Heterogeneous uav cluster self-organizing operation control method, system and device

Abstract

The application relates to a heterogeneous unmanned aerial vehicle cluster self-organizing operation control method, system and device, the method comprising: based on a Boids model, uniformly deploying task loads in a task space in a relative dispersion manner in an initial task stage; the task loads comprising each unmanned aerial vehicle in the heterogeneous unmanned aerial vehicle cluster; acquiring input parameters of a flight path planning algorithm and task demand information, and using a lightweight optimal control model to plan a flight path for each unmanned aerial vehicle in a demand response stage; the input parameters comprising initial state constraints and terminal state constraints of the unmanned aerial vehicle; controlling each unmanned aerial vehicle to enter a surrounding waiting stage according to the planned flight path; when all unmanned aerial vehicle members of a temporary task alliance enter the surrounding waiting stage, triggering the temporary task alliance to initiate a synchronous operation on a task target; after the synchronous operation is completed, the temporary task alliance is dissolved, and each unmanned aerial vehicle switches to the initial task stage and waits for demand response of a new task target. The task efficiency of the cluster system is significantly improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of unmanned aerial vehicle cluster control, and relates to a heterogeneous unmanned aerial vehicle cluster self-organizing operation control method, system and device. BACKGROUND

[0002] The unmanned aerial vehicle cluster network in the working environment is driven by the self-organizing behavior of each node to produce dynamic evolution effect. From the perspective of mapping from the cyber space to the physical space, the process involves nonlinear and emergent behaviors generated by the interaction between the task unit group and the environment. Therefore, from the perspective of task entity participating in the task process, the unmanned aerial vehicle cluster can be classified as a complex adaptive system (CAS). The traditional complex system modeling method based on differential equations, such as the Lanchester equation, often stays at the macro level for information feedback between task units, lacking the description of dynamic properties at the micro level of task entities. For example, the simulation method based on discrete events performs coarse-grained decomposition from top to bottom for the task process, which is difficult to fully consider the heterogeneous characteristics and action details of task units. Under the condition of dynamic confrontation, the situation of the working environment changes sharply, and new target tasks may be injected in a "pop-up" manner and the nodes of task units may fail or even the cascade failure of the task system at any time, which will affect the effective performance of the system task. However, the traditional unmanned aerial vehicle cluster technology has not yet solved the technical problem of low task performance of the unmanned aerial vehicle cluster system. SUMMARY

[0003] In view of the problems in the above-mentioned traditional method, the application provides a heterogeneous unmanned aerial vehicle cluster self-organizing operation control method, a heterogeneous unmanned aerial vehicle cluster self-organizing operation control system and a computer device, which can significantly improve the task performance of the unmanned aerial vehicle cluster system.

[0004] To achieve the above-mentioned purpose, the embodiments of the application adopt the following technical solutions:

[0005] On the one hand, a heterogeneous unmanned aerial vehicle cluster self-organizing operation control method is provided, comprising the steps of:

[0006] Based on the Boids model, the task load is relatively dispersed and uniformly deployed in the task space at the initial stage of the task; the task load includes each unmanned aerial vehicle in the heterogeneous unmanned aerial vehicle cluster;

[0007] The input parameters of the path planning algorithm and the task demand information are obtained, and a lightweight optimal control model is used to plan the path of each unmanned aerial vehicle in the demand response stage; the input parameters include the initial state constraint and the end state constraint of the unmanned aerial vehicle;

[0008] The flight path of each unmanned aerial vehicle is controlled to enter the encirclement waiting stage according to the planned flight path;

[0009] When all the unmanned aerial vehicle members of the temporary task alliance enter the encirclement waiting phase, the temporary task alliance is triggered to launch a synchronous operation on the task target;

[0010] After the synchronous operation is completed, the temporary task alliance is dissolved, and each unmanned aerial vehicle switches to the task initial phase to wait for the demand response of a new task target.

[0011] In another aspect, a heterogeneous unmanned aerial vehicle cluster self-organizing operation control system is also provided, comprising:

[0012] A task initial module is configured to uniformly deploy a task load in a task space in a task initial phase based on a Boids model; the task load includes each unmanned aerial vehicle in the heterogeneous unmanned aerial vehicle cluster;

[0013] A task response module is configured to obtain input parameters of a path planning algorithm and task demand information, and to use a lightweight optimal control model to plan a flight path for each unmanned aerial vehicle in a demand response phase; the input parameters include initial state constraints and end state constraints of the unmanned aerial vehicle;

[0014] An encirclement waiting module is configured to control each unmanned aerial vehicle to enter an encirclement waiting phase according to a planned flight path;

[0015] An operation execution module is configured to trigger the temporary task alliance to launch a synchronous operation on the task target when all the unmanned aerial vehicle members of the temporary task alliance enter the encirclement waiting phase;

[0016] A state switching module is configured to dissolve the temporary task alliance after the synchronous operation is completed, and to switch each unmanned aerial vehicle to the task initial phase to wait for the demand response of a new task target.

[0017] In yet another aspect, a computer device is also provided, comprising a memory and a processor; the memory stores a computer program; and the processor implements the steps of the above-mentioned heterogeneous unmanned aerial vehicle cluster self-organizing operation control method when executing the computer program.

[0018] In still another aspect, a computer readable storage medium is also provided, which stores a computer program; and the computer program implements the steps of the above-mentioned heterogeneous unmanned aerial vehicle cluster self-organizing operation control method when executed by a processor.

