Further, the principle of constructing the UAV set is as follows:

[0022] For any UAV, the remaining endurance time of the UAV itself and the remaining endurance time of all neighbor UAVs in the neighbor list of the UAV are obtained, if the remaining endurance time of the UAV itself is greater than the remaining endurance time of all neighbor UAVs, then the UAV is set as a temporary set center, a UAV set with the temporary set center as the center point is constructed, and the upper limit of the set members of the UAV set is preset.

[0023] ​For any drone not selected as a temporary assembly center, determine which temporary assembly center's neighbor list it belongs to. Remove drone sets that have reached the maximum number of members from the drone sets belonging to these temporary assembly centers. Sort the remaining drone sets in descending order of the remaining flight time of the temporary assembly centers. Assign the drone that was not selected as a temporary assembly center to the drone set ranked first. Construct several drone sets in this way. The drone with the highest remaining flight time in a drone set is designated as the center drone of that drone set, and the other drones are designated as member drones. Before assigning tasks, member drones in a drone set communicate with the center drone, but member drones do not communicate with each other.

[0024] Furthermore, the principle underlying the updating of the central drone in the drone ensemble is as follows:

[0025] The formula used to calculate the energy-load balancing factor of drones within a drone ensemble is as follows:

[0026]

[0027] in, Indicates the number of drones in the collection Energy-load balancing factor for drones This represents the index of a drone within a drone set. Indicates the number of drones in the collection The remaining flight time of the drone. This represents the maximum remaining flight time of all drones in the drone ensemble. Indicates the number of drones in the collection The current mission load of the drone. This represents the maximum current task load of all drones within the drone ensemble. Represents the weighting coefficient, and ;

[0028] During the monitoring period, the energy-load balancing factor ratio of each member drone in the drone ensemble and the central drone is calculated in real time. If the maximum value of the energy-load balancing factor ratio exceeds a preset threshold, the member drone corresponding to the maximum value of the energy-load balancing factor ratio is taken as the new central drone, thereby updating the central drone in real time during the monitoring period.

[0029] Furthermore, the principle of calculating consensus terms is as follows:

[0030] Each member drone in the drone swarm has a theoretically optimal position, or desired position, within the drone swarm, based on its assigned subtask.

[0031] Based on expert scoring, the task collaboration requirements of each member drone with other member drones are determined. Then, based on task collaboration, the task association weights between member drones are obtained. For the task association weights of each drone with other member drones in the drone set, an information consensus term is calculated. The formula used is as follows:

[0032]

[0033] in, Indicates the first Information consensus item for each member drone, Indicates the index of the member drone, Indicates the number of member drones. Indicates the first The member drone and the first Task association weights for each member drone Indicates the first The desired location of each member drone. Indicates the first The desired location of each member drone. This represents the adjustment coefficient of the consensus term, and ;

[0034] The formula used to calculate the energy consensus term is:

[0035]

[0036] in, Indicates the first Energy consensus term for each member drone Indicates the first The remaining battery life of each member's drone. This indicates the location of the central drone within the drone ensemble. This represents the adjustment coefficient of the energy consensus term.

[0037] Furthermore, the formula used to determine the actual position of a drone within the drone ensemble based on its information consensus term and energy consensus term is as follows:

[0038]

[0039] in, Indicates the first The actual location of each member drone.

[0040] The present invention also provides a distributed dynamic coordination device for unmanned aerial vehicle (UAV) swarms, the device being used to implement the above-mentioned distributed dynamic coordination method for UAV swarms, specifically including:

[0041] A data acquisition module is configured to acquire information state vectors and energy state vectors of each UAV in real time, and calculate the real-time remaining endurance time of the UAV based on the energy state vectors;

[0042] A feature monitoring module is configured to determine neighbor UAVs of each UAV based on the maximum information transmission radius, so as to construct a neighbor list of each UAV, and enable each UAV to acquire state beacons of each neighbor UAV in the neighbor list of the UAV;

[0043] A set construction module is configured to select temporary set centers from the UAVs and the neighbor lists of the UAVs based on the remaining endurance time, combine the UAVs that are not selected as the temporary set centers with the temporary set centers, and construct a plurality of UAV sets, wherein the UAV with the longest remaining endurance time in each UAV set is set as a center UAV, and the other UAVs are set as member UAVs.

[0044] A task allocation module is configured to set a monitoring period, and update the center UAV of each UAV set based on the remaining endurance time and the load condition of each UAV in the UAV set in each monitoring period, and allocate sub-tasks to the member UAVs by the center UAV.

