A self-organizing cluster cooperative control method for non-cooperative multi-target encirclement
By combining a target motion estimator, an encirclement center motion observer, and an encirclement radius estimator with distance tracking and phase uniform distribution control methods, the self-organized encirclement problem of multiple non-cooperative targets in unmanned swarms was solved, and reliable encirclement of UAVs in dynamic scenarios was achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIHANG UNIV
- Filing Date
- 2026-03-06
- Publication Date
- 2026-06-02
Smart Images
Figure CN122131817A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of unmanned aerial vehicle (UAV) control technology, and more specifically, to a self-organizing swarm cooperative control method for non-cooperative multi-target encirclement. Background Technology
[0002] Unmanned swarm collaborative systems consist of multiple unmanned aerial vehicles (UAVs) that cooperate through a specific mechanism to achieve collaborative missions. Each UAV in a swarm possesses autonomous operational capabilities, and through this collaborative mechanism, they can maximize combat effectiveness through information exchange and mission coordination. Therefore, they can bring non-linear performance gains and disruptive application breakthroughs to typical missions such as reconnaissance, detection, search and rescue, and mapping, thus becoming an indispensable force in future military and civilian fields.
[0003] The target encirclement problem in unmanned swarm cooperative systems is a key issue in research on unmanned swarm cooperative systems. It aims to encircle targets within the convex hull of the unmanned swarm, and can be applied to scenarios such as target protection, interception, and cooperative attacks. The most relevant technical field to this invention is unmanned swarm cooperative multi-target encirclement, which requires unmanned swarms to utilize their advantages in numbers and formation shape to maintain a good encirclement effect on multiple moving targets. Current research methods mainly include encirclement methods based on information-complete UAVs and encirclement methods without information-complete UAVs.
[0004] Encirclement methods based on information-complete UAVs assume the presence of information-complete UAVs (UAVs capable of measuring the positions of all targets) within the UAV swarm. Existing research often addresses multi-target encirclement problems through formation tracking methods and trajectory tracking strategies. However, the information-complete UAVs necessary for encirclement methods based on information-complete UAVs generally do not exist in practical applications and are difficult to implement in engineering practice.
[0005] Another type of approach is the encirclement method for scenarios where information-complete UAVs are absent. This type of method aims to achieve multi-target encirclement in scenarios without information-complete UAVs by designing an encirclement center observer and a cooperative encirclement controller. On one hand, inspired by average consistency theory, existing research has designed an encirclement center observer, enabling the UAV swarm to observe the geometric center of all targets as the encirclement center. On the other hand, existing research uses formation tracking methods to design cooperative encirclement controllers, forming a formation with the encirclement center as the formation center and a fixed encirclement radius to achieve multi-target encirclement. Encirclement methods for scenarios without information-complete UAVs often require global information such as the "number of UAVs and targets" or the "number of UAVs capable of measuring specific target locations," which is difficult to obtain in engineering practice. Furthermore, formation control methods have poor robustness to unexpected UAV exits and mid-course rejoins.
[0006] Existing encirclement methods based on information-complete UAVs and those without information-complete UAVs have addressed the multi-target encirclement problem, but a significant gap remains in scenarios involving multiple non-cooperative target groups. Multiple non-cooperative target groups are common in real-world applications, characterized by two core features: unmeasurable internal states and unacquirable dynamic target radii. However, these practical characteristics have not been addressed in existing research. Furthermore, current research often relies on global information (such as the number of UAVs and targets, and the number of UAVs capable of measuring specific target positions) to solve the multi-target encirclement problem. However, in real-world scenarios, individual UAVs often struggle to acquire this type of global information. Simultaneously, during actual encirclement operations, UAVs may join / leave the swarm due to malfunctions or replacements, and new non-cooperative targets may join the target group as needed for the mission. This can lead to unexpected changes in global information at unknown times, resulting in encirclement failure. Therefore, a self-organizing swarm cooperative control method for non-cooperative multi-target encirclement is urgently needed to ensure reliable multi-target encirclement even when UAVs join / leave or new targets are added. Summary of the Invention
[0007] To address the aforementioned problems in existing technologies and solve the problem of swarm encirclement of non-cooperative multi-target targets, this invention designs a self-organizing swarm cooperative control method for non-cooperative multi-target encirclement. First, based on a target motion estimator and an encirclement center motion observer, a distributed estimation of the encirclement center by the unmanned swarm is achieved. Then, based on an encirclement radius estimator, the UAV estimates the desired encirclement radius. Finally, based on a range tracking and phase-distributed control method, the acceleration control signal of the UAV is calculated, achieving cooperative encirclement of multiple non-cooperative targets.