[0019] One of the above technical solutions has the following advantages and beneficial effects:

[0020] The isomerous unmanned aerial vehicle cluster self-organizing operation control method, system and device, by means of the Boids model, the task load is relatively dispersed and uniformly deployed into the task space in the initial stage of the task, then the optimal control model is used for the path planning of the unmanned aerial vehicles in the demand response stage, the unmanned aerial vehicles enter the encirclement waiting stage according to the planned flight path, when all the unmanned aerial vehicles of the temporary task alliance enter the stage, the temporary task alliance is triggered to start the synchronous operation on the task target, and the temporary task alliance is dissolved after the operation is completed, so that the unmanned aerial vehicles switch to the initial stage of the task and wait for the demand response of the new task target. Thus, based on the information interaction network constructed in the unmanned aerial vehicle cluster, distributed sensing and information sharing of the operation situation are realized, on the basis of which, loose temporary task alliances are formed for multiple task targets, the task functions dispersed on multiple platforms are aggregated into an adaptive dynamic execution network, and finally the task is completed according to the preset cooperative task operation mode, and the task efficiency of the unmanned aerial vehicle cluster system is significantly improved. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0022] Figure 1 A flowchart of the isomerous unmanned aerial vehicle cluster self-organizing operation control method in an embodiment;

[0023] Figure 2 A schematic diagram of the motion state and transfer process of the unmanned aerial vehicle in an embodiment;

[0024] Figure 3 A schematic diagram of the planning process of the flight path of the unmanned aerial vehicle in an embodiment;

[0025] Figure 4 A flowchart of the isomerous unmanned aerial vehicle cluster self-organizing operation control method in another embodiment;

[0026] Figure 5 A flowchart of the isomerous unmanned aerial vehicle cluster self-organizing operation control method in another embodiment;

[0027] Figure 6 A schematic diagram of the module structure of the isomerous unmanned aerial vehicle cluster self-organizing operation control system in an embodiment. DETAILED DESCRIPTION

[0028] In order to make the purposes, technical solutions and advantages of the present application clearer, further detailed description will be made to the present application in combination with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not intended to limit the present application.

[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the description herein is for describing particular embodiments only and is not intended to be limiting of the present application.

[0030] It should be noted that a reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily mutually exclusive of one another.

[0031] It will be understood by those skilled in the art that the embodiments described herein can be combined with other embodiments. The term "and / or" used in the specification and appended claims of the present application means one or more of the associated listed items as well as all possible combinations of the items and includes the combinations.

[0032] In one embodiment, as shown in Figure 1 and Figure 2 A heterogeneous UAV cluster self-organizing operation control method is provided, including steps S11 to S15:

[0033] S11, based on the Boids model, the task load is uniformly deployed in the task space in the initial stage of the task; the task load includes each UAV in the heterogeneous UAV cluster.

[0034] It can be understood that in the working environment, the heterogeneous UAV cluster network can rely on the self-organizing behavior of each node to drive the dynamic evolution effect. From the perspective of mapping from the cyber space to the physical space, the process involves nonlinear and emergent behaviors generated by the interaction between the task unit group and the environment.

[0035] To this end, from the perspective of task entity participating in the task process, the UAV cluster can be classified as a complex adaptive system (CAS). Traditional complex system modeling methods based on differential equations, such as the Lanchester equation, often stay at the macro level for information feedback between task units, lacking the description of dynamic properties at the micro level of task entities. For example, the simulation method based on discrete events, which performs a top-down coarse-grained decomposition of the task process, has difficulty in fully considering the heterogeneous characteristics and action details of the task units.

[0036] And the agent-based modeling method (ABM) has a natural advantage in describing complex systems with heterogeneity, nonlinearity, emergence and large-scale self-organization and self-adaptation. In summary, the ABM (i.e. agent-based modeling) method can set the functional attributes of individual task units, simplify the description of the interaction process between task units by designing the "if-then" action set and distributed decision-making process, and effectively combine the micro actions of agents and the emergence of macro cluster intelligence.

[0037] Currently, the agent-based modeling method is widely used in the field of multi-agent task planning, and the core modules of its model establishment mainly include: situation awareness module, information processing module, decision-making module and behavior driving module. The present application mainly models the behavior driving module directly related to the self-organizing operation of the unmanned aerial vehicle cluster, which can be divided into state driving mode and trigger condition, i.e. the precondition and subsequent action of "if-then" rule. For ease of description, the ABM modeling process is introduced into the real task scene. For example, the task background can be set: manned platforms as high-value, multi-functional monolithic task units carry unmanned bee swarm dispensers, arrive at the safe task area outside the opponent's defense zone under the support of the preliminary sketch reconnaissance information, and release a large-scale low-value, modular functional Mosaic task platform to form a heterogeneous unmanned aerial vehicle cluster to approach the opponent's position. The task target of the heterogeneous unmanned aerial vehicle cluster is to build an information interaction network within the cluster, realize distributed situation awareness and information sharing, form a loose temporary task alliance for multiple task targets, aggregate the task functions dispersed on multiple platforms into a dynamic adaptive execution network, and finally complete the task according to the preset cooperative attack task style.

[0038] In the initial stage of the task, the heterogeneous UAV cluster has not obtained accurate target indication, and each UAV (hereinafter also referred to as a node or a task unit) does not need to plan a flight path but adopts a self-driven way to perform a serpentine maneuver, avoids collision between UAVs by perceiving the positions of surrounding friendly UAVs, and realizes relative dispersion and uniform deployment of the task load in the task space. Essentially, this is a bottom-up swarm intelligence emergence process, which realizes complex interaction and cooperation at the group level by setting relatively simple behavior rules for individuals. Typical research results of this kind of method include the Boids model that has been proposed, which abstracts the self-organizing process of biological clusters into three basic behavior rules: separation, cohesion, and alignment (Separation, Cohesion, Alignment), and finally realizes the self-organizing cluster behavior of multi-agent bird flocks. In this paper, the initial stage task resource is uniformly distributed in the task space based on the model, and each UAV in the heterogeneous UAV cluster includes UAVs that undertake different task functions and carry different operation resources.

[0039] In some embodiments, the task execution state of the UAV in the heterogeneous UAV cluster includes uncommitted target traction, committed target traction, and fixed target selection, and the motion state of the UAV includes serpentine maneuver, flight path flight, waiting in a circle, synchronous attack, and node failure.

[0040] It can be understood that, in the embodiment, under the action of the behavior driving module, the various motion states and transition rules of the UAV are as shown in the table. Figure 2 The task execution state es of the UAV in the task execution process is divided into three types, namely, uncommitted, committed, and sticky. The motion state ms is divided into five types, namely, serpentine maneuver, flight path flight, waiting in a circle, synchronous attack, and node failure.