[0045] A position optimization module is configured to determine the expected positions of the member UAVs in the UAV set and the task association weights of the member UAVs with other member UAVs based on the sub-tasks allocated to the member UAVs, calculate an information consensus term, calculate an energy consensus term based on the remaining endurance time of the UAV, and determine the actual positions of the UAVs in the UAV set based on the information consensus term and the energy consensus term of the UAV.

[0046] Compared with the prior art, the present application has the following advantages:

[0047] The present application can identify the nodes with insufficient energy at an early stage, adjust the task allocation in time, prevent the cluster from failing due to the energy depletion of a single node, and avoid the problems of uneven energy distribution and task interruption caused by static planning, by periodically acquiring the state and continuously monitoring the energy state.

[0048] The application also selects a set center according to the remaining endurance time, and dynamically constructs a UAV set, and a traditional centralized or fixed clustering method is easy to cause system collapse due to center node failure or energy depletion. The method disperses the control responsibility by dynamically selecting multiple set centers, and the system fault tolerance is stronger, the UAV with the longest remaining endurance time is selected as the set center, when the energy-load balance factor of other members is significantly higher than the current set center, the set center replacement is triggered, and the overall efficiency is avoided from being affected by the too large load or the energy decline of the set center; the application also determines the actual position of the UAV by comprehensively calculating the information consensus item and the energy consensus item based on the task cooperation demand and the remaining endurance time, so that the UAV meets the task cooperation demand and takes into account energy saving when adjusting the position, and the short board effect caused by single target optimization is avoided; the task allocation, energy state and position planning are fused, so that the UAV forms a distribution in the physical space which meets the task logic and satisfies the energy constraint. BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 It is a method flowchart of the embodiment of the application;

[0050] Figure 2 It is a fitting curve diagram of the energy-load balance factor changing with the remaining endurance time of the embodiment of the application;

[0051] Figure 3 It is a fitting curve diagram of the energy-load balance factor changing with the task load of the embodiment of the application;

[0052] Figure 4 It is a device module diagram of the embodiment of the application. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical scheme and advantages of the application clearer and more apparent, the application will be further described in detail below with specific embodiments.

[0054] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the application should be understood as the usual meaning understood by those skilled in the art to which the application belongs. The "first", "second" and similar words used in the application do not represent any order, quantity or importance, but are only used to distinguish different components. "Include" or "contain" and similar words mean that the elements or objects before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connected" or "connected" and similar words are not limited to physical or mechanical connection, but can include electrical connection, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to represent the relative positional relationship, when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0055] Example:

[0056] Please see Figures 1 to 4 The present invention provides a technical solution:

[0057] A distributed dynamic coordination method for unmanned aerial vehicle (UAV) swarms, comprising the following steps:

[0058] Step 1: Collect the information state vector and energy state vector of each UAV in real time, and calculate the remaining flight time of the UAV in real time based on the energy state vector;

[0059] In this embodiment, the information state vector is the position coordinates of the UAV in three-dimensional space, and the energy state vector includes the UAV's remaining available energy and instantaneous power; the UAV's information state vector is: ,in, Indicates the first The information state vector of the drone, Indicates the index of the drone. They represent the first The horizontal, vertical, and coordinates of the drone;

[0060] Each drone is equipped with onboard sensors to collect and update its own status data in real time;

[0061] Information state vectors are used to describe the real-time physical position of UAVs in three-dimensional space. The physical position of UAVs is the basis for path planning, task allocation and swarm formation. Without position information, UAVs cannot cooperate in space and are prone to collisions or blind spots. Based on position information, UAVs can calculate the distance to other UAVs, determine whether they are within communication range, and thus establish their respective neighbor lists.

[0062] The energy state vector is: ,in Indicates the first The energy state vector of the drone. Indicates the first The remaining available energy of the drone, Indicates the first The instantaneous power of a drone consists of its flight power, computing power, and communication power.

[0063] The energy state vector includes the UAV's remaining usable energy and instantaneous power. The remaining usable energy is collected in real time by the UAV's onboard battery management system and reflects the total usable energy remaining in the battery at the current moment. The instantaneous power is monitored in real time by power sensors installed on the UAV. The instantaneous power mainly consists of three parts: flight power, computing power, and communication power. Flight power represents the power consumed by the flight control system, such as motors and helix angle. Computational power represents the power consumed by the onboard processor to perform computational tasks such as mission planning and state processing. Communication power represents the power consumed by the wireless communication module to communicate with other UAVs.

[0064] The formula used to calculate the remaining flight time of a drone based on the energy state vector is:

[0065]

[0066] in, Indicates the first The remaining flight time of the drone.

