A method for group-to-group countermeasure information oriented guidance of distraction control of aircraft cluster

Abstract

The application particularly relates to a cluster-oriented distraction control method based on group counter-group countermeasure information guidance. The method comprises the following steps: acquiring a node set of a target unmanned cluster and an information interaction relationship between nodes at discrete time steps, and constructing an information interaction network of the target cluster; selecting a distraction node set from the target cluster according to the information interaction network, and applying external information input to the distraction nodes to obtain shielded real information input and deceptive information; fusing the shielded real residual information and the injected deceptive information to form equivalent input, which is applied to an information state updating process of the distraction nodes, so that local cognitive bias is coupled and propagated through the network, and distraction control of overall behavior evolution of the target cluster is realized. The method can realize continuous distraction and guidance of the target cluster without relying on large-scale physical destruction and without excessively relying on continuous high-power suppression.

Description

Technical Field

[0001] This invention relates to the field of information countermeasures technology for multiple unmanned systems, specifically to a restraining control method for "group-to-group" countermeasures against aircraft swarms. This method dynamically selects restraining nodes within the target swarm by randomly shielding and selectively injecting information into it, aiming to achieve restraining control over the overall situational awareness and behavioral evolution of the target swarm with low intervention costs. Background Technology

[0002] With the rapid development of unmanned systems technology, distributed collaborative decision-making, and swarm intelligence, the application of aircraft swarms in "countering swarms with swarms" is becoming increasingly widespread in real-world scenarios, and has become a crucial force that cannot be ignored in modern security patrols and other scenarios. In typical countermeasures missions, individual aircraft in a swarm often rely on local communication, sensor observation, and neighbor interaction to form situational awareness, and on this basis, complete decision-making behaviors such as formation maintenance, task allocation, collaborative search, collaborative penetration, or interception evasion. Since the overall behavior of an aircraft swarm is driven by both networked information flow and local situational awareness, if effective intervention can be implemented at the information layer on key nodes of the swarm to create controllable deviations in their perception-decision-making links, these deviations can be propagated and amplified through the swarm network, thereby achieving the restraint and guidance of the target swarm at a relatively low cost. However, factors such as unstable communication links, frequent node failures, and drastic topology changes in swarm countermeasure scenarios make it difficult to select targets, control the intensity, and deploy real-time interventions at the information layer, making it difficult to meet the requirements of real-time performance, robustness, and effectiveness for countermeasures missions. Therefore, developing information intervention and restraint control methods that can address the above challenges has become one of the core problems that urgently need to be solved in the field of multi-unmanned system information countermeasures technology.

[0003] In existing technologies, countermeasures against aircraft swarms typically rely on hard-kill and soft-kill methods. Hard-kill methods directly eliminate or damage enemy platforms to reduce the number of target swarms and their mission capabilities. While these methods are direct and effective, they often suffer from the problem of saturation consumption of large interception resources when facing swarm targets. Furthermore, hard-kill methods are limited by short interception windows and high hit difficulty in complex environments and under highly maneuverable conditions, making it difficult to achieve continuous, low-cost behavioral control and guidance of target swarms in engineering applications.

[0004] To reduce the resource consumption and constraints of hard-kill attacks, soft-kill methods employ electronic interference to reduce the signal-to-noise ratio, cause packet loss, or delay information, suppressing and disrupting the communication and sensing links at the communication level, thus weakening the cluster's information sharing capabilities and collaborative effects. Compared to hard-kill attacks, soft-kill attacks have advantages such as relatively low cost and repeatability, making them more suitable for the information intervention tasks requiring continuous pressure in cluster countermeasures. However, traditional soft-kill methods manifest as capability weakening rather than controllable restraint: their effects often manifest as reduced communication quality or disrupted information consistency in the target cluster, ultimately leading to collaborative degradation or partial failure, but they are unlikely to change the behavior of the opposing cluster according to the desired objective. Furthermore, soft-kill attacks typically require high power and spectrum resources, are susceptible to complex electromagnetic environments, obscured terrain, the opponent's anti-interference mechanisms, and multi-link redundancy strategies, resulting in significant uncertainty and scenario dependence in their actual effectiveness.

