Cooperative confrontation control method based on multiple UAVs

Abstract

The present invention relates to the technical field of drone group control, and in particular to a collaborative confrontation control method based on multiple drones. The method comprises: judging the reliability of trajectory prediction of each drone for the target according to the detection area of ​​each drone in the drone sub-cluster for the target, the cosine similarity between the predicted trajectory vector and the actual trajectory vector, and the spatial position distance; judging the success efficiency of each drone for the confrontation with the target according to the number of obstacles in the predicted confrontation path of each drone in the drone sub-cluster, the predicted confrontation time, the distance between the predicted impact position and the target position, and the state of each drone; selecting the first confrontation drone according to the reliability of trajectory prediction of each drone in the drone sub-cluster for the target and the success efficiency of the confrontation with the target. The present invention can solve the problem that in the process of system confrontation of multiple drones, the pressure of the communication network inside the drone cluster is high and the most reliable drone cannot be quickly screened out.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) swarm control, and in particular to a collaborative confrontation control method based on multiple UAVs. Background Art

[0002] As the application scope of drones continues to expand and the types of missions become more and more extensive, the application of drones has gradually developed from a single drone to a cluster of multiple drones. The collaborative confrontation control method based on multiple drones is an advanced technology developed in entertainment competitions. It aims to achieve reconnaissance, attack, defense and other tasks against targets in complex environments through the collaborative work of multiple drones. With the continuous advancement of drone technology and artificial intelligence algorithms, this method can achieve efficient and intelligent task allocation, path planning, information sharing and autonomous decision-making in real-time dynamic environments, enabling drone groups to cope with various complex and ever-changing confrontation scenarios.

[0003] For complex scenarios of multi-UAV coordinated confrontation, the communication topology network within the UAV cluster plays a vital role. This network is responsible for analyzing and processing relevant information of all UAVs to ensure the coordinated combat capability of the UAV cluster. However, as the scale of UAV clusters continues to expand, the pressure on the internal communication network is also increasing, which undoubtedly poses a challenge to the operational efficiency of UAVs. When conducting target attack and defense, it is crucial to quickly screen out the most reliable UAVs to improve combat effectiveness. However, existing systems often find it difficult to complete this task quickly and rely on a lot of manual settings and intervention. This not only increases the complexity and time cost of operations, but may also affect the timeliness and accuracy of combat decisions.

[0004] In addition, when a swarm of multiple drones performs a confrontation mission, in complex confrontation environments, existing technologies often use methods such as increasing the intelligence level of each drone in the drone swarm or using saturation confrontation to win by quantity to achieve successful confrontation mission execution. However, this method is not conducive to cost savings and cannot achieve the effect of successfully executing the confrontation mission at the minimum cost. Summary of the invention

[0005] In order to solve the technical problems in the prior art that the communication network within the drone cluster is under great pressure, the most reliable drones cannot be quickly screened out, and the cost of executing confrontation tasks is high, the purpose of the present invention is to provide a collaborative confrontation control method based on multiple drones, and the technical solutions adopted are as follows:

[0006] A collaborative confrontation control method based on multiple UAVs, the method comprising:

[0007] Step S1: for the drone group network, obtain the sensor data of the airborne equipment of each drone in real time;

[0008] Step S2: A hybrid control structure is adopted. The ground control station initially assigns tasks to each drone. When the drone group reaches a confrontation state, the drone cluster is divided into multiple drone sub-clusters according to the number of tasks. Each drone sub-cluster selects a main drone as the leader node, and the remaining slave drones are obedient nodes.

[0009] Step S3: The leader node of each drone sub-cluster scans and identifies the target, analyzes the target's motion trajectory based on the onboard equipment sensor data of each drone in the drone sub-cluster, and determines the reliability of the trajectory prediction of each drone in the drone sub-cluster for the target based on the detection area of ​​each drone in the drone sub-cluster, the cosine similarity between the predicted trajectory vector and the actual trajectory vector, and the spatial position distance;

[0010] Step S4: judging the success efficiency of each drone in the drone sub-cluster in confronting the target according to the number of obstacles in the predicted confrontation path of each drone in the drone sub-cluster, the predicted confrontation time, the distance between the predicted impact position and the target position, and the status of each drone;

[0011] Step S5: selecting a first confrontation drone according to the reliability of trajectory prediction of each drone in the drone sub-cluster for the target and the success efficiency of confrontation with the target;

[0012] Step S6: Use the first confrontation UAV to perform a confrontation mission on the target.