[0067] Remaining flight time reflects the ability of a drone to continuously perform tasks under its current operating conditions. Dividing the drone's remaining available energy by its instantaneous power gives how long the drone can sustain itself under the current power. Remaining available energy reflects the total energy required to maintain the drone's flight, while instantaneous power reflects the rate at which the drone consumes energy. The higher the remaining available energy and the lower the instantaneous power, the more energy the drone still has, and the lower the rate of energy consumption, resulting in a longer remaining flight time.

[0068] Step 2: Determine the neighboring drones of each drone based on the maximum information transmission radius to build a neighbor list for each drone, and enable each drone to obtain the status beacon of each neighboring drone in its neighbor list;

[0069] In this embodiment, the state beacon includes the UAV's information state vector and remaining flight time;

[0070] For any given drone, the logic for determining its neighbor list is as follows: take the drone as the target drone, and the other drones as candidates. Calculate the spatial distance between the target drone and the candidates. If the spatial distance between a candidate drone and the target drone is less than their maximum information transmission radius, then this candidate drone is considered a neighbor drone of the target drone. Summarize all the neighbor drones of the target drone to construct the target drone's neighbor list. The maximum information transmission radius represents the farthest distance that a drone can reach by transmitting its own information.

[0071] The neighbor list is a set of other UAVs that each UAV can identify and record based on its communication ability and spatial position; for any two UAVs , the Euclidean distance between them is calculated based on their information state vectors , and each UAV has a maximum information sending radius, which represents the farthest distance that its wireless communication module can effectively send information, and The maximum information sending radius of UAV , if and , then UAV are each other's neighbor UAVs, and all UAVs are combined in pairs to determine the neighbor list according to the above method until each UAV's neighbor list is determined; in the process of dynamic coordination, the communication ability directly affects the efficiency and reliability of information exchange, and each UAV has different effective communication ranges due to its communication module performance, environmental interference, energy state, and other factors. The two-way reachable determination mechanism based on the maximum information sending radius constructs the neighbor list, and only when both UAVs are within each other's effective communication range can they be identified as neighbor UAVs. If only one-way communication radius is used for determination, it may lead to one-way reachable information and cannot form effective feedback. The two-way determination ensures that neighbor UAVs can stably exchange state information and instructions; the communication ability of the UAV may fluctuate due to factors such as terrain, weather, electromagnetic interference, etc., affecting its maximum information sending radius, and each UAV adjusts the neighbor list according to its real-time communication ability to improve its adaptability to dynamic environments.

[0072] Step 3: Select a temporary set center from the UAV and its neighbor list based on the remaining endurance time, and combine the UAVs that are not selected as temporary set centers with the temporary set center to construct a number of UAV sets, and set the UAV with the longest remaining endurance time in the UAV set as the center UAV, and the other UAVs as member UAVs;

[0073] In this embodiment, the principle of constructing the UAV set is:

[0074] For any UAV, obtain its own remaining endurance time and the remaining endurance time of all neighbor UAVs in its neighbor list, if its own remaining endurance time is greater than that of all neighbor UAVs, then set the UAV as the temporary set center, construct the UAV set with the temporary set center as the center point, and preset the upper limit of the set members of the UAV set;

[0075] Each UAV is locally compared with the UAVs in its neighbor list, only the remaining endurance of the UAV is compared with the UAVs in the neighbor list, to ensure that the decision is based on the real-time state within the local communicable range, and in the local range, multiple UAVs may simultaneously satisfy the condition that "the remaining endurance of the UAV is greater than that of all neighbor UAVs", each UAV that meets the condition can independently become a temporary cluster center, and in the same monitoring period, multiple temporary cluster centers can exist, multiple UAV clusters are generated to realize a distributed multi-cluster collaborative architecture.

[0076] For any UAV that is not selected as a temporary cluster center, determine to which temporary cluster centers its neighbor list belongs, and after excluding the UAV clusters that reach the upper limit of cluster members from the UAV clusters in which the temporary cluster centers are located, arrange the remaining UAV clusters in descending order of the remaining endurance of the temporary cluster centers, and assign the UAV that is not selected as a temporary cluster center to the UAV cluster arranged at the top. In this way, a plurality of UAV clusters are constructed, and the UAV with the highest remaining endurance in the UAV cluster is taken as the center UAV of the UAV cluster, and the other UAVs are taken as member UAVs. Before the task is assigned, the member UAVs and the center UAV in the UAV cluster communicate with each other, and the member UAVs do not communicate with each other.

