Abnormal node identification method based on unmanned aerial vehicle swarm cooperative topology operation constraint verification

Abstract

This invention discloses an abnormal node identification method based on the verification of cooperative topology operation constraints in UAV swarms. The invention acquires the communication link status and location information during the cooperative flight of a UAV swarm in real time, constructs a cooperative topology graph describing the communication connections and spatial cooperative relationships between nodes, and verifies the operational status of each node based on a preset cooperative topology operation constraint model. Nodes that do not meet the constraints are identified as abnormal nodes. This method dynamically analyzes node behavior from the overall dimension of the swarm's cooperative operation status. Even with valid node authentication and normal communication link encryption, it can promptly detect abnormal behaviors that violate cooperative rules within the cooperative topology structure, achieving accurate identification of abnormal nodes without relying on known attack characteristics, thereby improving the stability and security of UAV swarm cooperative flight.

Description

Technical Field

[0001] This invention relates to the technical field of unmanned aerial vehicle (UAV) control, and specifically to an abnormal node identification method based on UAV swarm cooperative topology operation constraint verification. Background Technology

[0002] With the widespread application of drones in scenarios such as inspection, surveying, emergency rescue, and military reconnaissance, multi-drone coordinated formation flight has become the mainstream operational mode. Drone swarms typically establish cooperative communication links through wireless ad hoc networks and complete group coordinated control based on the position information and communication status between nodes. However, existing drone swarm control systems still mainly rely on single-node authentication or communication encryption mechanisms for security protection, lacking the ability to perform overall security analysis on the level of group coordinated behavior.

[0003] In real-world operating environments, if the flight control firmware of a node is tampered with, the communication protocol is hijacked, or abnormal data is injected into the navigation information, the operating state of that node in terms of communication topology and spatial coordination may deviate from normal coordination constraints. However, existing security mechanisms often fail to detect such anomalies in a timely manner, leading to the continuous participation of attacking nodes in group control decision-making, which in turn can cause serious consequences such as formation instability, mission failure, or even mass crashes.

[0004] Furthermore, existing drone security protection solutions are mostly based on static rules or single-dimensional indicators for detection, failing to conduct joint analysis of multi-dimensional characteristics such as drone swarm topology, communication relationships, and positional coordination at the level of collaborative operation constraints. This makes it impossible to dynamically assess the overall security of the swarm's collaborative state and thus difficult to meet the security requirements of drone swarm control systems in complex adversarial environments.

[0005] In summary, in the context of collaborative flight of UAV swarms, security mechanisms based on single-node authentication or communication link encryption cannot dynamically analyze abnormal behavior of nodes from the overall perspective of the collaborative operation status of the swarm. This makes it difficult to detect abnormal behavior of nodes that violate the overall collaborative rules in the collaborative topology in a timely manner. Summary of the Invention

[0006] The technical problem this invention aims to solve is that, in related technologies, in UAV swarm cooperative flight scenarios, security mechanisms based on single-node authentication or communication link encryption cannot dynamically analyze abnormal node behavior from the overall perspective of the swarm's cooperative operation state. This makes it difficult to detect abnormal behavior by nodes that violate overall cooperative rules in a timely manner. The purpose is to provide an abnormal node identification method based on UAV swarm cooperative topology operation constraint verification. This solves the technical problem of the difficulty in timely detection of abnormal behavior by nodes that violate overall cooperative rules in a cooperative topology.

[0007] This invention is achieved through the following technical solution:

[0008] In a first aspect, the present invention provides an abnormal node identification method based on UAV swarm cooperative topology operation constraint verification, the method comprising:

[0009] Acquire the communication link status and location information reported by each drone node during the coordinated flight of the drone swarm;

[0010] Based on the communication link status and the location information, a collaborative topology diagram is constructed to characterize the communication connection relationship and spatial collaboration relationship between the nodes of the UAV swarm;

[0011] Based on the cooperative topology graph and the preset cooperative topology operation constraint model, it is verified whether the operation state of each UAV node meets the corresponding cooperative topology operation constraints; wherein, the cooperative topology operation constraint model is used to define the constraint conditions that the node connection relationship and spatial cooperation relationship in the cooperative topology graph should meet under normal cooperative flight state;

[0012] Drone nodes that do not meet the cooperative topology operation constraints will be identified as abnormal nodes.