[0013] Furthermore, the sensor data includes: visual sensor data, radar sensor data, laser sensor data, and ultrasonic sensor data.

[0014] Furthermore, the real-time acquisition of the onboard equipment sensor data of each UAV specifically includes: the ground control station acquires in real time the signal after the equipment sensor data is fused and processed and transmitted by satellite relay.

[0015] Furthermore, the step S2 also includes: each obedient node communicates with the leader node in the corresponding drone sub-cluster at a preset time threshold 1, and when the leader node fails to detect the corresponding obedient node within a preset communication number threshold 2, the obedient node is deemed to be lost.

[0016] Furthermore, step S3 also includes: when the leader node of the drone sub-cluster scans and identifies the target, the leader node uploads the target position mark to the network, and other drone sub-clusters obtain it and select other targets to fight against.

[0017] Furthermore, step S3 also includes: when the detection area of ​​the target by the drones in the drone sub-cluster is larger, the cosine similarity between the predicted trajectory vector and the actual trajectory vector is higher, and the spatial position distance is closer, the reliability of the trajectory prediction of the target by the corresponding drones in the drone sub-cluster is higher.

[0018] Furthermore, the step S4 further includes:

[0019] The fewer the number of obstacles in the predicted confrontation path of the drones in the drone sub-cluster against the target, the shorter the predicted confrontation time, and the closer the distance between the predicted impact position and the target position, the higher the success efficiency of the corresponding drones in the drone sub-cluster against the target;

[0020] The status of each drone includes power status and communication stability. When the power of drones in the drone sub-cluster is higher and the communication stability is better, the corresponding drones in the drone sub-cluster have higher success efficiency in confronting the target.

[0021] Furthermore, the step S5 specifically includes:

[0022] According to the reliability ranking of trajectory prediction of each drone in the drone sub-cluster for the target, combined with the ranking of the successful efficiency of confrontation of each drone in the drone sub-cluster against the target, the drone with the highest reliability of successfully executing the confrontation task is obtained as the first confrontation drone, and the reliability of successfully executing the confrontation task of each drone in the drone sub-cluster is ranked from high to low.

[0023] Furthermore, after step S6, the method further includes: when the first confrontation drone fails to perform the confrontation task, ranking the drones in the drone sub-cluster according to their reliability of successfully performing the confrontation task from high to low, and continuing to perform the confrontation task in sequence until the task is successfully performed.

[0024] Furthermore, when the leader node signal is lost, all subordinate nodes in the corresponding drone sub-cluster select a new leader node from within by election; the election basis includes the controller computing resource usage, communication stability, and airborne equipment sensor response speed of all subordinate nodes.

[0025] The present invention has the following beneficial effects:

[0026] 1. It can solve the problem of high pressure on the communication network within the drone cluster during the multi-drone system confrontation process. By adopting a hybrid control structure, the ground control station initially assigns tasks to each drone. When the drone group reaches the confrontation state, the drone cluster is divided into multiple drone sub-clusters according to the number of tasks. Each drone sub-cluster selects the main drone as the leader node, and the remaining slave drones are obedient nodes. When the leader node of the drone sub-cluster scans and identifies the target, the leader node uploads the target location mark to the network. After other drone sub-clusters obtain it, they select other targets for confrontation. This effectively solves the problem of high pressure on the communication network within the drone cluster.

[0027] Second, it can solve the problem that the most reliable drone cannot be quickly screened out during the collaborative confrontation of multiple drones. The present invention analyzes the motion trajectory of the target based on the onboard equipment sensor data of each drone in the drone sub-cluster, and judges the reliability of the trajectory prediction of each drone in the drone sub-cluster for the target based on the detection area of ​​each drone in the drone sub-cluster for the target, the cosine similarity between the predicted trajectory vector and the actual trajectory vector, and the spatial position distance; judges the success efficiency of each drone in the drone sub-cluster against the target based on the number of obstacles in the predicted confrontation path of each drone in the drone sub-cluster, the predicted confrontation time, the distance between the predicted impact position and the target position, and the state of each drone; selects the first confrontation drone based on the reliability of the trajectory prediction of each drone in the drone sub-cluster for the target and the success efficiency of the confrontation with the target. In this way, the most reliable drone can be quickly screened out during the collaborative confrontation of multiple drones.