[0077] For each non-temporary cluster center UAV, scan all temporary cluster centers that list it in the neighbor list, and select the UAV clusters that do not reach the upper limit of cluster members to form an attributable cluster list. The attributable cluster list is arranged in descending order of the remaining endurance of the respective temporary cluster centers, and the UAV cluster arranged at the top is added. Each UAV cluster has a cluster member upper limit to prevent the size of a single cluster from being too large, causing the cluster center to be overloaded or the communication to be congested. The cluster member upper limit is dynamically set based on the task type and the communication bandwidth, and is usually set to 5-15. Before a specific work task is assigned, the communication mode in the UAV cluster is set to mutual communication between the member UAVs and the center UAV, and the member UAVs do not communicate with each other, thereby avoiding redundant state broadcasting between the members, reducing the overall communication complexity in the cluster, saving communication energy, and completing the task assignment by the center UAV, and the member UAVs only need to execute instructions and report states. If a single member UAV has a communication failure or a task execution error, the center UAV can perceive and handle it.

[0078] Step 4: Set a monitoring period, and in each monitoring period, update the center UAV of the UAV cluster based on the remaining endurance and load of each UAV in the UAV cluster, and assign sub-tasks to the member UAVs by the center UAV;

[0079] In this embodiment, the principle on which the center UAV of the UAV cluster is updated is as follows:

[0080] The formula used to calculate the energy-load balancing factor of drones within a drone ensemble is as follows:

[0081]

[0082] in, Indicates the number of drones in the collection Energy-load balancing factor for drones This represents the index of a drone within a drone set. Indicates the number of drones in the collection The remaining battery life of the drone. This represents the maximum remaining flight time of all drones in the drone ensemble. Indicates the number of drones in the collection The current mission load of the drone. This represents the maximum current task load of all drones within the drone ensemble. Represents the weighting coefficient, and ;

[0083] The energy-load balancing factor is used to dynamically assess whether a drone is suitable to serve as a rendezvous center. The energy-load balancing factor reflects the balance between the drone's energy sustainability and mission execution capability. A higher energy-load balancing factor value indicates that the drone has sufficient energy and a lighter load, making it more suitable as a rendezvous center. A lower energy-load balancing factor value indicates that the drone has insufficient energy or an excessive load, making it unsuitable as a rendezvous center. The calculation formula includes… and All are normalized values. This demonstrates that the longer the remaining flight time of a drone, the larger the energy-load balancing factor value. This reflects that the lower the drone's load, the higher the energy-load balancing factor value; remaining flight time and task load These are all dynamic variables collected and calculated in real time by the UAV, reflecting the UAV's real-time status; weighting coefficients are determined according to different mission stages. The value of [value] is determined at the beginning of a mission, when the drones have just started executing the mission, have sufficient energy, and a light mission load. The primary objective is rapid deployment and task allocation. At this stage, the load factor is relatively important, so drones with light loads are prioritized as the central drone to facilitate rapid task allocation. This reflects the primary objective of task allocation; in the middle of the mission, as the mission gradually unfolds and deployments, the load tends to stabilize, and the energy of some drones begins to decline, resulting in differences in endurance. At this point, it is necessary to balance endurance and load conditions to avoid excessive energy consumption at the central node. ; in the later stage of the task, the energy of the UAVs generally decreases, the task may still continue or end, the main goal at this time is to prevent the task from being interrupted or the cluster from failing due to energy depletion, the energy item becomes dominant, the UAV with more remaining energy is preferentially selected as the center, the weight of the energy item is increased, and the weight of the load item is reduced, and the UAV with the maximum energy-load balance factor is selected as the center .

[0084] In the monitoring period, if , the corresponding UAV is taken as the center UAV of the set in the monitoring period, wherein represents the energy-load balance factor of the center UAV.

[0085] In the process of the monitoring period, the energy-load balance factor ratio of each member UAV in the set to the center UAV is calculated in real time, if the maximum value of the energy-load balance factor ratio exceeds a preset threshold, the member UAV corresponding to the maximum value of the energy-load balance factor ratio is taken as the new center UAV, and the center UAV is updated in real time in the monitoring period; the threshold is set to 1.2, if the threshold is set too low, slight fluctuations will cause the center to be replaced frequently, introducing unnecessary communication overhead and causing the risk of task interruption, therefore, only the UAV with significantly better performance will be updated as the new set center; if the threshold is set too high, the performance change cannot be responded in time, resulting in slow updating, the threshold is set to 1.2 to ensure the response speed while reducing unnecessary updating.

[0086] Table 1 reflects the trade-off between energy and load in the selection of the center in the UAV swarm cooperation, and the corresponding task in the middle stage, , , reflects that and will independently and significantly affect , significantly decreases in the early stage of energy decrease or load increase, and is proportional to , and inversely proportional to , when is larger, is smaller, the corresponding is larger, and it is more likely to be selected as the center UAV.