[0013] Furthermore, the step of obtaining the communication link status and location information reported by each UAV node during the cooperative flight of the UAV swarm includes:

[0014] Receive communication link status and location information reported by each UAV node;

[0015] The communication link status includes at least one of the following: the link connectivity matrix collected by each UAV node through the airborne communication module, the rate of change of communication signal strength over time, the round-trip delay statistics between nodes, and the packet loss rate and retransmission count; the location information includes at least one of the following: the three-dimensional spatial coordinates collected by each UAV node through the airborne positioning module, the relative distance matrix between nodes, and the relative azimuth angle between nodes and its rate of change.

[0016] Furthermore, the step of constructing a cooperative topology map representing the communication connection relationship and spatial cooperation relationship between unmanned aerial vehicle (UAV) swarm nodes based on the communication link status and the location information includes:

[0017] A communication topology subgraph reflecting the communication connection relationship between UAV nodes is constructed based on the communication link status;

[0018] Construct a location collaboration subgraph based on location information to reflect the spatial collaboration relationships between UAV nodes;

[0019] By jointly modeling the communication topology subgraph and the location collaboration subgraph under a unified topology representation framework, a collaborative topology graph representing the communication connection relationship and spatial collaboration relationship between nodes in an unmanned aerial vehicle swarm is obtained.

[0020] Furthermore, the cooperative topology operation constraint model includes a static baseline constraint sub-model and a scenario dynamic adaptation sub-model;

[0021] The static baseline constraint sub-model is a constraint sub-model established based on historical normal cooperative flight data; wherein, the static baseline constraint sub-model includes at least one of the following: allowable range of node degree value change, allowable range of relative distance change between nodes, and allowable range of node communication status change;

[0022] The scenario dynamic adaptation sub-model is a sub-model with a pre-set power inspection scenario feature tag library; wherein, the power inspection scenario feature tag library includes at least one of terrain feature tags, electromagnetic environment feature tags, and formation operation feature tags; wherein, the scenario dynamic adaptation sub-model is used to dynamically adjust the constraint threshold in the static benchmark constraint sub-model according to the real-time identified current operation scenario feature tags.

[0023] Furthermore, the step of verifying whether the operating state of each UAV node satisfies the corresponding cooperative topology operating constraints based on the cooperative topology graph and the preset cooperative topology operating constraint model includes:

[0024] Identify the operational scenario feature tags of the current flight of the drone swarm;

[0025] Based on the identified operation scenario feature labels, the scenario dynamic adaptation sub-model is invoked to dynamically adjust the constraint thresholds in the static baseline constraint sub-model to obtain the currently applicable collaborative topology operation constraints.

[0026] For each UAV node, extract the real-time topological features of that UAV node in the collaborative topology graph; wherein, the real-time topological features include at least one of the following: node degree value, relative distance to neighboring nodes, and communication status;

[0027] The real-time topology features are compared with the currently applicable cooperative topology operation constraints;

[0028] If the real-time topology characteristics exceed the range defined by the currently applicable cooperative topology operation constraints, then the operating state of the UAV node is determined to not meet the corresponding cooperative topology operation constraints.

[0029] Furthermore, the step of identifying drone nodes that do not meet the cooperative topology operation constraints as abnormal nodes includes:

[0030] Mark drone nodes that do not meet the cooperative topology operation constraints as suspected abnormal nodes;

[0031] Obtain the operational status information of each UAV node in the neighboring node set of the suspected abnormal node; wherein, the neighboring node set includes at least one UAV node that has a direct or indirect communication connection with the suspected abnormal node;

[0032] Based on the running status information of the adjacent node set and the characteristics of the current work scenario, verify whether the suspected abnormal node is a true abnormal node.