[0028] Third, it can solve the problem of high cost of executing confrontation tasks caused by increasing the intelligence level of each drone in the drone cluster or using saturated confrontation to win by quantity in the existing technology. The present invention can effectively reduce the pressure of the communication network inside the drone cluster and reduce the hardware and software costs related to the communication and computing power of each drone by adopting a hybrid control structure, reasonable leader node and subordinate node settings and communication arrangements. At the same time, it adopts a reasonable and most reliable drone screening mechanism to select the drone with the highest reliability in executing confrontation tasks, avoiding the problem of high cost of executing confrontation tasks caused by increasing the intelligence level of each drone in the drone cluster or using saturated confrontation to win by quantity. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0030] Figure 1 A flow chart of a collaborative confrontation control method based on multiple UAVs provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0031] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation method, structure, characteristics and effects of the cooperative confrontation control method based on multiple UAVs proposed by the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. Unless otherwise defined, all technical and scientific terms used in this document have the same meaning as commonly understood by technicians in the technical field of the present invention.

[0032] Example

[0033] The specific scheme of the cooperative confrontation control method based on multiple UAVs provided by the present invention is described in detail below with reference to the accompanying drawings.

[0034] The specific scenario targeted by the embodiments of the present invention is: in the process of coordinated confrontation of multiple drones, through the collaborative work between multiple drones, tasks such as reconnaissance, attack, and defense are achieved against targets in complex environments. Due to the continuous expansion of the scale of drone clusters, the pressure on internal communication networks is also increasing, posing a challenge to the operating efficiency of drones. When conducting target attack and defense, quickly screening out the most reliable drones is crucial to improving combat effectiveness. In addition, on the premise of successfully executing confrontation tasks against targets, it is also very important to select the most economical drone cluster scale and hardware architecture. Based on this, the present invention optimizes the control structure of multi-drone collaboration and quickly analyzes and screens the most reliable drones to perform confrontation tasks.

[0035] See also Figure 1 , which shows a flow chart of a cooperative confrontation control method based on multiple UAVs provided by an embodiment of the present invention, the method comprising the following steps:

[0036] Step S1: For the drone group network, obtain the onboard equipment sensor data of each drone in real time.

[0037] When a drone swarm composed of multiple drones is performing a confrontation task, it is first necessary to self-organize the drone swarm so that the drone group can establish network communication. Then, the sensor data of the onboard equipment of each drone is obtained in real time.

[0038] Specifically, the ground control station obtains the signal of the device sensor data after fusion processing from the satellite relay transmission in real time. At the same time, according to the confrontation task processing process of the drone group, the task completion status is synchronized to the ground control station in real time by obtaining the satellite relay transmission.

[0039] In this embodiment, the sensor data includes: visual sensor data, radar sensor data, laser sensor data, and ultrasonic sensor data.

[0040] Step S2: A hybrid control structure is adopted. The ground control station initially assigns tasks to each drone. When the drone group reaches a confrontation state, the drone cluster is divided into multiple drone sub-clusters according to the number of tasks. Each drone sub-cluster selects the main drone as the leader node, and the remaining slave drones are obedient nodes.

[0041] When the drone swarm is performing a confrontation mission, there are mainly centralized methods and distributed methods for the control structure and task allocation of the drone swarm. The centralized control mainly allocates and schedules tasks through a centralized controller, and the ground control station centrally schedules tasks. However, this control method has high computational complexity and is easily affected by single point failures. For the distributed method, tasks are flexibly allocated to individual drones through distributed strategies, which is suitable for dynamic environments, and the task allocation scheme is determined through negotiation between drones. In order to better balance the globality and dynamism, this embodiment adopts a hybrid control structure. First, tasks are allocated and called for all drones in the drone swarm. The ground control station performs preliminary task allocation, and the corresponding allocation information is stored as a central command in the information corresponding to each drone. At the same time, each drone obtains its own action task.

[0042] Furthermore, when the drone swarm performs confrontation tasks, it adopts a centralized scheduling control method according to the preset path and moves forward along the set route. During this period, the drone cluster automatically identifies obstacles in the route and avoids them. When the drone swarm reaches the confrontation state or reaches the predetermined area for performing tasks, the drone cluster is divided into multiple drone sub-clusters according to the number of tasks. Each drone sub-cluster selects the main drone as the leader node, and the remaining slave drones are obedient nodes.

[0043] Specifically, the leader node acts as the master control node and is the instructor of the corresponding task of the corresponding drone sub-cluster. Other nodes act as subordinate nodes and plan tasks according to the instructions issued by the corresponding leader node.