[0087] Table 1. Energy-load balance factor change table

[0088]

[0089] Step 5: Determine the expected position of the member drone in the drone ensemble and the task association weight with other member drones based on the sub-tasks assigned to the member drones. Then calculate the information consensus term, calculate the energy consensus term based on the remaining flight time of the drones, and determine the actual position of the drones in the drone ensemble based on the information consensus term and the energy consensus term.

[0090] In this embodiment, the principle for calculating the consensus term is as follows:

[0091] Each member drone in the drone swarm has a theoretically optimal position, or desired position, within the drone swarm, based on its assigned subtask.

[0092] Based on expert scoring, the task collaboration requirements of each member drone with other member drones are determined. Then, based on task collaboration, the task association weights between member drones are obtained. For the task association weights of each drone with other member drones in the drone set, an information consensus term is calculated. The formula used is as follows:

[0093]

[0094] in, Indicates the first Information consensus item for each member drone, Indicates the index of the member drone, Indicates the number of member drones, Indicates the first The member drone and the first Task association weights for each member drone Indicates the first The desired location of each member drone. Indicates the first The desired location of each member drone. This represents the adjustment coefficient of the consensus term, and ;

[0095] The range of values ​​is The task correlation weight between the member UAVs is determined by the expert scoring method, which reflects the close degree of the two UAVs in performing the task. The greater the task correlation weight, the higher the task cooperation requirement, and the two UAVs tend to be closer to each other. The UAVs that need to cooperate closely in the task should also be close to each other in the physical space, because the strength, bandwidth and stability of wireless communication usually decrease with the increase of distance. The UAVs with a greater task correlation weight need to exchange data frequently, at high speed and with low delay. If the distance is too far, it will lead to unreliable communication link. Therefore, high task cooperation weight corresponds to high communication demand, and the physical distance needs to be shortened to ensure the communication quality. In the cooperative monitoring or search task, the two UAVs need to cover the same area or observe the target from different angles. Therefore, high task cooperation weight corresponds to high observation demand, and the two UAVs need to be kept within a certain distance to form an effective cooperation area. For the two member UAVs, if the task they perform is a cooperative task, such as area scanning, multi-UAV following and multi-UAV lifting, the two UAVs need to keep a relatively fixed distance and attitude to complete the task, and the task correlation weight is high. If the task they perform is an independent task, such as independent patrol, independent observation and independent data acquisition, the task correlation weight is low. First, it is determined whether the task between the member UAVs is cooperative or independent based on the expert scoring method, and then the corresponding task correlation weight is determined.

[0096] The calculation result of the task correlation weight is a vector, which indicates the direction from the expected position of the member UAV to the expected position of the member UAV . Multiplying the vector by means that the higher the task cooperation requirement of the member UAV and , the more they tend to be close to each other. It can be regarded as a direction of the member UAV towards the member UAV . The sum of the position difference between all other member UAVs in the set and the member UAV is summed up and multiplied by the corresponding task correlation weight, which reflects the task cooperation pull of the UAV in the UAV set by all other UAVs. It indicates the moving tendency of under the influence of all related UAVs in the UAV set; the coefficient is used to control the strength of the position adjustment due to the task cooperation, so as to avoid the position from changing sharply due to the excessive task cooperation requirement, which affects the flight stability; according to different task stages, the coefficient is adjusted. In the initial stage, the UAV set needs to quickly form a preliminary formation or spatial distribution that meets the task requirement from the initial scattered state. At this time, the energy is generally sufficient, and the load is zero or very light. A higher Value, requires rapid establishment of task collaboration, retrieval Midway through the mission, the drones have formed a stable formation and are executing the mission as planned. The layout has been largely optimized, and the main objective is fine-tuning rather than reconfiguration, while minimizing energy consumption. In the later stages of the mission, drones generally have low energy levels, making endurance the most pressing constraint. Some drones may be at the critical point of returning to base; any position adjustments made to improve mission coordination will consume energy. Therefore, unnecessary maneuvers should be minimized at this time. .

[0097] The information consensus term reflects the spatial adjustment trend of UAVs under the requirements of mission collaboration. It is used to guide UAVs to optimize the overall collaboration efficiency of the cluster by fine-tuning their positions while meeting the mission collaboration requirements. The information consensus term is proportional to the task association weight between tasks.