[0005] Therefore, there is an urgent need for a new method that can overcome the limitations of existing hard-kill and soft-kill methods: to achieve more efficient, stable, and directional control over the overall behavior of aircraft swarms without relying on large-scale platform destruction or excessive reliance on high-power suppression, thereby improving the controllability and effectiveness of countermeasures. Summary of the Invention

[0006] This invention aims to address the technical deficiencies in existing "group-against-group" information countermeasures methods for aircraft swarms. Existing technologies typically focus on reducing platform capabilities, treating direct damage and link suppression as relatively independent, phased measures. They lack a holistic restraint method aimed at guiding the coordinated behavior of the swarm, resulting in information countermeasures often manifesting as localized capability degradation or node disorder. This makes it difficult to provide stable and continuous guidance for the behavioral evolution of the target swarm, reducing the certainty and controllability of restraint. Furthermore, traditional methods often rely on fixed-pattern suppression or interception strategies, which are ill-suited to address the continuous changes in swarm structure and interaction relationships, node damage and exit, and dynamic events such as switching of coordination strategies during swarm countermeasures, resulting in severe lack of adaptability. Therefore, this invention provides a novel information countermeasure method that achieves proactive restraint and guidance of the behavioral evolution of the target swarm.

[0007] Other features and advantages of the invention will become apparent from the following detailed description, or may be learned in part by practice of the invention.

[0008] According to a first aspect of the present invention, a swarm control method for aircraft ensembles guided by swarm counter-swarm information is provided, the method comprising: Under discrete time steps, obtain the node set of the target unmanned cluster and the information interaction relationship between the nodes, and construct the information interaction network of the target cluster; Based on the information exchange network, a set of restraining nodes is selected from the target cluster, and the following external information inputs are applied to the restraining nodes: For each constrained node, a random blocking mechanism is used to probabilistically block information input from its real neighbors, so as to form a blocked real information input; For each restraint node, a targeted injection mechanism is used to generate deception information, where the injection strength of the deception information is adjusted by a continuously controlled gain parameter, and the executable injection scale is obtained through discrete mapping. By fusing the masked real information input with the injected deceptive information, the equivalent information input of the restraining node is obtained. This information is then used to influence the information state of the non-restraining node through the coupling propagation of the information interaction network, thereby achieving restraint and control over the overall behavioral evolution of the target unmanned cluster. The information interaction network, the set of restraining nodes, and the equivalent information input are iteratively updated at discrete time steps to output the state sequence and restraining result in the countermeasure process.

[0009] In some exemplary embodiments, the random masking mechanism is Bernoulli gated masking: a communication state random variable is introduced into the information link between the restraining node and each of its real neighbors, and it is made to follow a Bernoulli distribution; when the communication state is zero, the corresponding neighbor information input is masked, and otherwise the corresponding neighbor information input is retained.

[0010] In some exemplary embodiments, the targeted injection mechanism maps the continuous control gain to the injection size of discrete spoof nodes using an up-rounding operator.

[0011] In some exemplary embodiments, the directional injection mechanism generates the injection direction using a directional parameterization method with a fixed polar radius and an adjustable polar angle: the injection radius is set to a fixed constant value, and the injection direction is controlled by the polar angle offset parameter.

[0012] In some exemplary embodiments, the selection of the set of restraining nodes includes: The degree centrality, betweenness centrality, and eigenvector centrality of candidate nodes are calculated on the current information exchange network, and then fused according to weights to obtain a comprehensive score. Based on the comprehensive score and coverage gain, a comprehensive utility function is constructed. Nodes that maximize utility and improve network coverage are selected in turn and added to the set of restraining nodes until the preset number of restraining nodes is reached or all nodes are covered.

[0013] In some exemplary embodiments, the number of restraining nodes is a fixed size. During the countermeasure process, when the target cluster nodes exit or the number of nodes is less than the fixed size due to topology reconstruction, the restraining node index sequence is zero-filled or repeatedly mapped to maintain the consistency of the external information input channel dimension.