[0033] If the verification result is a true anomaly, then the suspected anomaly node is determined to be an anomaly node.

[0034] Furthermore, the step of verifying whether the suspected abnormal node is a true abnormal node based on the running status information of the adjacent node set and the characteristics of the current work scenario includes:

[0035] If the current work scenario feature label is identified to include a high-voltage electromagnetic environment feature label, then the high-voltage electromagnetic interference scenario verification sub-process is executed. This high-voltage electromagnetic interference scenario verification sub-process includes: determining whether the communication status of the suspected abnormal node exceeds the range defined by the currently applicable cooperative topology operation constraints; if it does, determining whether there are any neighboring nodes in the neighboring node set whose communication status fluctuates synchronously; if so, determining that the deviation in the communication status of the suspected abnormal node is caused by high-voltage electromagnetic environment interference, and the verification result is a false anomaly; if not, the verification result is a true anomaly.

[0036] If the current operation scenario feature label is identified as including the pole-and-tower formation operation feature label, then the formation maneuver scenario verification sub-process is executed. This sub-process includes: determining whether the relative distance between the suspected abnormal node and its adjacent nodes exceeds the range defined by the currently applicable cooperative topology operation constraints; if it does, determining whether the adjacent node set has undergone a consistent positional offset along the spatial axis direction associated with the formation maneuver command; wherein the spatial axis direction is determined based on the maneuver type and maneuver parameters included in the formation maneuver command; if consistent, determining that the positional deviation of the suspected abnormal node is caused by the overall formation maneuver, and the verification result is a false anomaly; if inconsistent, the verification result is a true anomaly.

[0037] Furthermore, the step of determining whether there are more than a preset proportion of neighboring nodes in the neighboring node set whose communication states experience synchronous fluctuations includes:

[0038] Determine the formation type of the current power line inspection operation of the drone swarm; wherein, the formation type includes linear formation along the power line and planar formation around the tower;

[0039] Based on the formation arrangement type and the interference coverage of the high-voltage electromagnetic environment, the judgment ratio of communication state synchronization fluctuation is dynamically determined as a preset ratio.

[0040] From the communication status of the suspected abnormal nodes and their neighboring nodes, indicators strongly correlated with the high-voltage electromagnetic environment are extracted as judgment indicators; wherein, the judgment indicators include communication signal strength and packet loss rate;

[0041] Based on the judgment index, determine whether the number of nodes in the adjacent node cluster whose judgment index fluctuation period matches the high-voltage transmission frequency and whose fluctuation trend is consistent with the suspected abnormal node reaches a dynamically determined judgment ratio.

[0042] If yes, it is determined that the communication status of more than a preset proportion of neighboring nodes has experienced synchronous fluctuations; if no, it is determined that none exist.

[0043] Secondly, the present invention provides a safety control method based on the cooperative flight of unmanned aerial vehicle swarms, the method comprising:

[0044] Identify abnormal nodes; among them, an abnormal node identification method based on the above-mentioned UAV swarm cooperative topology operation constraint verification is used to identify abnormal nodes;

[0045] A preset security control strategy is executed for the abnormal node to limit its impact on the coordinated flight of the drone swarm.

[0046] Thirdly, the present invention provides an electronic device, comprising: a memory, and one or more processors communicatively connected to the memory; the memory stores instructions executable by the one or more processors, the instructions being executed by the one or more processors to enable the one or more processors to implement the above-described method for identifying abnormal nodes based on UAV swarm cooperative topology operation constraint verification.