[0044] In this way, when the drone cluster performs task segmentation, the cluster is segmented according to the task situation. When the corresponding leader node is selected from the corresponding segmented cluster, its own communication topology is constructed in the corresponding segmented cluster to reduce the network nodes of the corresponding small topology and enhance the reliability of communication. At the same time, a communication topology is constructed between the corresponding leader nodes, and the task completion progress of each drone sub-cluster and the relevant corresponding obedient node information are synchronized in the communication topology. By dividing the corresponding drone cluster into multiple sub-task nodes and the main control node, the communication topology information of each of them is reduced, solving the problem of high pressure on the communication network within the drone cluster. In addition, in this way, the drone cluster can be expanded and reduced at any time, which increases the scalability of the drone cluster.

[0045] It should be further explained that step S2 also includes: each obedient node communicates with the leader node in the corresponding drone sub-cluster at a preset time threshold 1. When the leader node fails to detect the corresponding obedient node within the preset communication number threshold 2, the obedient node is deemed to be lost.

[0046] Specifically, in the task communication topology network within the drone cluster, the leader node will detect the heartbeat information of the corresponding obedient node, and each obedient node will communicate with the corresponding leader node every 0.02s. When the corresponding leader node still fails to detect the heartbeat information of the corresponding relevant obedient node after 50 heartbeats, the obedient node is considered lost. If the heartbeat is restored within 50 heartbeats, it is considered that the obedient node still exists. If the obedient node reapplies for connection after being judged to be lost, a separate task is calibrated for the obedient node at this time, and the obedient node performs the confrontation task separately according to its status. In this way, it can be ensured that all drones in the drone sub-cluster within the drone cluster can be fully utilized.

[0047] Step S3: The leader node of each drone sub-cluster scans and identifies the target, analyzes the target's motion trajectory based on the onboard equipment sensor data of each drone in the drone sub-cluster, and determines the reliability of the trajectory prediction of each drone in the drone sub-cluster for the target based on the detection area of ​​each drone in the drone sub-cluster, the cosine similarity between the predicted trajectory vector and the actual trajectory vector, and the spatial position distance.

[0048] Specifically, when the drone swarm reaches the confrontation state or arrives at the predetermined area for executing the mission, the onboard equipment sensors of each drone will detect surrounding targets in real time. When there are multiple targets, the leader node of the drone sub-cluster scans and identifies the target. At the same time, the leader node uploads the target location mark to the network. After obtaining it, other drone sub-clusters select other targets for confrontation.

[0049] In the process of analyzing the target's motion trajectory by the onboard equipment sensor data of each drone in the drone sub-cluster, when a drone sub-cluster locks on one of all the targets as the confrontation target, each drone in the drone sub-cluster detects and identifies the other party's three-dimensional spatial position, speed, maximum acceleration and other information based on multiple onboard equipment sensors, and predicts its subsequent motion trajectory based on the relevant neural network model. For example, the running speed and maximum acceleration of the relevant target are obtained based on the visual sensor data, and when the drones in the drone sub-cluster capture the target, the running trajectory of the target begins to be predicted. For another example, when the target is an attacking drone, the motion trajectory is predicted based on the angle and direction of the target and the ground captured by the drones in our drone sub-cluster through sensors (the angle and direction of the target body relative to the ground during flight, including pitch angle, roll angle and yaw angle, these attitude parameters directly affect the speed and motion trajectory of the target). Since there are multiple drones in the drone sub-cluster to detect the target, the relevant position information of the target can be fused according to the relevant observation angles, and then the trajectory of the target is predicted to obtain the difference between the predicted result and the actual motion result of the target in unit time.

[0050] It should be noted that when the drone group arrives at the predetermined area for performing the mission, it is considered to have reached a confrontation state. The target movement process scanned and identified by the leader node forms a continuous area. The front end of the target movement area for each drone in the corresponding drone sub-cluster is defined as the detection area, the area in the middle of the target movement area where confrontation with our drone may occur is defined as the early warning area, and the area at the back end of the target movement area where a collision with our drone may occur is defined as the collision area.

[0051] In the detection area, each drone in the drone sub-cluster detects and identifies the target information based on various airborne equipment sensors, such as the flight speed and corresponding activity range at each moment, and then performs Kalman filtering operations, and then predicts the target trajectory. At the same time, the target motion trajectory at the next moment is detected through the airborne equipment sensor. In this way, the error between the first-stage prediction result and the actual motion result of the target can be obtained. The error vector between the first-stage prediction result and the actual motion result of the target is regarded as the second-stage Kalman filter excitation noise, and the error between the second-stage prediction result and the actual motion result of the target is corrected. Repeat the above operation, after multiple stages of iterative correction of the error vector between the prediction result and the actual motion result of the target, to predict the subsequent motion trajectory of the target in the warning area, so that the drones that perform subsequent tasks can carry out collision confrontation with the target in the collision area.