[0014] According to a second aspect of the present invention, a swarm control system for aircraft ensembles guided by swarm counter-swarm information is provided, the system comprising: The information interaction network construction module is used to construct a network coupling evolution model of the information interaction network topology and node information status based on the interaction rules between target cluster nodes. The module for selecting key control nodes is used to determine the set of control nodes by combining node centrality and coverage gain under a fixed control size constraint. The random blocking module is used to probabilistically block the input of real neighbor information of the restraining node in a Bernoulli gating manner; The directional injection module is used to generate discrete injection scales by mapping continuous injection gain with rounding up, and to generate directional deception information using a fixed polar radius and an adjustable polar angle bias. The countermeasure information-guided restraint control module is used to merge the shielded real residual information with the injected deceptive information to form an equivalent input, and to act on the information state update process of the restraint node, so that local cognitive biases are propagated through network coupling, thereby achieving restraint control over the overall behavioral evolution of the target cluster.

[0015] In some exemplary embodiments, the control key node selection module further includes a dynamic update unit, which is used to recalculate the centrality score and coverage gain based on the current network structure and update the control node set when the target cluster topology changes.

[0016] In some exemplary embodiments, the random shielding module and the directional injection module act on the same set of restraining nodes, and support setting different shielding probabilities and injection direction biases for different types of interactive links.

[0017] In some exemplary embodiments, the system is deployed in a central processing unit or a distributed processing unit and is able to access target cluster situational parameters, interactive network data and node information status data, and output intervention strategies and the updated results of the restraint node set.

[0018] Compared with existing technologies, this invention, by constructing a collaborative intervention mechanism of random shielding and targeted injection, can achieve the restraint control of countermeasure information guidance for target clusters without relying on large-scale physical damage or excessive reliance on high-power continuous suppression. Simultaneously, this invention provides a key restraint node selection and update mechanism oriented towards the dynamic topology of the target cluster. This mechanism can rapidly update the set of restraint nodes based on the current network structure in cases of node damage and exit, link fluctuations, and topology reconstruction, ensuring effective coverage of the restraint control target and maintaining the continuity and stability of the restraint effect. In summary, the information intervention restraint control method proposed in this invention, while ensuring the "group against group" effect, possesses reliable engineering feasibility: on the one hand, the random shielding mechanism of real information conforms to the objective characteristic of information loss caused by suppression interference, facilitating payload application deployment; on the other hand, the targeted injection of deceptive information allows for free configuration of the information intervention intensity and scope to adapt to the task requirements in complex cluster countermeasure scenarios, demonstrating significant practical value.

[0019] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0020] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention. It is obvious that the drawings described below are merely some embodiments of the invention, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0021] Figure 1 This is a schematic diagram of the information interaction network model of the target cluster in an embodiment of the present invention.

[0022] Figure 2 This is a schematic diagram of the random masking intervention process for real information in an embodiment of the present invention.

[0023] Figure 3 This is a schematic diagram of the false information injection intervention process in an embodiment of the present invention.

[0024] Figure 4 This is a flowchart of the method in an embodiment of the present invention. Detailed Implementation

[0025] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the invention will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0026] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0027] To address the shortcomings and deficiencies of existing technologies, this example implementation provides a restraint and control method for aircraft swarms that uses "group against group" to counter information guidance. By modeling the information interaction network of the target swarm, an information intervention mechanism oriented towards the target nodes is established. Virtual node intervention channels are introduced into the information interaction network. Key nodes of the target swarm are selected, and real neighbor information is randomly masked while false virtual entities are injected in a targeted manner, thereby restraining and guiding the overall behavioral evolution of the target swarm.

[0028] The aircraft swarm of the present invention consists of multiple platforms, including but not limited to unmanned aerial vehicles (UAVs) or heterogeneous combinations thereof; an information intervention unit for generating and outputting intervention information streams for target nodes; a central or distributed processing unit for executing the information intervention and containment control method of the present invention, wherein the processing unit is capable of accessing target swarm situational parameters, target swarm interaction network data, and target node information status data, and outputting intervention strategies and target node set update results; and a communication network for ensuring the delivery of intervention information streams and the acquisition of target swarm interaction data, forming a closed-loop containment control system in the scenario of swarm information intervention countermeasures.