[0041] The uncommitted target traction indicates that the UAV currently does not respond to the demand of any target and randomly walks in the task space in the serpentine maneuver state. The committed target traction indicates that the UAV currently responds to the task demand of a certain target and is in the flight path flight state. In the process of going to the task execution position, the UAV can find a target with higher benefits and switch to the flight path to the task execution position of the target. In addition, the UAV can find that there is a UAV that has already gone to execute the task and is better than itself, and then it is considered that it has failed in the "bidding" for the task and switches to the serpentine maneuver state.

[0042] The fixed target selection means that the UAV is "sticky" to the currently selected target task and does not switch tasks until it reaches a specific task execution position around the target, enters a surrounding waiting state, and waits to form a temporary task alliance with the remaining heterogeneous UAVs for the target. When the alliance is formed, the heterogeneous UAVs in the alliance will launch a synchronous attack on the target, and after the attack is completed, the alliance is dissolved, and each UAV returns to the snake-like maneuvering state. In some cases, the UAV in the surrounding waiting state may be stuck for a long time without new UAVs joining the alliance, thus falling into a "stuck" state. At this time, the UAV can actively break this state (such as setting a waiting countdown) and re-enter the snake-like maneuvering state to respond to the needs of the remaining targets. In addition, under dynamic confrontation conditions, the UAV may be attacked by the opponent at any time during the entire task execution process and enter a node failure state (dead).

[0043] S12, obtaining input parameters of a flight path planning algorithm and task demand information, and using a lightweight optimal control model to perform flight path planning for each UAV in the demand response stage; the input parameters include initial state constraints and terminal state constraints of the UAV, the initial state constraints being the current position and speed of the UAV, and the terminal state constraints being the best attack position and speed direction into the best attack position calculated from the task demand information.

[0044] It can be understood that the flight path planning algorithm refers to an existing UAV flight path planning algorithm in the art, and the input parameters can be obtained by pre-computing, inputting or setting. The task demand information refers to parameter information such as the type, number of UAVs and the type and number of resources carried by the UAVs required to complete the operation on the specified target task, which can be set according to actual task needs. The lightweight optimal control model is a pre-established UAV control quantity calculation model, which can be directly loaded and used. The flight path planning, i.e., the flight path planning of the UAV in the demand response stage, is used to determine the trajectory of the UAV in this stage.

[0045] S13, controlling each UAV to enter the surrounding waiting stage according to the planned flight path;

[0046] S14, when all UAV members of the temporary task alliance enter the surrounding waiting stage, triggering the temporary task alliance to launch a synchronous operation on the target task;

[0047] S15, after completing the synchronous operation, dissolving the temporary task alliance, and each UAV switches to the task initial stage and waits for the demand response of a new target.

[0048] Understandably, after trajectory planning, for each UAV in a temporary mission coalition targeting a specific mission objective, each UAV flies according to its own flight path and enters a containment and waiting phase, waiting for other UAVs in the same temporary mission coalition to enter this phase. When all UAV members in the same temporary mission coalition have entered the containment and waiting phase, it triggers all UAVs in that temporary mission coalition to initiate synchronized operations on the mission objective. After the operations are completed, the temporary mission coalition disbands, and each UAV switches to the initial mission phase, enters a serpentine maneuver state, and waits for the response to the new mission objective.

[0049] The aforementioned self-organizing operation control method for heterogeneous UAV swarms, based on the Boids model, distributes the mission payload relatively evenly across the mission space in the initial stage. Then, a lightweight optimal control model is used to plan the flight paths of each UAV during the demand response phase. Each UAV is controlled to enter the encirclement and waiting phase according to the planned flight path. When all UAV members of the temporary mission alliance have entered this phase, the temporary mission alliance initiates synchronous operations on the mission target. Upon completion of the operations, the temporary mission alliance is disbanded, allowing each UAV to switch back to the initial mission phase, awaiting the demand response of the new mission target. Thus, based on the constructed information interaction network within the UAV swarm, distributed perception and information sharing of the operational situation are achieved. On this basis, loosely linked temporary mission alliances can be formed for multiple mission targets, aggregating mission functions scattered across multiple platforms into an adaptive dynamic execution network. Finally, the mission is completed according to a preset collaborative task operation pattern, significantly improving the mission efficiency of the UAV swarm system.

[0050] In one embodiment, further, during the mission execution process in the initial stage of the mission, the deployment process of each UAV in the mission space may specifically include the following processing:

[0051] Based on the interaction force data between UAVs in the mission space, the repulsive force data and the attractive force data experienced by the UAVs, the resultant force data experienced by the UAVs is calculated.

[0052] The position and velocity data of the UAV are updated using Euler integrals based on the combined force data at each time step.

[0053] It is understandable that, in practical applications, the behavior of the aforementioned self-organizing heterogeneous UAV swarms needs to be transformed into a mechanical model of interactions between point masses:

[0054]

[0055] Among them, F i s For U drones i The repulsive force experienced; x iand respectively, are the UAVs i current position and average position of other nodes in the separation region; ||·|| represents the Euclidean distance; ε s is a bias factor. Under the dynamic confrontation condition, the UAV individuals need to frequently and quickly avoid collision and respond to demand, and can be given sufficient freedom rather than fixed grouping of multiple UAVs, so the use of the two behavior rules of aggregation and alignment within the cluster is not involved.

[0056] In addition, in the task execution phase, in order to maintain the network patch, it is necessary to attract UAVs without task demand traction to relatively aggregate distribution in the task space at the self-organizing level of the cluster, to serve as routing nodes for information transmission of the task unit and backup nodes when attacked. In addition, it is necessary to limit the movement range of the UAV individuals within the task space. Therefore, two driving force models are added:

[0057]

[0058] where F i c is the attraction force received by the UAV U i . is the attraction force received by the UAV U i . b is the average position of the current known target. The size of the working space is set to [0, L], the task area is [L b , L-L b ], and L i is the gap size between the task area boundary and the working space boundary. When the UAV enters the gap, it will receive a boundary repulsion:

[0059]

[0060]

[0061] where, and are the components of the boundary repulsion received by the UAV U x in the x and y directions; gap y and gap i are the distances of the UAV U b to the working space boundary in the x and y directions; ε i is a bias factor.