[0008] The complete technical solution of this invention includes: S1 Computing Unmanned Aerial Vehicle exist Neighbor set at time
[0009] drones exist Neighbor set at time This is fundamental to collaborative target encirclement control in unmanned swarms, enabling information exchange between drones. This step calculates the distance between drones... exist Neighbor set at time .
[0010] Consider a A swarm of drones needs to be surrounded. A non-cooperative target. The group of drones is denoted as... The set of targets is denoted as Select For each drone's sensing range, calculate the drone's... With drones Between Relative distance at time ,in drones With drones exist The position vector at that moment. Then, the drone... exist Neighbor set at time The calculation is as follows:
[0011] S2 Computing Drone exist The observations of the position, velocity, and acceleration of the encircling center at all times.
[0012] This step is based on distributed information interaction between drones, calculating the drone... exist The observations of the position, velocity, and acceleration of the encircling center at all times. Record of drones exist The set of targets whose location information can be measured at any time is For any target For drones Design the following target motion estimator to calculate the motion of a UAV. For the goal exist Location at any moment ,speed With acceleration The estimated values are denoted as follows: and :
[0013] in For the drone algorithm's operating cycle, drones For the goal Location ,speed With acceleration The estimation error, For positive design parameters, ,in for One of the Lipschitz constants. For vectors Sign function, for vector Define vector of The symbolic function is ,in For a scalar, the sign function is... Its values are as follows:
[0014] For vectors , .
[0015] Then, based on the estimates generated by the target motion estimator For drones Design the following surrounding center motion observer to calculate the drone's motion. For the surrounding center position ,speed With acceleration The estimated value is denoted as :
[0016] in In order to be in Time Drone The internal state of the surrounding center motion observer. They are respectively in Time Drone The estimated average position, velocity, and acceleration of the multiple targets, This represents the gain of the motion observer.
[0017] S3 Computing Drone exist The estimated value of the enclosing radius at time 1
[0018] This step is based on distributed information interaction between drones, calculating the drone... exist The estimated value of the enclosing radius at time 1 This enables distributed estimation of the dynamic encirclement radius, ensuring that the encirclement radius changes dynamically with the radius occupied by the target.
[0019] remember and drones exist Time for global quantity and The distributed estimate, where For multi-target radius, The desired enclosing radius is defined as follows:
[0020] in This represents the minimum number of nodes in the cluster. To ensure the target is surrounded, an auxiliary distance is provided. Then, for drones... Design the following bounding radius estimator to calculate the bounding radius of a UAV. For the desired bounding radius The estimated value :
[0021] in For drones exist A generalized set of neighbors at any given moment. For drones only The set, For drones Local radius of detectable targets across multiple targets. For drones exist The set of generalized neighbors at time 1000 contains all the elements with the largest value. A set of drones with values, The gain of the enclosing radius estimator, A function with a positive sign for a scalar, for a scalar Its values are as follows:
[0022] S4 Computing Drone exist Acceleration control commands at any time Update the drone's pose status This step is based on distributed information interaction between drones, calculating the drone... exist Acceleration control commands at any time This enables all drones to surround the center at equal distances and angles.
[0023] First, calculate the drone exist Linear acceleration between the moment and the center of gravity With linear velocity as follows:
[0024] in To control the gain, For drones exist The distance between the moment and the center of the encirclement. For drones exist The linear velocity between the moment and the center of the surrounding area.