[0098] The formula used to calculate the energy consensus term is:

[0099]

[0100] in, Indicates the first Energy consensus term for each member drone Indicates the first The remaining battery life of each member's drone. This indicates the location of the central drone within the drone ensemble. This represents the adjustment coefficient of the energy consensus term.

[0101] The energy consensus term reflects the spatial adjustment tendency of UAVs due to differences in energy state. Specifically, UAVs with lower energy tend to move closer to the central UAV. Indicates member drone The normalized value of the remaining flight time; the larger the value, the better the drone's performance. The more energy available, This reflects the relative insufficiency of the drone's energy; the higher the value, the more inadequate the drone's power. The smaller the value, the closer the drone's energy is to being depleted, and the greater the adjustment range due to energy limitations; Indicates from member drone The desired location points to the location of the central drone. The central drone is usually the node with the most energy and lightest load in the group. It has strong task coordination and resource scheduling capabilities, which allows the lower-energy member drones to move closer to the central drone, reducing their long-distance communication and maneuvering consumption. At the same time, it is convenient for the central drone to adjust the tasks of the lower-energy member drones in emergency situations, and prevent the low-energy drones from losing contact due to energy depletion. The value of the task phase is related to the task phase, and in different task phases, the value is determined according to pushes out the value.

[0102] The formula for determining the actual position of the unmanned aerial vehicle in the unmanned aerial vehicle set based on the information consensus item and the energy consensus item of the unmanned aerial vehicle is:

[0103]

[0104] Wherein, The actual position of the first member unmanned aerial vehicle.

[0105] The expected position of the member unmanned aerial vehicle , that is, only considering the allocated subtask, the best position of the member unmanned aerial vehicle , in the actual flight process, the task coordination of the member unmanned aerial vehicle and other member unmanned aerial vehicles and the endurance of the member unmanned aerial vehicle itself are considered, both of which will affect the position of the member unmanned aerial vehicle , which are the influence of task coordination demand and the influence of energy state respectively. Based on , the spatial layout between the unmanned aerial vehicles is optimized through to improve the task coordination efficiency, and the energy state change is responded through to avoid premature failure of the low-energy node, and finally the reflects the comprehensive optimal position of the unmanned aerial vehicle under the current time, the current task and the energy state, which not only meets the basic demand of task execution, but also enhances the coordination density between the unmanned aerial vehicles through position adjustment; the worse the energy state is, the more the unmanned aerial vehicle tends to be close to the center, reduces the energy consumption of itself, and facilitates the cluster management.

[0106] Please refer to Figure 4 , the application also provides a kind of unmanned aerial vehicle group distributed dynamic coordination device, the device is used to realize the above-mentioned unmanned aerial vehicle group distributed dynamic coordination method, specifically includes:

[0107] Data acquisition module, for real-time acquisition of information state vector and energy state vector of each unmanned aerial vehicle, and calculating the real-time remaining endurance time of the unmanned aerial vehicle based on the energy state vector;

[0108] Feature monitoring module, for determining the neighbor unmanned aerial vehicle of each unmanned aerial vehicle based on the maximum information sending radius, to construct the neighbor list of each unmanned aerial vehicle, so that each unmanned aerial vehicle obtains the state beacon of each neighbor unmanned aerial vehicle in the neighbor list;

[0109] The collection building module is configured to select a temporary collection center from the UAVs and the neighbor list based on the residual endurance time, combine the UAVs that are not selected as the temporary collection center with the temporary collection center, build a plurality of UAV collections, and set the UAV with the longest residual endurance time in the UAV collection as a center UAV and the other UAVs as member UAVs;

[0110] The task allocation module is configured to set a monitoring period, update the center UAV of the UAV collection based on the residual endurance time and load condition of each UAV in the UAV collection in each monitoring period, and allocate sub-tasks to the member UAVs by the center UAV.

[0111] The position optimization module is configured to determine the expected position of the member UAVs in the UAV collection and the task association weight with other member UAVs based on the sub-tasks allocated to the member UAVs, calculate an information consensus term, calculate an energy consensus term based on the residual endurance time of the UAVs, and determine the actual position of the UAVs in the UAV collection based on the information consensus term and the energy consensus term of the UAVs.

[0112] The above formulas are all dimensionless values calculated, the formulas are obtained by collecting a large amount of data to simulate a formula of the most recent real situation, and the preset parameters in the formulas are set by a person skilled in the art according to the actual situation.

[0113] The above embodiments can be realized wholly or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized wholly or partially in the form of a computer program product. Those skilled in the art can realize that the units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized by hardware or software methods depends on the specific application and design constraints of the technical solutions.

[0114] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, which can be located in one place or distributed on multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiments according to actual needs.