[0029] refer to Figure 4 As shown in the example embodiment of the present invention, a countermeasure control method for aircraft swarms using "swarm-to-swarm" counter-information guidance may specifically include the following steps: Step S1: Under discrete time steps, obtain the node set of the target unmanned cluster and the information interaction relationship between the nodes, and construct the information interaction network topology of the target cluster; Step S2: Select a set of restraining nodes from the target cluster based on the information exchange network, and apply the following external information input to the restraining nodes: For each constrained node, a random blocking mechanism is used to probabilistically block information input from its real neighbors, so as to form a blocked real information input; For each restraint node, a targeted injection mechanism is used to generate deception information, where the injection strength of the deception information is adjusted by a continuously controlled gain parameter, and the executable injection scale is obtained through discrete mapping. Step S3: The masked real information input is fused with the injected deceptive information to obtain the equivalent information input of the restraining node, and the information state of the non-restraining node is affected through the coupling propagation of the information interaction network, thereby realizing the restraint and control of the overall behavior evolution of the target unmanned cluster. Step S4: Iteratively update the information interaction network topology, the set of restraining nodes, and the equivalent information input at discrete time steps to output the state sequence and restraining result in the countermeasure process.

[0030] The method of this invention can achieve continuous containment and guidance of target clusters without relying on large-scale physical destruction or excessive reliance on continuous high-power suppression. It is applicable to application scenarios such as multi-unmanned system countermeasure simulation, training simulation and aircraft cluster control.

[0031] The steps in this exemplary embodiment will now be described in more detail with reference to the accompanying drawings and embodiments.

[0032] The information interaction network model in step S1 is specifically as follows: In a scenario of information guidance for countering aircraft swarm attacks, the maneuvering behavior of individual swarm members is driven by their perception of the local environment, neighboring nodes, and threat situation. To characterize the information propagation process of the target swarm in the perception-decision closed loop, an information interaction network model of the target swarm is established, such as... Figure 1 As shown.

[0033] Let the set of individuals in the target cluster be... ,in This represents the number of individuals in the target cluster. The local information interaction relationships formed between individuals in the target cluster under the sensing communication link can be represented as an undirected simple graph. edge set This represents the connectivity relationships that allow information exchange between individuals within the target cluster. Accordingly, an adjacency matrix is ​​defined. If and only if season ,otherwise In this network description, nodes The degree is defined as: (1) To characterize the topological density and local propagation capability of the target cluster information interaction network, the maximum degree corresponding to the current target cluster network is defined as: (2) During the movement of the target cluster, each individual in the target cluster The physical state can be determined by position and velocity. The description states that its acceleration is composed of both a self-driving term and an interaction potential field term, which then evolves into complex swarm cooperative behavior. Building upon this, to further characterize the target swarm's comprehensive perception of information such as strength comparison, threat situation, and cooperative structure at the information layer, the information layer situational awareness state of individual members of the target swarm is introduced. Characterizing individual members of the target cluster Local situational awareness formed under the influence of information. Since the interaction topology and potential function of individual members in the target cluster depend on the local information of their neighboring node sets, the situational awareness of the target cluster's information layer can be written in the form of general network coupling dynamics: (3) in, , Representing individuals in the target cluster There exists a set of friendly neighbors and a set of target neighbors that interact with each other, and the function... Determined by the adopted population dynamics model and interaction potential field function; To act on individual members of the target cluster External information input items represent information interventions applied to the information links of the target cluster. Therefore, based on the target cluster information interaction network, a set of virtual nodes with information interaction capabilities is introduced. and in the target cluster node set A subset of nodes selected that are subject to information intervention Its cardinality satisfies For any individual in the target cluster ,when At that time, there exists a corresponding virtual node. Establish information channels to ensure that the information input of individual nodes in the target cluster meets the requirements. ;when At that time, it is assumed that the node is not affected by external information, that is Based on this, the information layer states and inputs of all individual members of the target cluster are represented as vectors by nodes. , The relationship between the selection of virtual nodes and their corresponding target cluster nodes, influenced by information intervention, can be represented by a constraint indication matrix. , among which when hour ,when hour .