[0062] Finally, the resultant force received by the UAV U i is:

[0063] F s (t)=ω i F s +ωc F i c +ω b F i b (5)

[0064] Wherein, the coefficient ω s These are the adjustment parameters for each component force. Assume that during mission execution, the UAV U... i The mass remains unchanged, and its position and velocity are updated for each time step using Euler integrals:

[0065]

[0066] Among them, v i For U drones i The velocity is Δt, where Δt is the time step.

[0067] In one embodiment, such as Figure 3 As shown, the process of using a lightweight optimal control model to plan the flight path for each UAV during the demand response phase in step S12 above can specifically include the following processing steps:

[0068] S121, based on the input parameters, instructs each UAV to determine a temporary mission alliance that meets the operational requirements of the mission objective through consensus negotiation.

[0069] It is understood that in this embodiment, it is assumed that each UAV is equipped with the same existing airborne expert system, which can calculate the operational requirements of the mission target and the optimal attack position based on information such as target attributes, platform performance and the situation of the operating environment, by comprehensively using fuzzy theory, the "Peresvet" special calculation program and power field theory. Each UAV then determines a temporary mission alliance that meets the operational requirements of the mission target through consensus negotiation.

[0070] S122, based on the solution information from the airborne expert system, instructs the UAVs in the temporary mission coalition to draw a circular path around the mission objective; the circular path consists of multiple equidistant scattered points, used for the UAVs to fly and wait for the other members of the temporary mission coalition to arrive.

[0071] The coordinated action process of a temporary mission alliance can be understood to consist of three steps: stealthy approach to the target, encirclement and waiting, and synchronized operation. After determining its response to the mission target, the UAV, based on information calculated by its onboard expert system, delineates a circular path around the target consisting of multiple equidistant points to await the arrival of the remaining members of the temporary mission alliance—this is the encirclement and waiting phase. It should be noted that the radii of the circular paths delineated by the heterogeneous UAVs, based on the characteristics of the mission target and their own functional attributes, exhibit gradient differences to form an optimal operational formation when initiating synchronized operations.

[0072] S123, using the light-weight optimal control model to plan a flight path for the UAV at multiple discrete points in the loop path and determine the end state of the planned flight path;

[0073] S124, instructing the UAV to select the discrete point with the minimum predicted arrival time as the optimal entry point and determine the flight path of the optimal entry point as the planned flight path;

[0074] S125, controlling the UAV to switch from the snake maneuvering state to the flight path state and fly according to the flight path.

[0075] It can be understood that the UAV plans a flight path for multiple discrete points in the loop path, and distinguishes the movement direction after entering the loop path, i.e., determines the end state of the planned flight path. Finally, the discrete point with the minimum predicted arrival time is selected as the optimal entry point, the corresponding movement direction is selected as the optimal entry direction, and the corresponding flight path is selected as the planned flight path. After the flight path is planned, the UAV switches from the snake maneuvering state to the flight path state and flies according to the flight path, waiting for the arrival of the rest of the temporary task alliance members.

[0076] Specifically, considering the cluster scale effect, the application uses the light-weight optimal control model of formula (7) to plan a flight path for a single UAV, and calculates the optimal control input at any time as:

[0077]

[0078] Wherein, u i is the control quantity of the UAV U i ; is the task execution position of the final arrival target T j , i.e., the optimal entry point; is the estimated time of arrival at the optimal entry point position of the target. The flight path is differentiated according to the time sequence, i.e., the entire flight path can be divided into multiple discrete sequence points. According to formula (7), the optimal control input of the UAV U i at each discrete point can be obtained, and the corresponding state quantity of the UAV at the point, including position and speed, can be obtained through integration operation.

[0079] In order to ensure that the flight path is real and flyable, the earliest estimated arrival time is selected as , and then the value of is appropriately increased, the maximum flight speed in the entire flight path is limited to be less than the set maximum flight speed value v m of the task, and the maximum angular velocity is limited to be less than the set maximum angular velocity value ω m of the task, i.e., the final planned flight path is considered to be real and flyable.

[0080] In the closing waiting phase, the UAVs enter their respective ring paths (tracks) and visit the next discrete point one by one to wait for the remaining members in the temporary task alliance. When all members of the temporary task alliance enter this state, the temporary task alliance is triggered to initiate a synchronous operation. After that, the temporary task alliance is dissolved, and each UAV switches back to the snake-shaped maneuvering state to wait for the demand response of a new target.

[0081] Further, in the dynamic confrontation condition, the situation of the operation environment changes sharply, and at any moment, new target tasks can be injected in a "pop-up" manner, and nodes of the task unit can fail or even the cascade failure of the task system. In addition, in actual applications, the topology of the heterogeneous UAV cluster communication network also changes dramatically, which will indirectly affect the performance of the system task. Based on such a task environment, the execution network of the heterogeneous UAV cluster is constructed in the present application. The essence of the process is to complete the good matching of multiple target tasks in the task space and multiple temporary task alliances formed by heterogeneous multi-UAVs.

[0082] Under the distributed architecture, each UAV decides to form a loose temporary task alliance with which UAVs based on the local information it masters, by calculating the marginal benefit of participating in the temporary task alliance for each target task, thereby forming a heterogeneous execution network for a specific target task. From the process of intelligent emergence, this process is not a traditional top-down task assignment, but rather a bottom-up task selection process.

[0083] Set the initial stage, the heterogeneous UAV cluster has N U UAVs , among which are N S reconnaissance UAVs, N A attack UAVs, N J electronic warfare UAVs, and N D decoy UAVs, N U = N S + N A + N J + N D ; in the present application, the initial arrangement order of each UAV in the cluster is used as its exclusive identity, i.e. ID(U i ) = i, and type(U i ) ∈ {S, A, J, D} represents the type of UAV U i (the UAV adopts modular design, and a single UAV only carries one type of task resource), and the specific heterogeneous ratio can refer to the ratio of the existing task system in the art.