[0052] By tracking the trajectory of the target in its first stage, the corresponding detection data and trajectory prediction difference data are used as Kalman filtering to iteratively predict the subsequent motion trajectory. Since the motion trajectory corresponding to the target will change with its own posture changes, the corresponding target confrontation drones in the drone sub-cluster are in different spatial positions, and there are certain differences in the target posture monitoring, and the specific situation of the actual corresponding target's motion line is affected by many factors, so the reliability of the corresponding prediction result has a great influence on the final confrontation success rate of the target. In this regard, when the target is in the early warning area, this embodiment proposes to judge the reliability of the trajectory prediction of each drone in the drone sub-cluster for the target based on the detection area of ​​each drone in the drone sub-cluster, the cosine similarity between the predicted trajectory vector and the actual trajectory vector, and the spatial position distance.

[0053] Specifically, the larger the detection area of ​​the target by the drones in the drone sub-cluster, the higher the relative accuracy of the prediction of the target trajectory change; the higher the cosine similarity between the predicted trajectory vector and the actual trajectory vector of the drones in the drone sub-cluster, the higher the relative accuracy of the prediction of the target trajectory change; the closer the spatial position distance of the drones in the drone sub-cluster to the target, the higher the relative accuracy of the prediction of the target trajectory change.

[0054] In this embodiment, the calculation formula for the reliability of trajectory prediction of each drone in the drone sub-cluster for the target can be expressed as:

[0055]

[0056] Among them, i represents the detected target, n represents the nth drone in the corresponding drone sub-cluster, m represents the mth moment, and KP i,n It represents the reliability of the target trajectory prediction of the nth UAV in the UAV sub-cluster, MT i,nIndicates the number of detection results when the target is running in the warning area, z i,n,m represents the predicted change vector of the nth UAV to its target at its mth moment, w i,n,m represents the actual detection vector of the nth UAV on its target at the mth moment, SK m,n represents the detection area of ​​the target by the nth UAV at the mth time, O n,m represents the corresponding spatial position of the nth UAV at its mth moment, K i,m represents the spatial position of the target at the mth moment, cos(z i,n,m ,w i,n,m ) represents the cosine similarity between the target detection trajectory vector of the nth UAV in the UAV sub-cluster and the target actual change trajectory vector at the mth moment, D(O n,m ,K i,m ) represents the spatial position distance between the nth UAV in the UAV sub-cluster and the target at the mth moment, and morn represents the normalization of the results.

[0057] In the above calculation formula for the reliability of trajectory prediction of each drone in the drone sub-cluster for the target, firstly, the predicted change vector z of the nth drone for its target at its mth moment is obtained. i,n,m and the actual detection vector w of the nth UAV to its target at its mth moment i,n,m The cosine similarity cos(z i,n,m ,w i,n,m ), which plays a role in the reliability KP of the target trajectory prediction of the nth UAV in the UAV sub-cluster i,n The most critical variable to evaluate can be used as the reliability KP of the target trajectory prediction i,n Calculated correction coefficient; Due to the detection area SK of the target by the nth UAV at the mth moment m,n Directly positively correlated with reliability KP i,n , the detection area SK of the target by the nth UAV at the mth moment m,n The larger the value, the higher the relative accuracy of the prediction of the target trajectory change; in addition, at the mth moment, the spatial position distance D(O n,m ,K i,m ), which is directly negatively correlated with the reliability KP i,n , the farther the spatial position distance of the nth UAV in the UAV sub-cluster from the target at the mth moment, the lower the relative accuracy of the prediction of the target trajectory change; the detection area SK of the nth UAV at the mth moment to the target m,n The spatial position distance D(O) between the nth UAV in the UAV sub-cluster and the target at the corresponding mth moment n,m ,K i,m) multiplied by the correction factor, that is, multiplied by the predicted change vector z of the nth UAV to its target at its mth moment i,n,m and the actual detection vector w of the nth UAV to its target at its mth moment i,n,m The cosine similarity cos(z i,n,m ,w i,n,m ), the result is the reliability KP of the trajectory prediction of the nth UAV to the target at a single moment (the mth moment) in the first stage i,n ,Then the reliability levels from the single moment in the first stage to the single moment in the last stage are averaged, and the results are normalized to obtain the reliability of the trajectory prediction of each drone in the drone sub-cluster for the target.