[0115] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered within the protection scope of the present application.

Claims

1. A method for distributed dynamic coordination of a UAV swarm, characterized in that, The specific steps include: Step 1: collecting the information state vector and the energy state vector of each unmanned aerial vehicle in real time, and calculating the remaining endurance time of the unmanned aerial vehicle in real time based on the energy state vector; Step 2: determining the neighbor unmanned aerial vehicles of each unmanned aerial vehicle based on the maximum information transmission radius to construct a neighbor list of each unmanned aerial vehicle, and enabling each unmanned aerial vehicle to obtain the state beacon of each neighbor unmanned aerial vehicle in the neighbor list of the unmanned aerial vehicle; Step 3: selecting a temporary cluster center from the unmanned aerial vehicles and the neighbor list based on the remaining endurance time, combining the unmanned aerial vehicles that are not selected as the temporary cluster center with the temporary cluster center, constructing a plurality of unmanned aerial vehicle clusters, setting the unmanned aerial vehicle with the longest remaining endurance time in the unmanned aerial vehicle cluster as a center unmanned aerial vehicle, and setting other unmanned aerial vehicles as member unmanned aerial vehicles; Step 4: setting a monitoring period, updating the center unmanned aerial vehicle of the unmanned aerial vehicle cluster based on the remaining endurance time and the load condition of each unmanned aerial vehicle in the unmanned aerial vehicle cluster in each monitoring period, and distributing sub-tasks to the member unmanned aerial vehicles by the center unmanned aerial vehicle; Step 5: determining the expected position of the member unmanned aerial vehicle in the unmanned aerial vehicle cluster and the task association weight with other member unmanned aerial vehicles based on the sub-tasks distributed to the member unmanned aerial vehicles, calculating an information consensus term based on the expected position, calculating an energy consensus term based on the remaining endurance time of the unmanned aerial vehicle, and determining the actual position of the unmanned aerial vehicle in the unmanned aerial vehicle cluster based on the information consensus term and the energy consensus term of the unmanned aerial vehicle; The principle of calculating the information consensus term is that each member unmanned aerial vehicle in the unmanned aerial vehicle cluster has a theoretically relatively optimal position in the unmanned aerial vehicle cluster, i.e., an expected position, according to the sub-task distributed to the member unmanned aerial vehicle; The task coordination demand of each member unmanned aerial vehicle with other member unmanned aerial vehicles is determined based on the expert scoring method, and the task association weight between the member unmanned aerial vehicles is obtained according to the task coordination, and the information consensus term is calculated for each unmanned aerial vehicle and other member unmanned aerial vehicles in the unmanned aerial vehicle cluster based on the task association weight, and the formula is: The formula for calculating the energy consensus term is: wherein, represents an information consensus item of the first member unmanned aerial vehicle, represents an index of the member unmanned aerial vehicle, represents a number of the member unmanned aerial vehicles, represents a task association weight of the first member unmanned aerial vehicle and the second member unmanned aerial vehicle, represents a desired position of the first member unmanned aerial vehicle, represents a desired position of the first member unmanned aerial vehicle, represents a regulation coefficient of the information consensus item, and ; The formula for determining the actual position of the unmanned aerial vehicle in the unmanned aerial vehicle cluster based on the information consensus term and the energy consensus term of the unmanned aerial vehicle is: wherein, represents the energy consensus item of the i-th member UAV, represents the remaining endurance time of the i-th member UAV, represents the maximum value of the remaining endurance time of all UAVs in the UAV set, represents the position of the central UAV in the UAV set, represents the adjustment coefficient of the energy consensus item;​​ The formula for calculating the remaining endurance time of the unmanned aerial vehicle based on the energy state vector is: wherein, represents the actual position of the nth member drone. 2.The method of claim 1, wherein: The information state vector in step 1 is the position coordinate of the UAV in three-dimensional space, and the energy state vector includes the remaining available energy and the instantaneous power of the UAV; the information state vector of the UAV is: Wherein, represents the index of the i-th UAV, represents the information state vector of the i-th UAV, represents the index of the UAV, represents the horizontal, vertical and vertical coordinates of the i-th UAV, respectively. the horizontal, vertical and vertical coordinates of the i-th UAV, respectively. The energy state vector is: ,in Indicates the first The energy state vector of the drone. Indicates the first The remaining available energy of the drone, Indicates the first The instantaneous power of a drone consists of its flight power, computing power, and communication power. The state beacon in step 2 includes the information state vector and the remaining endurance time of the unmanned aerial vehicle; wherein, represents the first remaining endurance time of the drone. 3.The method of claim 2, wherein: For any unmanned aerial vehicle, the logic for determining the neighbor list of the unmanned aerial vehicle is that the unmanned aerial vehicle is taken as a target unmanned aerial vehicle, and other unmanned aerial vehicles are taken as to-be-selected unmanned aerial vehicles, the spatial distance between the target unmanned aerial vehicle and the to-be-selected unmanned aerial vehicles is calculated, if the spatial distance between a to-be-selected unmanned aerial vehicle and the target unmanned aerial vehicle is less than the maximum information transmission radius of the two unmanned aerial vehicles, the to-be-selected unmanned aerial vehicle is taken as a neighbor unmanned aerial vehicle of the target unmanned aerial vehicle, and all neighbor unmanned aerial vehicles of the target unmanned aerial vehicle are collected to construct a neighbor list of the target unmanned aerial vehicle, and the maximum information transmission radius represents the farthest distance that the unmanned aerial vehicle can transmit its own information. The principle of constructing the unmanned aerial vehicle cluster in step 3 is that 4. The method of claim 1, wherein: ​ For any one of the UAVs, the residual endurance time of the UAV itself and the residual endurance time of all neighbor UAVs in the neighbor list of the UAV are obtained, if the residual endurance time of the UAV itself is greater than the residual endurance time of all neighbor UAVs, the UAV is set as a temporary cluster center, a UAV cluster with the temporary cluster center as the center point is constructed, and the upper limit of the cluster members of the UAV cluster is preset; For any one of the UAVs which is not selected as the temporary cluster center, it is determined to which temporary cluster centers the UAV belongs, after the UAV cluster reaching the upper limit of the cluster members is removed from the UAV clusters of the temporary cluster centers, the residual UAV clusters are arranged in descending order according to the residual endurance time of the temporary cluster centers, the UAV which is not selected as the temporary cluster center is distributed to the UAV cluster arranged in the first place, a plurality of UAV clusters are constructed, and the UAV with the highest residual endurance time in the UAV cluster is taken as the center UAV of the UAV cluster, and the other UAVs are taken as the member UAVs, before the task is distributed, the member UAVs and the center UAV in the UAV cluster communicate with each other, and the member UAVs do not communicate with each other.