[0084] There are N T initially preset heterogeneous target objects in the task scenario Assuming the solution of the onboard expert system of the UAV, each task target needs a certain number of heterogeneous UAVs to cooperate to complete the task execution, namely target T j There are multiple task requirements For example Indicates target T j Need to perform reconnaissance and attack tasks; target T j The task requirements of each target T For example Indicates that The reconnaissance aircraft and The attack aircraft jointly complete the task execution on target T j The task requirements of target T j The current task alliance is

[0085] The task requirements of target T j The grouping of various types of heterogeneous UAVs is For example Indicates that there are currently 2 attack aircraft targeting target T j As the work object. The target selection of each UAV at the global level is A = {a i ,..., a N}, element a i Indicates that the UAV U i selects the target number, that is, if U i ∈G(T j ), then Indicates that the UAV U i selects to execute the subtask of target T j and its own functional attribute is consistent. In particular, if the current UAV U i does not select any target object, then is an empty set, and the value of A reflects the execution network construction of the UAV cluster at the global level.

[0086] During the execution network construction process, the UAV U i realizes flexible switching between multiple states according to the decision result. According to the above content, the task execution state of the UAV U i is defined as es i ∈{0, 1, 2}, in order to indicate no target traction, target traction and fixed target selection; the motion state is defined as ms i∈ {-1, 0, 1, 2, 3}, which represent node failure, snake maneuver, track flight, circle waiting and synchronous operation in order. The state switching of multiple UAVs in the cluster finally presents intelligent emergent behavior of dynamic execution network construction for different targets at the global level.

[0087] In one embodiment, as shown in FIG. 16, the above method further comprises the following step S16: Figure 4

[0088] S16, calling the task performance model of the heterogeneous UAV cluster under the established dynamic time-varying condition to perform global task performance calculation on the task execution network composed of the heterogeneous UAV cluster, to obtain global task performance data; wherein the task performance model includes the nonlinear task performance of the temporary task alliance constructed for the first and second types of target tasks.

[0089] It can be understood that in the process of execution network construction, each UAV still follows the "if-then" rule to make action selection. Therefore, before studying the allocation method, it is necessary to first clarify the decision information and decision criteria of the UAV when making action selection. Under the distributed decision architecture, each UAV needs to calculate the marginal benefit brought to itself by each target according to: (1) the damage requirement of each target task; (2) the temporary task alliance formation situation for each target; (3) the ability attribute of the UAV itself; (4) the relative spatial position of the UAV itself and the target, and other local information within its own perspective, to make the best target selection. Among them, the key is to reasonably describe the temporary task alliance formation situation.

[0090] In order to reflect the nonlinear superposition of heterogeneous task performance achieved under the current alliance formation situation, the closed execution chain performance is used to describe the alliance task performance. A closed execution chain composed of heterogeneous UAVs should at least include reconnaissance and attack nodes. A task system composed of multiple nodes can form multiple closed execution chains, i.e. an execution network is constructed for a specific target, and the more execution chains in it represent the more operation modes of the task system to the target, which embodies the redundancy characteristics of the task system and is also a guarantee for effective damage to the target.

[0091] First, the task performance of the task node is valued according to the task attribute of the task node. For a reconnaissance node, its reconnaissance performance (sense) E S can be represented by the reconnaissance capability factor of the far boundary of its reconnaissance area, which embodies its ability to discover targets and monitor targets as far as possible. For an attack node, its attack performance (attack) E A ​The description embodies the ability of the node to attack the target. For the electronic warfare node, the jamming ability factor of the node to the target in the best attack formation represents the active jamming effectiveness E J . For the decoy node, the deception ability factor of the node to the target in the best attack formation represents the effectiveness E D .

[0092] Secondly, the response speed between the nodes is used to transfer the effectiveness between the nodes. Under the dynamic condition, the rapid closure of the execution chain mainly depends on the information cooperation of the nodes in the task system, which depends on the shortest communication topological distance between the nodes. The Dijkstra algorithm can be used to obtain the shortest communication distance d ik between any two nodes i and k in the communication network, and the response speed between the nodes is represented as where Δ is the distance sensitivity coefficient.

[0093] The targets in the task space can be divided into two categories. The first category of task targets is the targets without countermeasures such as airport runways, oil depots and repair plants, etc. The second category of task targets is the targets with countermeasures, mainly the enemy air defense positions, etc. The temporary task alliance for the first category of task targets only needs to include reconnaissance and attack nodes to achieve the expected operation effect, i.e. The node transition matrix M is defined to represent the conduction of the execution chain effectiveness between the nodes. The transition matrix between the target and the reconnaissance node is The transition matrix between the reconnaissance node and the attack node is The transition matrix between the attack node and the target is The nonlinear task effectiveness of the temporary task alliance constructed for the target is:

[0094]

[0095] For the second category of task targets, the attack effectiveness E A of the attack node is affected by the cover effectiveness E C of the electronic interference node and the decoy node, so the electronic warfare unmanned aerial vehicle and the decoy unmanned aerial vehicle are needed to cover the attack unmanned aerial vehicle to complete the task, i.e. Taking the electronic warfare node k covering the attack node i as an example, the attack effectiveness of the final attack node i is:

[0096]

[0097] wherein, is the response speed of the electronic warfare node k to any attack node i in the alliance. The attack effectiveness of the attack node i under the cover of all electronic warfare nodes and decoy nodes in the alliance is:

[0098]

[0099] wherein, node k and node l represent electronic warfare node and decoy node in the alliance respectively; σ represents the performance release level of attack node i under the condition of no cover task resource. The cover performance received by node i is defined as It should be noted that when the cover performance takes value E Cover > 1, E Cover = 1 is taken, which can avoid excessive occupation of cover resources in the cluster by a small number of attack nodes in the alliance. Then for the second type of task target, the transfer matrix between the attack node and the target is The formation of the alliance can be regarded as the performance release degree of the current task system in the task space, and the global task performance of the execution network formed by the current heterogeneous unmanned aerial vehicle cluster is:

[0100]

[0101] In one embodiment, as shown in Figure 5 the above method further comprises the following step S17:

[0102] S17, according to the global task performance data and the expected time of each type of unmanned aerial vehicle to reach the corresponding task execution position, the task expected income data of the temporary task alliance is calculated.