[0034] In step S2, the random masking mechanism is used to probabilistically block information input from real neighbors, specifically as follows: Based on the information interaction network model of the target cluster, the first type of real information intervention based on random shielding is considered: by reducing the signal-to-noise ratio of the individual communication links of the target cluster to simulate channel fading and data loss under electromagnetic suppression, the real neighbor information of the target node is blocked on the information input side, so that the perception data packets received by the target node are lost or misverified with a certain probability, thereby weakening the target node's acquisition of real situation information and causing deviations in the process of local situation cognition.

[0035] Let the target node be Its true neighbor set is The actual number of neighbors is For any real neighbor , set the target node under information intervention With real neighbors The connectivity state of the sensing link is represented as a binary random variable. Follows Bernoulli distribution: (4) in The probability of masking real node information. The larger the value, the stronger the shielding, and the greater the probability of the communication link being blocked. When When, it indicates the target node Strong interference caused real neighbors Information masking occurs due to packet loss or checksum errors; when When, it indicates the target node Able to receive real neighbors normally The true situation information.

[0036] target node under random shielding The set of neighbors can be represented as: (5) The number of neighbors under the corresponding random shielding effect is denoted as . It can be represented as: (6) Although the specific neighbor nodes that are blocked at each moment change randomly, the number of real neighbors that the target node ultimately effectively perceives satisfies the following statistical expectation: (7) Therefore, through parameters Achieve continuously adjustable probabilistic control based on the real neighbor information available to the target node: when As the size increases, the target node retains less real neighbor information; when When the target node is reduced, it retains more real neighbor information. Unlike deterministic masking of specific neighbors, random masking does not require pre-specifying which neighbors to mask. Instead, it probabilistically blocks adjacent links, causing the masked objects to change randomly over time. This aligns with the uncertainty of information transmission caused by suppressive interference in actual countermeasures, and avoids solidifying information intervention into a dependence on a fixed set of neighbors. This facilitates continuous control and guidance of the target cluster under conditions such as link fluctuations and topology changes.

[0037] In the specific process of random blocking intervention of real information, random blocking variables and blocking probability parameters can be set for different types of information interaction links, such as Figure 2 As shown, by applying different methods to the collaborative neighborhood links and the target cluster neighborhood links respectively... Adjustments are made to differentiate and weaken the input of real information from different sources. Thus, the command unit can adjust its approach based on the target cluster situation and intervention resource constraints. The value of the value is used for online scheduling, thereby achieving a balance between intervention intensity and resource consumption while ensuring that intervention is feasible.

[0038] In step S2, the deceptive information is generated using a targeted injection mechanism as follows: Based on the information interaction network model of the target cluster, a second type of deceptive information intervention based on targeted injection is further considered: false situational interaction information is projected directionally from virtual nodes to the target cluster nodes, allowing the deceptive information to be injected into the communication link between the target node and its allies in the form of real node interaction information, thereby causing a deviation from the true situational information during the local situational awareness process. To address this, continuous control gain is employed. Characterization of target node The strength of the deceptive information injection. Let the target node be... The number of real neighbors at the current moment is The discrete scale increment of deception injection of fake nodes is defined as: (8) in, This represents the floor operator, which is adjusted... This allows for continuous control over the discrete scale of deceptive injected fake nodes. .

[0039] Based on this, the target node The relative positions of the injected spoof nodes are represented in polar coordinates. This takes into account the spoof nodes relative to the target node. The influence of distance on the interaction potential field, and the continuous control gain corresponding to the information injection intensity. Consistent. Therefore, the polar radius of spurious nodes is a fixed constant value. And through polar angle bias To achieve directional control of the interaction potential field. For the ... A fake node injected. Its relative position parameters are: (9) in, Inject the radius to a preset constant value; The reference direction is determined by the direction constrained by the task's expected direction; This is the polar angle offset parameter, used to control the deflection angle of the deception injection relative to the reference direction.