[0058] Step S4: Determine the success efficiency of each drone in the drone sub-cluster in confronting the target based on the number of obstacles in the predicted confrontation path of each drone in the drone sub-cluster, the predicted confrontation time, the distance between the predicted impact position and the target position, and the status of each drone.

[0059] Among them, the success efficiency of the confrontation with the target, taking the interception state of the target as an example, is affected by its interception time, interception trajectory, the distance between the interception point and the target area, and the state of the intercepted drone. Therefore, this embodiment considers the success efficiency of the confrontation with a single target on its predicted trajectory from the above-mentioned directions.

[0060] Considering the factors affecting the successful efficiency of each drone in the drone sub-cluster against the target, the fewer obstacles there are in the predicted confrontation path of each drone in the drone sub-cluster, and the closer the confrontation trajectory of the drones in the drone sub-cluster against the target is to a straight line, the smaller the error generated by the corresponding confrontation drone in its movement against the target, and the higher the successful efficiency of confrontation; the shorter the predicted confrontation time of each drone in the drone sub-cluster, the smaller the change in the speed of the attitude adjustment performed by the target during this time, and the lower the deviation of the corresponding trajectory of the target, so the higher the successful efficiency of confrontation; the closer the distance between the predicted impact position of each drone in the drone sub-cluster and the target position, the more accurate the reliability of the trajectory prediction of each drone in the drone sub-cluster for the target, resulting in a higher successful efficiency of confrontation; the better the state of each drone in the drone sub-cluster, the higher the stability of each drone in the drone sub-cluster when performing confrontation tasks, and the higher the successful efficiency of confrontation.

[0061] It should be noted that the farther the distance between the predicted impact position of each drone in the drone sub-cluster and the target position, the lower the threat level of each drone in the drone sub-cluster to the target. If the confrontation fails, a secondary confrontation can be carried out.

[0062] Furthermore, the status of each drone includes power status and communication stability. When the power of the drones in the drone sub-cluster is higher and the communication stability is better, the corresponding drones in the drone sub-cluster have a higher success efficiency in confronting the target.

[0063] In this embodiment, the calculation formula of the successful efficiency of each drone in the drone sub-cluster against the target can be expressed as:

[0064]

[0065] Among them, R i,n ZL represents the expected countermeasure efficiency of the nth UAV against its target UAV. n,i represents the number of obstacles in the expected confrontation path of the nth UAV, MB represents the spatial position of the target, and ZC i,n represents the spatial position of the expected confrontation point, Zt n,i represents the expected confrontation time of the corresponding drone, w n Indicates the state of the nth drone (the state is obtained by obtaining the power status and internal communication stability), D(MB,ZC i,n ) indicates obtaining the distance between the predicted impact position and the target position, and morn indicates normalizing the result.

[0066] In the above calculation formula of the success efficiency of each drone in the drone sub-cluster against the target, since the number of obstacles ZL in the expected confrontation path of the nth drone is n,i Negatively correlated with the success efficiency of confrontation, the corresponding number of obstacles ZL n,i The more, the lower the efficiency of the confrontation success; the corresponding drone’s estimated confrontation time Zt n,i Negatively correlated with the success efficiency of confrontation, corresponding to the estimated confrontation time Zt of the drone n,i The longer it is, the lower the efficiency of the confrontation success; the distance D(MB,ZC i,n ) is negatively correlated with the success efficiency of the confrontation, and predicts the distance D(MB,ZC) between the impact position and the target position. i,n ) is farther away, the lower the efficiency of the confrontation success; in addition, the state w corresponding to the nth drone n Positively correlated with the success efficiency of the confrontation, corresponding to the state w of the nth drone n The better, the higher the interception efficiency. In this way, through the above formula, the interception efficiency of the corresponding drone can be obtained from the perspectives of confrontation time, confrontation position, confrontation path and the stability of the confrontation drone itself.

[0067] Step S5: Select the first confrontation drone according to the reliability of trajectory prediction of each drone in the drone sub-cluster for the target and the success efficiency of confrontation with the target.

[0068] Specifically, based on the reliability ranking of the trajectory prediction of each drone in the drone sub-cluster for the target, combined with the ranking of the successful confrontation efficiency of each drone in the drone sub-cluster against the target, the drone with the highest reliability in successfully performing the confrontation task is obtained as the first confrontation drone, and the reliability of successfully performing the confrontation task of each drone in the drone sub-cluster is ranked from high to low.