[0047] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0048] This invention acquires real-time communication link status and location information during the cooperative flight of a UAV swarm, constructs a cooperative topology graph to describe the communication connections and spatial cooperative relationships between nodes, and verifies the operational status of each node based on a preset cooperative topology operation constraint model, identifying nodes that do not meet the constraints as abnormal nodes. This method dynamically analyzes node behavior from the overall perspective of the swarm's cooperative operation status. Even with valid node authentication and normal communication link encryption, it can promptly detect abnormal behaviors that violate cooperative rules within the cooperative topology structure, achieving accurate identification of abnormal nodes without relying on known attack characteristics, thereby improving the stability and security of UAV swarm cooperative flight. Attached Figure Description

[0049] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort. In the drawings:

[0050] Figure 1 This is one of the flowcharts for an anomaly node identification method based on UAV swarm cooperative topology operation constraint verification provided in the embodiments of this specification;

[0051] Figure 2 This is an architecture diagram of an abnormal node identification method based on UAV swarm cooperative topology operation constraint verification provided in the embodiments of this specification;

[0052] Figure 3 The second flowchart illustrates an abnormal node identification method based on UAV swarm cooperative topology operation constraint verification, as provided in the embodiments of this specification.

[0053] Figure 4 This is a block diagram of an electronic device provided in the embodiments of this specification. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0055] In related technologies, a drone swarm cooperative flight system may include: several drone nodes and a swarm control center. Each drone node may be equipped with an onboard flight control module, a communication module and a positioning module. The swarm control center and each drone node, as well as the drone nodes themselves, establish communication connections through a wireless ad hoc network to form a cooperative communication link for the drone swarm.

[0056] In the field of security protection for drone swarm collaborative flight, the protection mechanisms of relevant technologies are all implemented with individual drone nodes as the core. The protection methods mainly include two categories: single-node authentication and communication link encryption. Single-node authentication verifies the identity information of a drone node before it joins the network and participates in swarm collaborative communication, allowing only legitimate nodes to join the drone swarm's collaborative system. Communication link encryption encrypts the communication data between drone nodes and the swarm control center, as well as between drone nodes themselves, to prevent the theft or tampering of communication data. Some related technologies can also perform single-dimensional anomaly detection on the flight control and positioning data of individual drone nodes, but the detection object is still the independent operational data of a single node, failing to correlate the node's operational status with the overall collaborative operational status of the drone swarm.

[0057] The aforementioned technologies have significant technical limitations in practical applications. They can only verify the security of individual drone nodes and cannot analyze whether the behavior of nodes within the collaborative topology conforms to the rules of collaborative operation within the swarm as a whole. During actual drone swarm operation, drone nodes may exhibit abnormal behavior that disrupts the collaborative operation rules, even if their identity information is legitimate and communication links are properly encrypted. This is manifested in deviations from normal communication connections and spatial positioning relationships within the collaborative topology. Because these technologies lack the ability to monitor and verify dynamic changes in the collaborative topology, they cannot identify such abnormal behaviors that disrupt the overall collaborative rules. As a result, abnormal nodes remain undetected and continue to participate in the collaborative control and task execution of the drone swarm, thereby disrupting the overall collaborative topology, leading to swarm instability, mission failure, and even the entire swarm crashing.

[0058] The core technical problem this invention aims to solve is as follows: In related technologies, in the scenario of UAV swarm cooperative flight, security protection mechanisms based on single-node authentication or communication link encryption cannot dynamically analyze abnormal node behavior from the overall dimension of the swarm's cooperative operation state. This makes it difficult to detect abnormal behavior by nodes that violate the overall cooperative rules in a timely manner. The core reason for this technical problem is that the protection logic of related technologies always revolves around the independent operating state of individual UAV nodes, without establishing a correlation between the operating state of UAV nodes and the swarm cooperative topology. It lacks a dynamic analysis and verification system for the swarm cooperative topology, ignores the fact that the core of UAV swarm cooperative flight is the stability of the swarm cooperative topology, and fails to use whether a node conforms to the swarm cooperative operation rules as a criterion for determining the node's security status. Therefore, even if a node's own identity and communication are normal, its behavior that violates the swarm cooperative rules cannot be identified.