[0040] To uniformly describe the effect of deception information on the updating of target node information, the deception information generated by virtual nodes is represented as an injection vector. This represents the total deception information injected into the target node during information state updates: (10) in, Used to construct false friendly cooperative situations and interaction information, causing the target node to deviate from the relative position, number, or cooperative behavior of its friendly neighbors; Used to construct a false threat posture of the target cluster, causing the target nodes to deviate from the direction or structure of the threat.

[0041] Specific processes of injecting false information into interventions, such as Figure 2 As shown. For each target node Virtual nodes Inject deceptive information using the following process: First, obtain the target node. Current information status and local reference direction Then, based on the control gain... Calculate the injection scale of discrete spurious nodes Next, a fixed polar diameter is adopted. With polar angle bias Generate relative position parameters for dummy nodes This process generates corresponding virtual situational information; finally, the virtual information is encapsulated into an injection vector. Injecting false information into target nodes allows them to incorporate it into their situational awareness during information updates, thereby inducing their local perception and collaborative decision-making.

[0042] In step S3, the specific method for selecting and controlling the target restraint node is as follows: Based on the information network model and information intervention mechanism of the target cluster, a method for selecting and controlling the restraining nodes of the target cluster is further designed: by selecting a set of restraining nodes in the information network of the target cluster. This ensures that external information input only affects the set. The corresponding target node, through the coupling effect of the target cluster network, spreads its restraining effect to all nodes, thereby achieving the goal of restraining and controlling the overall situation and behavioral evolution of the target cluster. Considering that the damage and exit of individual nodes in the target cluster, link changes, and topology reconstruction during the countermeasure process will cause the network to change instantaneously, an offline design is adopted to fix the number of target nodes and an online update of the target node set to achieve the selection and control of the target restraining node.

[0043] During the offline design phase, the initial information layer network of the target cluster at the initial moment is determined. Number of nodes The maximum degree of each corresponding target node Provide a lower bound estimate for the initial number of restraining nodes. Based on this, a redundancy coefficient is introduced. The fixed number of restraining nodes is set as follows: (11) During the online update phase, a centrality assessment of the target cluster network at the current moment is required, and a fixed set of restraining nodes, determined in the offline design phase, is selected based on its overall utility. For each moment... Target cluster network Calculate candidate target nodes The three centrality indices are degree centrality, betweenness centrality, and eigenvector centrality, and they are normalized to the [0,1] interval.

[0044] The method for calculating degree centrality is as follows: (12) The method for calculating betweenness centrality is as follows: (13) in, For nodes To the node The number of shortest paths, For the nodes The number of shortest paths, This is the normalization constant.

[0045] The method for calculating the eigenvector centrality is as follows: (14) in, It is the right eigenvector corresponding to the largest eigenvalue of the adjacency matrix of the current target cluster network.

[0046] Based on this, the three types of centrality mentioned above are weighted and superimposed to construct a comprehensive node centrality score: (15) in, and This is a weighting coefficient, which can be configured according to the specific task.

[0047] For the current set of restraining nodes With its already covered set of nodes ,initialization , For any candidate node Considering its marginal coverage gain under the current state of the constrained node set coverage, it is: (16) Based on this, define the comprehensive utility function corresponding to the set of restraining nodes at the current moment: (17) Therefore, in each step of dynamically selecting the restraining node, the previously unselected key node is chosen to make it... The largest node is added to the restraining node set, and the restraining node coverage set is updated synchronously until one of the following conditions is met: the restraining node set covers all nodes in the target cluster. Or the number of restraining nodes reaches a fixed number. .

[0048] In summary, the pseudocode for the dynamic key node selection algorithm for the target cluster is as follows:

[0049] Example 1: Step 1: Countermeasure Task Initialization and Target Cluster Information Interaction Network Modeling 1) Input parameters: Receives the countermeasure task input, including the set of observable nodes of the target cluster and their identifiers. The current position of each node ,speed Status information such as heading.