[0025] Next step, computing drones exist angular acceleration between the moment and the center of the enclosure With angular velocity Therefore, the first step is to calculate the drone. exist The previous neighbor number of the moment Numbering of next neighbor as follows:
[0026] in In order to be in Time Drone With drones The relative angle between them is calculated as follows:
[0027] in , They are respectively in Time Drone drones The phase angle with the difference vector surrounding the center is defined as follows:
[0028] Then, design drones exist angular acceleration between time and the center of the enclosure With angular velocity as follows:
[0029] in To control the gain.
[0030] Then, calculate the drone exist Acceleration control commands at any time as follows:
[0031] in .
[0032] Finally, update the drone. exist Location at any moment ,speed and drones exist Distance between time and the center of the enclosure With phase angle as follows:
[0033] The beneficial effects of this invention are as follows: Unlike existing methods that rely on global information to solve multi-target encirclement problems and cannot address the two core characteristics of multiple non-cooperative targets—unpredictable internal states and unobtainable dynamic target radii—this invention proposes a self-organizing swarm cooperative control method for non-cooperative multi-target encirclement. This method can solve the multi-target encirclement problem of multiple non-cooperative target groups without requiring global information. It supports the entry / exit of UAVs and the addition of new non-cooperative targets during the encirclement process, and the remaining UAVs can maintain reliable encirclement without adjusting parameters. In summary, this invention addresses the limitations of existing cooperative methods in dealing with the characteristics of multiple non-cooperative targets and the dynamic scenarios involving UAV entry / exit and the addition of new non-cooperative targets. The proposed self-organizing encirclement method has the potential to be applied to achieve non-cooperative target encirclement in both dynamic and static scenarios. Numerical simulations have tested the effectiveness of the encirclement method in both dynamic and static non-cooperative target encirclement scenarios. Attached Figure Description
[0034] Figure 1 The center position of the enclosure in a static scene Its estimated value The simulation results.
[0035] Figure 2 The center position of the enclosure in a static scene Its estimated value The simulation results.
[0036] Figure 3 The center position of the enclosure in a static scene Its estimated value The simulation results.
[0037] Figure 4 Desired radius in static scene Its estimated value The simulation results.
[0038] Figure 5 The simulation results show the motion trajectories of an unmanned swarm and multiple non-cooperative targets in a static scene.
[0039] Figure 6 Desired radius in static scene Distance between the drone and the center of the encirclement The simulation results.
[0040] Figure 7 Phase angle between the drone and the surrounding center in a static scene The simulation results.
[0041] Figure 8 Enclosing center position in dynamic scenes Its estimated value The simulation results.
[0042] Figure 9 Enclosing center position in dynamic scenes Its estimated value The simulation results.
[0043] Figure 10 Enclosing center position in dynamic scenes Its estimated value The simulation results.
[0044] Figure 11 Desired radius in dynamic scenarios Its estimated value The simulation results.
[0045] Figure 12 The simulation results show the motion trajectories of unmanned swarms and multiple non-cooperative targets in a dynamic scenario.
[0046] Figure 13 Desired radius in dynamic scenarios Distance between the drone and the center of the encirclement The simulation results.
[0047] Figure 14 Phase angle between the drone and the surrounding center in a dynamic scene The simulation results. Detailed Implementation
[0048] The present invention will now be described in detail with reference to embodiments and accompanying drawings. However, it should be understood that the embodiments and drawings are for illustrative purposes only and do not constitute any limitation on the scope of protection of the present invention. All reasonable modifications and combinations included within the inventive spirit of the present invention fall within the scope of protection of the present invention.