[0069] When the drone sub-cluster arrives at the predetermined area for executing the mission, each drone in the drone sub-cluster sends the corresponding relevant detection information to its leader node, which processes it and selects the corresponding confrontation drones. Based on the above-obtained reliability of the trajectory prediction of the single drone for the target and the efficiency of the confrontation success against the target, the final reliability of the single drone in executing the confrontation mission is obtained.

[0070] In this embodiment, the calculation formula for the success reliability of each drone in the drone sub-cluster in executing the confrontation task against the target can be expressed as:

[0071] H i,n =R i,n +KP i,n

[0072] Among them, H i,n represents the reliability of the nth adversarial UAV in successfully executing the adversarial mission against target i, R i,n represents the success efficiency of the nth UAV against its target, KP i,n It indicates the reliability of the trajectory prediction of the nth UAV for its target.

[0073] In the above calculation formula for the reliability of each drone in the drone sub-cluster to execute the confrontation mission against the target, the reliability of the predicted confrontation success efficiency of the target and the trajectory prediction of the target are obtained, and the two are accumulated and summed to obtain the reliability of each drone in the drone sub-cluster to execute the confrontation mission against the target, H i,n The drone with the highest value has the highest reliability in successfully executing the confrontation task against the target. Before each drone in the drone sub-cluster enters the collision area, the drone with the highest reliability in successfully executing the confrontation task is obtained from the corresponding drone sub-cluster and used as the first confrontation drone to perform collision confrontation against the target.

[0074] Step S6: Use the first confrontation UAV to perform a confrontation mission on the target.

[0075] Specifically, confrontation tasks include attack, interception, etc.

[0076] Furthermore, after step S6, it also includes, when the first confrontation drone fails to perform the confrontation task, ranking the drones in the drone sub-cluster according to their reliability of successfully performing the confrontation task from high to low, and continuing to perform the confrontation task in sequence until the task is successfully performed.

[0077] It should be noted that when the leader node signal is lost, all subordinate nodes in the corresponding drone sub-cluster select a new leader node from within through election; the election basis includes the controller computing resource usage, communication stability, and airborne device sensor response speed of all subordinate nodes. When the controller computing resource usage of the subordinate node is low, the communication stability is high, and the airborne device sensor response speed is fast, the corresponding subordinate node is the optimal new leader node.

[0078] Specifically, according to the actual target tasks, the three parameters of the obedient node's controller computing resource usage, communication stability, and airborne equipment sensor response speed are considered with different weights in the election process. In this way, it can be ensured that each drone sub-cluster always has a leader node, ensuring that each drone sub-cluster can engage in the confrontation task against the target in an organized manner.

[0079] Through the above-mentioned collaborative confrontation control method based on multiple UAVs, by adopting a hybrid control structure, the ground control station preliminarily assigns tasks to each UAV. When the UAV group reaches the confrontation state, the UAV cluster is divided into multiple UAV sub-clusters according to the number of tasks. Each UAV sub-cluster selects a main UAV as a leader node, and the remaining slave UAVs are obedient nodes. This can solve the problem of high pressure on the communication network within the UAV cluster during the system confrontation of multiple UAVs; by analyzing the motion trajectory of the target according to the onboard equipment sensor data of each UAV in the UAV sub-cluster, judging the UAV according to the detection area of ​​each UAV in the UAV sub-cluster for the target, the cosine similarity between the predicted trajectory vector and the actual trajectory vector, and the spatial position distance. The reliability of trajectory prediction of each drone in the human-machine sub-cluster for the target; judging the success efficiency of each drone in the drone sub-cluster against the target according to the number of obstacles in the predicted interception path of each drone in the drone sub-cluster, the predicted interception time, the distance between the predicted impact position and the target position and the status of each drone; selecting the first confrontation drone according to the reliability of trajectory prediction of each drone in the drone sub-cluster for the target and the success efficiency of confrontation against the target can solve the problem that the most reliable drone cannot be quickly screened out during the coordinated confrontation of multiple drones; and further solves the problem of high cost of executing confrontation tasks caused by the existing technology that often adopts methods such as increasing the intelligence level of each drone in the drone cluster or adopting saturation confrontation that wins by quantity.