[0059] like Figure 1 , Figure 2 and Figure 3 As shown, this embodiment provides an abnormal node identification method based on UAV swarm collaborative topology operation constraint verification. The execution subject of the method can be the UAV swarm ground control center, which can be an industrial computer, server or dedicated control terminal deployed on the ground. When the ground control center is the execution subject, each UAV node reports the collected communication link status and location information to the center in real time through the airborne communication module. The center uniformly completes the construction of the collaborative topology map, the verification of the node operation status based on the collaborative topology operation constraint model, and the identification and judgment of abnormal nodes. The identification results can be directly sent from the center to each UAV node for subsequent security strategy execution.

[0060] The subject of the method can also be the lead drone in a drone swarm or a cloud server, etc.

[0061] The method may include:

[0062] Step S12: Obtain the communication link status and location information reported by each UAV node during the collaborative flight of the UAV swarm.

[0063] In this embodiment, the drone swarm can be a collaborative flight group composed of at least two drone nodes with autonomous flight, wireless communication and swarm coordination capabilities. Each node can interact with each other's data and complete collaborative operations in accordance with a unified formation control strategy.

[0064] In this embodiment, the UAV node is a single UAV carrier that constitutes a UAV swarm. Each UAV node can be equipped with an onboard flight control module, an onboard communication module, and an onboard positioning module, and has basic functions such as flight data acquisition, wireless communication, autonomous flight status control, and swarm collaborative data interaction.

[0065] In this embodiment, the communication link status can be the operational characteristic parameters of the wireless communication links between UAV nodes and between UAV nodes and the execution entity.

[0066] In this embodiment, the location information can be a set of parameters representing the spatial location characteristics of the UAV node in three-dimensional space and the relative spatial relationship characteristics between nodes.

[0067] In one possible and specific implementation scheme, the implementing entity can receive communication link status and location information actively reported by each UAV node during the coordinated flight of the UAV swarm via wireless communication. Each UAV node collects data and reports it according to a preset sampling period. The sampling period can be flexibly set according to the coordinated operation requirements of the UAV swarm, and the sampling periods of each UAV node can be kept synchronized. Specifically, the communication link status of the UAV nodes can be collected by the onboard communication module. The collected communication link status can include one or more of the following: link connectivity matrix, rate of change of communication signal strength over time, round-trip delay statistics between nodes, packet loss rate, and number of retransmissions. The link connectivity matrix is ​​used to characterize whether there is a valid wireless communication connection between any two UAV nodes. The rate of change of communication signal strength over time is used to characterize the stability of the communication signal between nodes. The round-trip delay statistics are used to characterize the delay characteristics of data transmission between nodes. The packet loss rate and number of retransmissions are used to characterize the effectiveness of data transmission between nodes.

[0068] The location information of UAV nodes can be collected by the onboard positioning module. The collected location information may include one or more of the following: the three-dimensional spatial coordinates of each UAV node, the relative distance matrix between nodes, and the relative azimuth angles and their rate of change between nodes. The three-dimensional spatial coordinates represent the absolute position of a single UAV node in three-dimensional space. The relative distance matrix represents the straight-line distance between any two UAV nodes. The relative azimuth angles and their rate of change represent the relative spatial orientation and orientation change characteristics between any two UAV nodes.

[0069] After collecting communication link status and location information, each UAV node encapsulates the collected data and reports it to the method execution entity via a wireless ad hoc network.

[0070] Step S14: Based on the communication link status and the location information, construct a collaborative topology map that represents the communication connection relationship and spatial collaboration relationship between the nodes of the UAV swarm.

[0071] In this embodiment, the construction action can be represented as using the acquired communication link status and location information to generate a data structure that can comprehensively describe the communication connection and spatial cooperation relationship between nodes, namely a cooperative topology graph.

[0072] Step S16: Based on the cooperative topology graph and the preset cooperative topology operation constraint model, verify whether the operation status of each UAV node meets the corresponding cooperative topology operation constraints; wherein, the cooperative topology operation constraint model is used to define the constraint conditions that the node connection relationship and spatial cooperation relationship in the cooperative topology graph should meet under normal cooperative flight state.