[0103] Specifically, on the basis of nonlinear description of heterogeneous task performance, the task expected income that can be obtained by the current temporary task alliance is:

[0104]

[0105] wherein, is the nominal return of target T j ; is the nonlinear performance sum formed by the demand D(T j ) of each type of unmanned aerial vehicle for target T j , if is taken λ is an adjustment coefficient; is the expected time of each type of unmanned aerial vehicle formation to reach its designated task execution position, the value is the arrival time of the unmanned aerial vehicle that arrives at the task position latest in the formation, if the formation size of this type of unmanned aerial vehicle is 0, then set For a larger value Q. The coalition formation in equation (12) as an exponential term can achieve the "reverse marginal diminishing effect", that is, with the increase of the overall task performance of the alliance, the increase speed of the expected return of the alliance is faster, which can attract unmanned aerial vehicles to preferentially join larger alliances in the subsequent execution network construction to complete the task as soon as possible, reduce the overall waiting time of the alliance, and thus improve the overall operation efficiency of the execution network.

[0106] In one embodiment, the above method further comprises the following steps:

[0107] Based on the potential game method, the marginal revenue function generated by the single unmanned aerial vehicle joining the temporary task alliance to execute the subtask of the task target is used to drive the single unmanned aerial vehicle to pursue the maximization of its current revenue, so as to realize the maximization of the global task execution revenue of the unmanned aerial vehicle cluster.

[0108] Specifically, the current global task revenue of the unmanned aerial vehicle cluster is set as From the perspective of potential game, the change trend of the global task expected revenue and the marginal revenue of the single unmanned aerial vehicle can be kept relatively consistent, that is, the global revenue of the cluster can be improved by driving the single unmanned aerial vehicle to greedily pursue the maximization of its current revenue. The marginal revenue function generated by the single unmanned aerial vehicle U i joining the alliance G(T j ) to execute the subtask of the target T j is set as:

[0109]

[0110] Wherein, A -i represents the task allocation of the remaining unmanned aerial vehicles in the cluster except the unmanned aerial vehicle U i ; represents the task alliance of the target containing the unmanned aerial vehicle U i , On the contrary. It should be noted that when the unmanned aerial vehicle individual calculates the task revenue by using the formula (13), the global task allocation A and the target alliance formation G(T j ) are based on the latest information in its own perspective, which may deviate from the true situation.

[0111] It should be understood that although Figures 1 to 5 the steps in the flowchart are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this article, the execution of these steps has no strict order limitation, and these steps can be executed in other orders. Moreover Figures 1 to 5At least some of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0112] In one embodiment, such as Figure 6 As shown, a heterogeneous UAV swarm self-organizing operation control system 100 is also provided, including a task initiation module 11, a task response module 13, an encirclement and waiting module 15, a job execution module 17, and a state switching module 19. The task initiation module 11 is used to deploy the task payload in a relatively dispersed and uniform manner in the task space based on the Boids model during the initial stage of the task; the task payload includes each UAV in the heterogeneous UAV swarm. The task response module 13 is used to obtain the input parameters of the trajectory planning algorithm and the task requirement information, and to perform trajectory planning for each UAV in the requirement response stage using a lightweight optimal control model; the input parameters include the initial state constraints and terminal state constraints of the UAV. The encirclement and waiting module 15 is used to control each UAV to enter the encirclement and waiting stage according to the planned flight trajectory. The job execution module 17 is used to trigger the temporary task alliance to initiate synchronous operations on the task target when all UAV members of the temporary task alliance have entered the encirclement and waiting stage. The state switching module 19 is used to disband the temporary task alliance after the synchronous operations are completed, and each UAV switches to the task initiation stage to wait for the requirement response of the new task target.

[0113] The aforementioned heterogeneous UAV swarm self-organizing operation control system 100, based on the Boids model, distributes the mission payload relatively evenly across the mission space in the initial stage of the mission. Then, a lightweight optimal control model is used to plan the flight paths of each UAV during the demand response phase. Each UAV is controlled to enter the encirclement and waiting phase according to the planned flight path. When all UAV members of the temporary mission alliance have entered this phase, the temporary mission alliance is triggered to initiate synchronous operations on the mission target. Upon completion of the operations, the temporary mission alliance is disbanded, allowing each UAV to switch back to the initial mission phase, awaiting the demand response of a new mission target. Thus, based on the constructed information interaction network within the UAV swarm, distributed perception and information sharing of the operational situation are achieved. On this basis, loosely linked temporary mission alliances can be formed for multiple mission targets, aggregating mission functions scattered across multiple platforms into an adaptive dynamic execution network. Finally, the mission is completed according to a preset collaborative task operation pattern, significantly improving the mission efficiency of the UAV swarm system.

[0114] In one embodiment, the task execution state of the unmanned aerial vehicle in the heterogeneous unmanned aerial vehicle cluster includes no target traction, target traction and fixed target selection, and the motion state of the unmanned aerial vehicle includes snake maneuvering, flight along a track, waiting in a circle, synchronous attack and node failure.

[0115] In one embodiment, the task initial module 11 can be specifically configured to calculate the resultant force data of the unmanned aerial vehicle according to the interaction force data between the unmanned aerial vehicles in the task space, the repulsion force data and the attraction force data of the unmanned aerial vehicle; and update the position data and the speed data of the unmanned aerial vehicle after each time step by using Euler integral method according to the resultant force data.