[0080] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0081] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A cooperative confrontation control method based on multiple UAVs, characterized in that: The method comprises: Step S1: for the drone group network, obtain the sensor data of the airborne equipment of each drone in real time; Step S2: A hybrid control structure is adopted. The ground control station initially assigns tasks to each drone. When the drone group reaches a confrontation state, the drone cluster is divided into multiple drone sub-clusters according to the number of tasks. Each drone sub-cluster selects a main drone as the leader node, and the remaining slave drones are obedient nodes. Step S3: The leader node of each drone sub-cluster scans and identifies the target, analyzes the target's motion trajectory based on the onboard equipment sensor data of each drone in the drone sub-cluster, and determines the reliability of the trajectory prediction of each drone in the drone sub-cluster for the target based on the detection area of ​​each drone in the drone sub-cluster, the cosine similarity between the predicted trajectory vector and the actual trajectory vector, and the spatial position distance; Step S4: judging the success efficiency of each drone in the drone sub-cluster in confronting the target according to the number of obstacles in the predicted confrontation path of each drone in the drone sub-cluster, the predicted confrontation time, the distance between the predicted impact position and the target position, and the status of each drone; Step S5: selecting a first confrontation drone according to the reliability of trajectory prediction of each drone in the drone sub-cluster for the target and the success efficiency of confrontation with the target; Step S6: Use the first confrontation UAV to perform a confrontation mission on the target.

2. The method for cooperative confrontation control based on multiple UAVs according to claim 1 is characterized in that: The sensor data includes: visual sensor data, radar sensor data, laser sensor data, and ultrasonic sensor data.

3. The method for cooperative confrontation control based on multiple UAVs according to claim 2 is characterized in that: The real-time acquisition of the onboard equipment sensor data of each UAV specifically includes: the ground control station acquires in real time the signal after the equipment sensor data is fused and processed and transmitted by satellite relay.

4. The method for cooperative confrontation control based on multiple UAVs according to claim 1 is characterized in that: The step S2 also includes: each obedient node communicates with the leader node in the corresponding drone sub-cluster at a preset time threshold 1, and when the leader node fails to detect the corresponding obedient node within a preset communication number threshold 2, the obedient node is deemed lost.

5. The method for cooperative confrontation control based on multiple UAVs according to claim 1 is characterized in that: The step S3 also includes: when the leader node of the drone sub-cluster scans and identifies the target, the leader node uploads the target position mark to the network, and other drone sub-clusters obtain it and select other targets to fight against.

6. The method for cooperative confrontation control based on multiple UAVs according to claim 1 is characterized in that: The step S3 also includes: when the detection area of ​​the drones in the drone sub-cluster for the target is larger, the cosine similarity between the predicted trajectory vector and the actual trajectory vector is higher, and the spatial position distance is closer, the reliability of the trajectory prediction of the corresponding drones in the drone sub-cluster for the target is higher.

7. The method for cooperative confrontation control based on multiple UAVs according to claim 1 is characterized in that: The step S4 further comprises: The fewer the number of obstacles in the predicted confrontation path of the drones in the drone sub-cluster against the target, the shorter the predicted confrontation time, and the closer the distance between the predicted impact position and the target position, the higher the success efficiency of the corresponding drones in the drone sub-cluster against the target; The status of each drone includes power status and communication stability. When the power of drones in the drone sub-cluster is higher and the communication stability is better, the corresponding drones in the drone sub-cluster have higher success efficiency in confronting the target.

8. The method for cooperative confrontation control based on multiple UAVs according to claim 1 is characterized in that: The step S5 specifically includes: According to the reliability ranking of trajectory prediction of each drone in the drone sub-cluster for the target, combined with the ranking of the successful efficiency of confrontation of each drone in the drone sub-cluster against the target, the drone with the highest reliability of successfully executing the confrontation task is obtained as the first confrontation drone, and the reliability of successfully executing the confrontation task of each drone in the drone sub-cluster is ranked from high to low.

9. The method for cooperative confrontation control based on multiple UAVs according to claim 8 is characterized in that: The step S6 also includes, when the first confrontation UAV fails to perform the confrontation task, ranking the UAVs in the UAV sub-cluster according to their reliability of successfully performing the confrontation task from high to low, and continuing to perform the confrontation task in sequence until the task is successfully performed.

10. The method for cooperative confrontation control based on multiple UAVs according to claim 4 or 5, characterized in that: When the leader node signal is lost, all subordinate nodes in the corresponding drone sub-cluster select a new leader node from within through election; the election basis includes the controller computing resource usage, communication stability, and airborne equipment sensor response speed of all subordinate nodes.

Images