[0073] In this embodiment, the executing entity can match the constructed collaborative topology graph with the pre-configured collaborative topology operation constraint model, verify the operation status of each drone node in the drone swarm one by one, and determine whether the operation status of a single node meets the compliance conditions defined by the collaborative topology operation constraint model.

[0074] In one possible and specific implementation, firstly, the executing entity retrieves a pre-configured cooperative topology operation constraint model. This model is constructed based on historical operational data of normal cooperative flight of the UAV swarm. The model predefines the constraints that the node connection relationships and spatial cooperative relationships in the cooperative topology graph of the UAV swarm should satisfy under normal cooperative flight conditions. Specifically, this may include the allowable range of node degree value changes, the constraint range of connectivity relationships between nodes, and cooperative constraints on relative position changes between nodes. Next, for each UAV node in the cooperative topology graph, the executing entity extracts the real-time topology features of that node as its operational state. These real-time topology features may include one or more of the following: node degree value, relative distance to neighboring nodes, and communication state. The node degree value is the number of neighboring nodes with which the node has a valid communication connection in the communication topology subgraph. The relative distance to neighboring nodes is the spatial straight-line distance between the node and each neighboring node in the positional cooperative subgraph. The communication state includes operational characteristics such as signal strength, latency, and packet loss rate of the communication links between the node and its neighboring nodes.

[0075] Then, the executing entity can compare the extracted real-time topology features of a single UAV node with the corresponding predefined constraints in the cooperative topology operation constraint model one by one to determine whether the node's real-time topology features are within the compliance range defined by the model. If all real-time topology features of a single UAV node are within the compliance range defined by the cooperative topology operation constraint model, then the UAV node's operating state is determined to satisfy the corresponding cooperative topology operation constraint. If at least one real-time topology feature of a single UAV node exceeds the compliance range defined by the cooperative topology operation constraint model, then the UAV node's operating state is determined to not satisfy the corresponding cooperative topology operation constraint.

[0076] Step S18: Identify drone nodes that do not meet the cooperative topology operation constraints as abnormal nodes.

[0077] In this implementation, when a node is determined to be in an abnormal operating state, the executing entity can identify it as an abnormal node and record the time of the abnormality, the abnormal topological characteristics (specific constraints violated, degree of deviation, etc.), and related environmental data. The identification results can be used to trigger subsequent security control strategies, such as sending alarms to the ground station, reducing the node's weight in collaborative control, restricting its communication permissions, or logically isolating it from the collaborative topology.

[0078] This embodiment acquires the communication link status and location information during the cooperative flight of a UAV swarm in real time, constructs a cooperative topology graph to describe the communication connection and spatial cooperation relationships between nodes, and verifies the operational status of each node based on a preset cooperative topology operation constraint model, identifying nodes that do not meet the constraints as abnormal nodes. This method dynamically analyzes node behavior from the overall dimension of the swarm's cooperative operation status. Even with valid node authentication and normal communication link encryption, it can promptly detect abnormal behaviors that violate cooperative rules within the cooperative topology structure, achieving accurate identification of abnormal nodes without relying on known attack characteristics, thereby improving the stability and security of UAV swarm cooperative flight.

[0079] In some implementations, the step of obtaining the communication link status and location information reported by each UAV node during the coordinated flight of the UAV swarm includes:

[0080] Step S122: Receive the communication link status and location information reported by each UAV node.

[0081] The communication link status includes at least one of the following: the link connectivity matrix collected by each UAV node through the airborne communication module, the rate of change of communication signal strength over time, the round-trip delay statistics between nodes, and the packet loss rate and retransmission count; the location information includes at least one of the following: the three-dimensional spatial coordinates collected by each UAV node through the airborne positioning module, the relative distance matrix between nodes, and the relative azimuth angle between nodes and its rate of change.