[0049] To facilitate understanding of the proposed cluster multi-target cooperative encirclement control method, the technical solution is illustrated below through a set of specific implementation cases. Simulation results are used to verify the effectiveness and self-organizing characteristics of the self-organizing cluster cooperative control method for non-cooperative multi-target encirclement in two scenarios: a static scenario where there is no drone withdrawal, access, or new target appearance (hereinafter referred to as a static scenario) and a dynamic scenario where there is drone withdrawal, access, and new target appearance (hereinafter referred to as a dynamic scenario). The specific implementation steps are as follows: 1. Motion modeling of unmanned swarms and non-cooperative targets: Suppose there are unmanned aerial vehicles in an unmanned swarm. The kinematic model is as follows:
[0050] in drones Position and velocity, and For drones The acceleration and control input.
[0051] Let the non-cooperative multi-objective goal be... The kinematic model is as follows:
[0052] in The target Position, velocity, and acceleration.
[0053] 2. Brief description of the implementation case scenario This implementation example performs simulations for both static and dynamic scenarios to verify the effectiveness of the technical solution proposed in this invention under both conditions. The scenarios are summarized below: In a static scenario, there is no drone withdrawal, joining, or new target appearance. Consider a drone swarm consisting of 6 drones, which needs to surround 2 non-cooperative targets. The acceleration settings for the two non-cooperative targets are as follows:
[0054] The unmanned swarm's detection status towards the target is as follows:
[0055] The initial states of the unmanned swarm and the non-cooperative target are as follows:
[0056] In dynamic scenarios, drones may leave, join, or new targets may appear. Consider a drone swarm initially consisting of 5 drones, which initially need to surround 2 non-cooperative targets. The acceleration of the two non-cooperative targets and the detection state of the 5 drones towards the targets, along with their initial states, remain consistent with static scenarios. Based on this, consider the following scenarios of drones leaving, joining, or new targets appearing: 1) Drones joining the unmanned swarm: Consider drone number 6, with the initial state as follows: It initially has no neighbors and remains stationary.
[0057] 2) Drones leaving the swarm: Consider drones numbered 3 and 4 experiencing a complete actuator failure at an unknown time, described as follows:
[0058] 3) New targets emerging: Consider the new target numbered 3 in When this occurs, the detection status of the unmanned swarm is as follows:
[0059] Its in The initial conditions at that time are: , Its acceleration is set as follows: . 3. Simulation parameter settings During the simulation, the same simulation parameters are used for both static and dynamic scenes, as follows:
[0060] 4. Based on the technical solution in the invention, design a program using the MATLAB platform and perform simulation.
[0061] 5. Output and analyze simulation results Simulation results in static scenarios during the simulation process are as follows: Figure 1-7 As shown.
[0062] Figure 1 –3 indicates that in static scenarios, the proposed encircling center observer can be implemented for each UAV. Observations of the location, velocity, and acceleration surrounding the center. , , Tracking the corresponding truth value , , .
[0063] Figure 4 –7 indicates that, in static scenarios, the proposed bounding radius estimator and acceleration control input can achieve bounding for multiple non-cooperative targets. Specifically, Figure 4 Displays the radius observation values of each UAV. Converging to the desired dynamic bounding radius This verifies that all drones have achieved dynamic enclosing radius control. The finite-time estimation. Figure 5 depicts the trajectory of the UAV with multiple non-cooperative targets, showing that when The unmanned swarm has achieved encirclement of all targets, and the encirclement radius dynamically changes with the radius of the multiple targets. Figure 6 Displays the distance between each drone and the center of the encirclement. Converging to the desired radius , Figure 7 This indicates the phase angle between all drones and the surrounding center. To achieve uniform distribution.
[0064] Simulation results in dynamic scenarios are as follows Figure 8-14 As shown in the figure, The corresponding signal indicates an active drone or a non-cooperative target that has appeared. This indicates that the drone is not inactive.
[0065] Figure 8 –10 validates that in dynamic scenarios involving UAV joining / leaving and the addition of new non-cooperative targets, all active UAVs achieve observations of the perimeter center's position, velocity, and acceleration through the proposed perimeter center observer. , , Tracking the truth , , And there is no need to adjust any parameters for activating the drone.