[0050] 2) Initial network modeling: Construct the information interaction network topology diagram of the target cluster based on the interaction rules. Adjacency Matrix Calculate each target node degree and maximum Define each target node Information layer state .

[0051] 3) Intervention Channel Configuration: Define external information input items for target node information intervention. .

[0052] Step Two: Setting the Scale of Key Control Nodes and Selecting Target Nodes 1) Fixed constraint scale setting: Set redundancy coefficient based on task resource constraints. Fixed size with restraining nodes Design centrality weights according to specific task requirements. .

[0053] 2) Centrality metric calculation: For the current network Each candidate node Calculate degree centrality, betweenness centrality, and eigenvector centrality, and then normalize them.

[0054] 3) Construction of comprehensive score: Construct a comprehensive centrality score according to preset weights. .

[0055] 4) Selection of restraint nodes: Initialize the set of restraint nodes Covering set Calculate the marginal coverage gain for unselected nodes. And calculate the overall utility. In each iteration, select to make The largest node added and update the covered set. until all nodes are covered or .

[0056] Step 3: Randomly filtering out real information and injecting deceptive information in a targeted manner. 1) Parameter initialization: for each restraint node Set random blocking probability With directional injection intensity gain Directional offset With polar radius constant .

[0057] 2) Randomly masking real information: for restraining nodes Every real neighbor Introducing link Bernoulli gated variables .when When the neighbor's input is blocked; The neighbor input is retained at that time. This yields the masked set of valid real neighbors. and the number of effective real neighbors .

[0058] 3) Targeted injection of false information: Calculating the control node The number of equivalent real information sources Based on injection intensity gain Calculate the injection scale of discrete spurious nodes For the first The polar angle of the interaction potential is calculated using virtual information sources. .

[0059] 4) Equivalent information intervention input: For each restraining node The system integrates the real neighbor input retained after random masking with the deceptive input generated by targeted injection to obtain equivalent information input. The fusion method can be vector superposition, weighted superposition, or equivalent mapping, ensuring that equivalent inputs can be directly used as input items for the information layer state update relationship.

[0060] Step 4: Execution of the closed-loop control mechanism guided by countermeasure information 1) Constraint control execution: This involves using equivalent inputs... During the information state update process applied to the restraining node, the state of the individual is recorded at each time step. Record intermediate results such as the target cluster network structure.

[0061] 2) Termination Condition Judgment: Within each control cycle, the task completion conditions are determined based on observable cluster coordination indicators or situational response indicators; if not met, the control parameters for information intervention are updated. , or Then return to step two and continue the loop.

[0062] Output results: Output the entire trajectory, speed, and multi-layer network topology of the target cluster over time, and record the configuration of the control parameters used for information intervention in this operation.

[0063] This invention also provides a containment control system for aircraft swarms to counter information guidance in a "group-against-group" manner, including a parameter input and initialization module, a target swarm information interaction network construction module, a containment key node selection module, a real information intervention module based on random shielding, a deception information generation module based on targeted injection, and a containment control module for counter information guidance. The information interaction network construction module constructs the topology of the target cluster based on the interaction rules and establishes a network coupling evolution model of node information states; The module for selecting key control nodes determines the set of control nodes by comprehensively considering node centrality and coverage gain under a fixed control scale constraint. The random shielding module uses Bernoulli gating to probabilistically block the input of real neighbor information of the restraining node; the directional injection module generates discrete injection scale by mapping continuous injection gain and rounding up, and generates directional deception information by using fixed polar radius and adjustable polar angle offset. The countermeasure information-guided control module merges the shielded real residual information with the injected deceptive information to form an equivalent input, which is then applied to the information state update process of the control node. This allows local cognitive biases to propagate through network coupling, thereby achieving control over the overall behavioral evolution of the target cluster.

[0064] Furthermore, the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of the present invention, and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Additionally, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0065] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention herein. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention 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 the invention are indicated by the claims.

[0066] It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is defined only by the appended claims.