5. The method of claim 1, wherein: The principle for updating the center UAV of the UAV cluster in step 4 is that: The energy-load balance factor of the UAV in the UAV cluster is calculated according to the formula: wherein, represents the index of the i-th drone in the set of drones, represents the energy-load balancing factor of the j-th drone, represents the index of the drone in the set of drones, represents the remaining endurance time of the i-th drone, represents the remaining endurance time of the j-th drone, represents the current task load of the i-th drone, represents the current task load of the j-th drone, represents the maximum value of the current task load of all drones in the set of drones, represents the weight coefficient, and ; During the monitoring period, the energy-load balance factor ratio of each member UAV to the center UAV in the UAV cluster is calculated in real time, if the maximum value of the energy-load balance factor ratio exceeds the preset threshold, the member UAV corresponding to the maximum value of the energy-load balance factor ratio is taken as the new center UAV, so that the center UAV is updated in real time within the monitoring period.

6. An apparatus for distributed dynamic coordination of UAV swarm, characterized in that: The device is used to implement the UAV swarm distributed dynamic coordination method of any one of claims 1-5, and specifically comprises: A data acquisition module is configured to acquire the information state vector and the energy state vector of each UAV in real time, and calculate the real-time residual endurance time of the UAV based on the energy state vector; A feature monitoring module is configured to determine the neighbor UAV of each UAV based on the maximum information transmission radius, so as to construct the neighbor list of each UAV, and make each UAV obtain the state beacon of each neighbor UAV in the neighbor list of the UAV; A cluster construction module is configured to select a temporary cluster center from the UAV and the neighbor list of the UAV based on the residual endurance time, combine the UAV which is not selected as the temporary cluster center with the temporary cluster center, construct a plurality of UAV clusters, and set the UAV with the longest residual endurance time in the UAV cluster as the center UAV, and set the other UAVs as the member UAVs; A task distribution module is configured to set a monitoring period, and update the center UAV of the UAV cluster based on the residual endurance time and the load of each UAV in the UAV cluster in each monitoring period, and distribute sub-tasks to the member UAVs by the center UAV. A position optimization module is configured to determine an expected position of the member UAV in the UAV set and a task association weight with other member UAVs based on a subtask assigned to the member UAV, and to calculate an information consensus term, an energy consensus term based on a remaining endurance time of the UAV, and an actual position of the UAV in the UAV set based on the information consensus term and the energy consensus term of the UAV.

Images