[0116] In one embodiment, the task response module 13 can be specifically configured to instruct the unmanned aerial vehicles to determine a temporary task alliance meeting the task target operation requirement through consistent negotiation according to the input parameters; instruct the unmanned aerial vehicles in the temporary task alliance to set a circular path around the task target according to the information calculated by the on-board expert system; the circular path is composed of a plurality of equidistant discrete points, which are used for the unmanned aerial vehicle to wait for the arrival of the remaining members in the temporary task alliance; perform track planning for the plurality of discrete points in the circular path of the unmanned aerial vehicle by using a lightweight optimal control model and determine the end state of the planned flight track; instruct the unmanned aerial vehicle to select the discrete point with the minimum predicted arrival time as the best entry point and determine the flight track of the best entry point as the planned flight track; control the unmanned aerial vehicle to switch from the snake maneuvering state to the flight along a track state and fly along the flight track.

[0117] In one embodiment, the self-organizing operation control system 100 of the heterogeneous unmanned aerial vehicle cluster described above can further include a global task performance calculation module, configured to call the task performance model of the heterogeneous unmanned aerial vehicle cluster under the dynamic time-varying condition to perform global task performance calculation on the task execution network composed of the heterogeneous unmanned aerial vehicle cluster, and obtain global task performance data; wherein the task performance model includes the nonlinear task performance of the temporary task alliance constructed for the first type of task target and the second type of task target, respectively.

[0118] In one embodiment, the self-organizing operation control system 100 of the heterogeneous unmanned aerial vehicle cluster described above can further include a benefit calculation module, configured to calculate the task expected benefit data of the temporary task alliance according to the global task performance data and the predicted time of each type of unmanned aerial vehicle to arrive at the corresponding task execution position.

[0119] In one embodiment, the self-organizing operation control system 100 of the heterogeneous unmanned aerial vehicle cluster described above can further include a benefit driven module, configured to drive a single unmanned aerial vehicle to pursue the maximization of its current benefit based on the marginal benefit function generated by the execution of the subtask of the task target after the single unmanned aerial vehicle joins the temporary task alliance, based on the potential game method.

[0120] With regard to the specific definition of the heterogeneous unmanned aerial vehicle cluster self-organizing operation control system 100, reference can be made to the corresponding definition of the heterogeneous unmanned aerial vehicle cluster self-organizing operation control method in the foregoing, which will not be described here again. Each module in the above-mentioned heterogeneous unmanned aerial vehicle cluster self-organizing operation control system 100 can be realized by software, hardware and a combination thereof in whole or in part. The above-mentioned modules can be embedded in or independent of a device with a specific data processing function in hardware form, or can be stored in the memory of the aforementioned device in software form, so as to be called and executed by the processor to perform the operation corresponding to each of the above-mentioned modules. The aforementioned device can be, but is not limited to, various types of portable, vehicle-mounted or ship-mounted unmanned aerial vehicle cluster control devices in the prior art.

[0121] In one embodiment, a computer device is also provided, comprising a memory and a processor, the memory storing a computer program, and the processor implementing the following processing steps when executing the computer program: based on the Boids model, uniformly deploying task loads in a task space in a relative dispersion manner in an initial stage of a task; the task loads comprising unmanned aerial vehicles in a heterogeneous unmanned aerial vehicle cluster; obtaining input parameters of a path planning algorithm and task demand information, and using a lightweight optimal control model to perform path planning for the unmanned aerial vehicles in a demand response stage; the input parameters comprising initial state constraints and terminal state constraints of the unmanned aerial vehicles; controlling the unmanned aerial vehicles to enter a siege waiting stage according to the planned flight path; when all unmanned aerial vehicle members of a temporary task alliance enter the siege waiting stage, triggering the temporary task alliance to initiate a synchronous operation on a task target; after completing the synchronous operation, disbanding the temporary task alliance, and each unmanned aerial vehicle switching to the initial stage of the task to wait for a demand response of a new task target.

[0122] It can be understood that, in addition to the memory and the processor mentioned above, the above-mentioned computer device also comprises other software and hardware components not listed in the present specification, which can be determined according to the specific data processing and control device model in different application scenarios, and the present specification will not be listed and described in detail.

[0123] In one embodiment, the processor executing the computer program can also implement the steps or sub-steps added in each embodiment of the above-mentioned heterogeneous unmanned aerial vehicle cluster self-organizing operation control method.

[0124] In one embodiment, a computer readable storage medium is also provided, and the computer program is stored on the computer readable storage medium, and the computer program is executed by a processor to implement the following processing steps: uniformly deploying task loads in a task space in an initial stage of a task based on a Boids model, the task loads comprising each unmanned aerial vehicle in a heterogeneous unmanned aerial vehicle cluster; obtaining input parameters of a path planning algorithm and task demand information, and using a lightweight optimal control model to plan a flight path for each unmanned aerial vehicle in a demand response stage; the input parameters comprising initial state constraints and end state constraints of the unmanned aerial vehicle; controlling each unmanned aerial vehicle to enter a siege waiting stage according to the planned flight path; when all unmanned aerial vehicle members of a temporary task alliance enter the siege waiting stage, triggering the temporary task alliance to launch a synchronous operation on a task target; after completing the synchronous operation, disbanding the temporary task alliance, and each unmanned aerial vehicle switching to the initial stage of the task to wait for a demand response of a new task target.

[0125] In one embodiment, the computer program, when executed by the processor, can also implement the steps or sub-steps added in each embodiment of the self-organizing operation control method of the heterogeneous unmanned aerial vehicle cluster.

[0126] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiments. Any reference to memory, storage, database or other medium used in each embodiment provided by the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus dynamic random access memory (Rambus DRAM, RDRAM) and interface dynamic random access memory (DRDRAM).

[0127] Each technical feature of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of each technical feature in the above embodiments are described, however, as long as the combinations of the technical features do not exist, they should be considered as the scope of the present disclosure.