Claims

1. A method for containment control of aircraft swarms guided by swarm counter-swarm information, characterized in that, The method includes: Under discrete time steps, obtain the node set of the target unmanned cluster and the information interaction relationship between the nodes, and construct the information interaction network of the target cluster; Based on the information exchange network, a set of restraining nodes is selected from the target cluster, and the following external information inputs are applied to the restraining nodes: For each constrained node, a random blocking mechanism is used to probabilistically block information input from its real neighbors, so as to form a blocked real information input; For each restraint node, a targeted injection mechanism is used to generate deception information, where the injection strength of the deception information is adjusted by a continuously controlled gain parameter, and the executable injection scale is obtained through discrete mapping. By fusing the masked real information input with the injected deceptive information, the equivalent information input of the restraining node is obtained. This information is then used to influence the information state of the non-restraining node through the coupling propagation of the information interaction network, thereby achieving restraint and control over the overall behavioral evolution of the target unmanned cluster. The information interaction network, the set of restraining nodes, and the equivalent information input are iteratively updated at discrete time steps to output the state sequence and restraining result in the countermeasure process.

2. The method according to claim 1, characterized in that, The random masking mechanism is Bernoulli gated masking: a communication state random variable is introduced into the information link between the restraining node and each of its real neighbors, and it follows a Bernoulli distribution; when the communication state is zero, the corresponding neighbor information input is masked, and otherwise the corresponding neighbor information input is retained.

3. The method according to claim 1, characterized in that, In the targeted injection mechanism, the continuous control gain is mapped to the injection scale of discrete spurious nodes through the floor operator.

4. The method according to claim 1, characterized in that, The directional injection mechanism generates the injection direction using a directional parameterization method with a fixed polar radius and an adjustable polar angle: the injection radius is set to a fixed constant value, and the injection direction is controlled by the polar angle offset parameter.

5. The method according to claim 1, characterized in that, The selection of the set of restraining nodes includes: The degree centrality, betweenness centrality, and eigenvector centrality of candidate nodes are calculated on the current information exchange network, and then fused according to weights to obtain a comprehensive score. Based on the comprehensive score and coverage gain, a comprehensive utility function is constructed. Nodes that maximize utility and improve network coverage are selected in turn and added to the set of restraining nodes until the preset number of restraining nodes is reached or all nodes are covered.

6. The method according to claim 1, characterized in that, The number of restraining nodes is fixed. During the countermeasure process, when the target cluster nodes exit or the number of nodes is less than the fixed size due to topology reconstruction, the restraining node index sequence is zero-filled or repeatedly mapped to maintain the consistency of the external information input channel dimension.

7. A containment control system for aircraft swarms guided by swarm counter-countermeasure information, characterized in that, include: The information interaction network construction module is used to construct a network coupling evolution model of the information interaction network topology and node information status based on the interaction rules between target cluster nodes. The module for selecting key control nodes is used to determine the set of control nodes by combining node centrality and coverage gain under a fixed control size constraint. The random blocking module is used to probabilistically block the input of real neighbor information of the restraining node in a Bernoulli gating manner; The directional injection module is used to generate discrete injection scales by mapping continuous injection gain to the floor function, and to generate directional deception information using a fixed polar radius and an adjustable polar angle bias. The countermeasure information-guided restraint control module is used to merge the shielded real residual information with the injected deceptive information to form an equivalent input, and to act on the information state update process of the restraint node, so that local cognitive biases are propagated through network coupling, thereby achieving restraint control over the overall behavioral evolution of the target cluster.

8. The system according to claim 7, characterized in that, The key node selection module also includes a dynamic update unit, which is used to recalculate the centrality score and coverage gain based on the current network structure and update the set of key nodes when the target cluster topology changes.

9. The system according to claim 7, characterized in that, The random shielding module and the directional injection module operate on the same set of restraining nodes, and support setting different shielding probabilities and injection direction biases for different types of interactive links.

10. The system according to claim 7, characterized in that, The system is deployed in a central processing unit or a distributed processing unit and can access target cluster situational parameters, interactive network data and node information status data, and output intervention strategies and the updated results of the restraint node set.

Images