[0066] Figure 11 –14 indicates that in dynamic scenarios, all activated UAVs achieve cooperative encirclement of multiple non-cooperative targets through the proposed encirclement radius estimator and acceleration control inputs. Specifically, Figure 11 It was verified that all activated drones achieved coverage of the encirclement radius. Finite-time estimation. Figure 12 The display shows that when After that, all detected targets All can be surrounded by activated drones, and whenever a drone joins or leaves, all activated drones will self-organize into a non-preset formation to maintain the encirclement of the target. Figure 13 Displays the distance between each active drone and the center of the encirclement. Converging to the desired radius , Figure 14 The phase angle between all activated drones and the surrounding center was verified. To achieve uniform distribution.
[0067] The simulation results above verify that the self-organizing cluster cooperative control method for non-cooperative multi-objective encirclement proposed in this invention can be applied to both static and dynamic scenarios involving multi-non-cooperative object encirclement.
[0068] The foregoing has only described preferred embodiments of the present invention in detail and is not intended to limit the invention. Those skilled in the art will readily conceive of other embodiments of this disclosure upon considering the disclosure in the specification and embodiments. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.
Claims
1. A self-organizing cluster cooperative control method for non-cooperative multi-objective encirclement, characterized in that, Includes the following steps: S1: Computational Drone exist Neighbor set at time ; S2: Computational Drone exist The observations of the position, velocity, and acceleration of the encircling center at all times. ; S3: Computational Drone exist The estimated value of the enclosing radius at time 1 ; S4: Computational Drone exist Acceleration control commands at any time Update the drone's pose status.
2. The self-organizing cluster cooperative control method for non-cooperative multi-objective encirclement as described in claim 1, characterized in that, The self-organizing cluster is composed of A swarm of unmanned aerial vehicles (UAVs), wherein the non-cooperative multi-target group includes targets that require encirclement by a self-organizing swarm. Each non-cooperative target, drone, and non-cooperative multi-target group constitutes its own set.
3. The self-organizing cluster cooperative control method for non-cooperative multi-objective encirclement as described in claim 1, characterized in that, In step S1, the drone is calculated. In contrast to other drones The relative distance at any given moment, and based on the drone The perception range of the drone exist The collection of neighbors at any given moment.
4. The self-organizing cluster cooperative control method for non-cooperative multi-objective encirclement as described in claim 1, characterized in that, In step S2, the drone is first established. exist A set of targets whose location information can be measured at any time is defined, and a target motion estimator is designed for any target in the set to calculate the motion of the UAV. For any target in The position estimate, velocity estimate, and acceleration estimate at each moment are then used, and based on these estimates, the UAV... The surrounding center motion observer is used to calculate the drone's motion. For the estimated position surrounding the center, the estimated velocity and the estimated acceleration.
5. A self-organizing cluster cooperative control method for non-cooperative multi-objective encirclement as described in claim 4, characterized in that, In step S3, based on the drone exist For multiple targets and desired encirclement radius at any time, and for UAVs Design a bounding radius estimator to calculate the bounding radius of a drone. For the estimated value of the desired enclosing radius.
6. The self-organizing cluster cooperative control method for non-cooperative multi-objective encirclement as described in claim 1, characterized in that, In step S4, the drone is first calculated. exist The linear acceleration and linear velocity between the moment and the center of the enclosure are then calculated for the UAV. exist Angular acceleration and angular velocity between the moment and the center of the enclosure.
7. A self-organizing cluster cooperative control method for non-cooperative multi-objective encirclement as described in claim 6, characterized in that, Based on drones exist Calculate the linear acceleration, linear velocity, angular acceleration, and angular velocity of the UAV between the time point and the center of gravity. exist Acceleration control commands at any given time.
8. A self-organizing cluster cooperative control method for non-cooperative multi-objective encirclement as described in claim 6, characterized in that, Based on drones exist Real-time acceleration control commands for human-machine interface exist Position, speed and drones at any time exist The distance and phase angle between the time and the surrounding center are updated.