[0128] The above embodiments only express several implementation ways of the present application, and the description is specific and detailed, but it should not be understood as a limitation to the patent scope of the application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, several modifications and improvements can be made, which all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A heterogeneous unmanned aerial vehicle cluster self-organizing operation control method, characterized in that, The method comprises the steps of: Based on the Boids model, the task load is uniformly deployed in the task space in the initial stage of the task; the task load includes each unmanned aerial vehicle in a heterogeneous unmanned aerial vehicle cluster, wherein in the initial stage of the task, each unmanned aerial vehicle is in a self-driven snake-like maneuvering mode without planning a flight path, and inter-machine collision avoidance is achieved by sensing the positions of surrounding friendly neighbor unmanned aerial vehicles; Input parameters of a flight path planning algorithm and task demand information are obtained, and a lightweight optimal control model is used to plan a flight path for each unmanned aerial vehicle in a demand response stage; the input parameters include initial state constraints and end state constraints of the unmanned aerial vehicle; Each unmanned aerial vehicle enters a surrounding waiting stage according to the planned flight path; When all unmanned aerial vehicle members of the temporary task alliance enter the surrounding waiting stage, the temporary task alliance initiates synchronous operation on the task target; the unmanned aerial vehicle in the surrounding waiting state reenters the snake-like maneuvering state after a set waiting countdown ends; After the synchronous operation is completed, the temporary task alliance is dissolved, and each unmanned aerial vehicle switches to the initial stage of the task to wait for a demand response of a new task target; wherein the lightweight optimal control model is as follows: in, For drones The control quantity, t For time, For drones speed, To ultimately reach the goal The task execution location, For drones Current location The estimated time to reach the optimal entry point of the target; A task performance model of a heterogeneous unmanned aerial vehicle cluster under dynamic time-varying conditions is called to calculate the global task performance of a task execution network composed of the heterogeneous unmanned aerial vehicle cluster to obtain global task performance data; wherein the task performance model includes a nonlinear task performance of a temporary task alliance constructed for a first type of task target and a second type of task target; Based on a potential game method, a marginal revenue function generated by a single unmanned aerial vehicle performing a subtask of a task target after joining the temporary task alliance is used to drive the single unmanned aerial vehicle to maximize its current revenue.

2. The heterogeneous UAV swarm self-organizing operation control method according to claim 1, wherein, The task execution state of the unmanned aerial vehicle in the heterogeneous unmanned aerial vehicle cluster includes no target traction, target traction, and fixed target selection, and the motion state of the unmanned aerial vehicle includes snake-like maneuvering, flight path flying, surrounding waiting, synchronous attack, and node failure.

3. The heterogeneous UAV swarm self-organizing operation control method according to claim 1 or 2, characterized in that, In the task execution process in the initial stage of the task, the deployment process of each unmanned aerial vehicle in the task space comprises: According to the interaction force data between the unmanned aerial vehicles in the task space, the repulsion force data and the attraction force data received by the unmanned aerial vehicles, the resultant force data received by the unmanned aerial vehicles is calculated; According to the resultant force data, the position data and the speed data of the unmanned aerial vehicles after each time step are updated in an Euler integral manner.

4. The heterogeneous UAV swarm self-organizing operation control method according to claim 1, wherein, The process of planning a flight path for each unmanned aerial vehicle in a demand response stage by using a lightweight optimal control model comprises: According to the input parameters, each unmanned aerial vehicle determines a temporary task alliance that meets the operation demand of the task target through consistent negotiation; According to the information calculated by the on-board expert system, the unmanned aerial vehicles in the temporary task alliance set a ring-shaped path around the task target; the ring-shaped path is composed of multiple equidistant discrete points, and is used for the unmanned aerial vehicles to fly and wait for the arrival of the remaining members in the temporary task alliance; The lightweight optimal control model is used to plan a flight path for the multiple discrete points in the ring-shaped path of the unmanned aerial vehicle and determine the end state of the planned flight path. The UAV is instructed to select a discrete point with the minimum predicted arrival time as the optimal entry point and determine a flight path of the optimal entry point as a planned flight path; The UAV is controlled to switch from the snake-like maneuvering state to the flight path flying state and fly along the flight path.

5. The heterogeneous UAV swarm self-organizing operation control method according to claim 1, wherein, The method further comprises: Based on the global task performance data and the predicted time of each type of UAV to reach the corresponding task execution location, the task expected benefit data of the temporary task alliance is calculated.

6. A heterogeneous unmanned aerial vehicle cluster self-organizing operation control system, characterized in that, It comprises: A task initial module is configured to relatively disperse and uniformly deploy task loads in a task space in a task initial stage based on a Boids model; The task loads comprise each UAV in a heterogeneous UAV cluster, wherein in the task initial stage, each UAV does not need to plan a flight path but adopts a self-driven manner to perform snake-like maneuvering and avoid collision between UAVs by perceiving the positions of surrounding friendly neighbor UAVs; A task response module is configured to obtain input parameters of a flight path planning algorithm and task demand information and plan a flight path for each UAV in a demand response stage by using a lightweight optimal control model; the input parameters comprise initial state constraints and terminal state constraints of the UAVs; A surrounding waiting module is configured to control each UAV to enter a surrounding waiting stage according to the planned flight path; A work execution module is configured to trigger the temporary task alliance to launch a synchronous work on a task target when all UAV members of the temporary task alliance enter the surrounding waiting stage; and the UAV in the surrounding waiting state reenters the snake-like maneuvering state after a set waiting countdown ends. A state switching module is configured to disband the temporary task alliance after the synchronous work is completed, each UAV switches to the task initial stage and waits for a demand response of a new task target; wherein the lightweight optimal control model is as follows: in, For drones The control quantity, t For time, For drones speed, To ultimately reach the goal The task execution location, For drones Current location The estimated time to reach the optimal entry point of the target; A task performance model of a heterogeneous UAV cluster under dynamic time-varying conditions is called to perform global task performance calculation on a task execution network composed of the heterogeneous UAV cluster to obtain global task performance data; wherein the task performance model comprises nonlinear task performance of a temporary task alliance constructed for a first type of task target and a second type of task target respectively; Based on a potential game method, a marginal benefit function generated by a single UAV performing a subtask of a task target after joining the temporary task alliance is used to drive the single UAV to pursue the maximization of its current benefit. 7.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-6 when the computer program is executed by the processor. The processor executes the computer program to implement the steps of the self-organizing operation control method of the heterogeneous UAV cluster in any one of claims 1 to 5.

8. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the self-organizing operation control method of the heterogeneous UAV cluster in any one of claims 1 to 5.

Images