[0082] In this embodiment, the executing entity (e.g., a ground control station, a lead drone, or a cloud server) can acquire data packets actively sent by each UAV node via a wireless communication link. Specifically, a periodic reception mode or an event-triggered reception mode can be used for data reception. Specifically, when a UAV node detects a sudden change in communication quality (a sharp drop in signal strength), a position change exceeding a threshold (deviation from the predetermined flight path), or a change in formation status, it actively triggers a report, and the executing entity receives the event-triggered data.

[0083] In this embodiment, the airborne communication module can be a hardware module integrated into each UAV node, possessing basic functions such as wireless communication and acquisition of communication link characteristic parameters. It can realize wireless data interaction between UAV nodes and between UAV nodes and the execution entity, and can also collect various operational characteristic parameters of the communication link between itself and other nodes in real time. The airborne positioning module can also be a hardware module integrated into each UAV node, possessing basic functions such as spatial position detection and acquisition of spatial relationship characteristic parameters. It can collect the spatial position coordinates of the UAV node itself in real time, and can also calculate the relative spatial relationship characteristic parameters between nodes based on its own position data and that of other nodes.

[0084] In one possible and specific implementation, the link connectivity matrix can be an N×N two-dimensional matrix, where N is the total number of nodes in the UAV swarm. Each element in the matrix indicates whether a valid communication link exists between the node at the corresponding row index and the node at the corresponding column index. The criteria for determining a valid link can be: signal strength higher than a preset threshold (e.g., -90dBm), successful communication protocol handshake, or receiving a heartbeat packet within a certain time window.

[0085] In one possible and specific implementation, the rate of change of the communication signal strength over time is used to measure the dynamic fluctuations in signal strength. The onboard communication module of the UAV node can continuously sample the received signal strength indication (RSSI) and calculate the amount of change in signal strength per unit time (per second), i.e., the rate of change. A large rate of change often indicates that the node is rapidly moving away from or entering an interference area, or that the communication module is malfunctioning.

[0086] In one possible and specific implementation, the round-trip time (RTT) between the nodes is statistically analyzed. RTT is the time elapsed from when a node sends a probe packet to when it receives an acknowledgment from the other node. Nodes can periodically send probe packets to their neighbors, recording and statistically analyzing the RTT values. The statistical values ​​can be averages (reflecting the overall latency level), minimums (reflecting the optimal path latency), maximums (reflecting the worst-case scenario), or percentiles (90th percentile latency).

[0087] In one possible and specific implementation, the packet loss rate is the proportion of data packets sent by a node that are not successfully received within a certain period of time, and can be expressed as a percentage. The retransmission count is the number of times data packets are retransmitted due to loss, and can be a cumulative value or the retransmission frequency per unit time.

[0088] In one possible and specific implementation, the relative distance matrix between the nodes can also be an N×N two-dimensional matrix, where each element can represent the Euclidean distance between the corresponding row index node and the corresponding column index node, and the relative distance can be calculated from the three-dimensional coordinates of the two nodes.

[0089] In one possible and specific implementation, relative azimuth refers to the directional angle of one node relative to another. For example, taking node i as the observation point, the azimuth of node j relative to node i can be the heading angle in the horizontal plane and the pitch angle in the vertical plane. Azimuth can also be calculated based on three-dimensional coordinates. The rate of change is the amount of change in azimuth per unit time, reflecting the relative motion trend between nodes. When the formation performs maneuvers (turning, detouring), the azimuth and its rate of change will change regularly; however, abnormal abrupt changes in azimuth may indicate that a node has deviated from the formation.

[0090] In some implementations, the step of constructing a cooperative topology map representing the communication connection relationship and spatial cooperation relationship between UAV swarm nodes based on the communication link status and the location information includes:

[0091] Step S142: Construct a communication topology subgraph reflecting the communication connection relationship between UAV nodes based on the communication link status.

[0092] In this embodiment, the communication topology subgraph can be represented as a topology structure with UAV nodes as topology nodes and effective communication connections between nodes as topology edges, used to characterize the communication connection relationship between UAV swarm nodes. The topology edges can carry communication link status related attribute information.

[0093] In one possible and specific implementation, the implementing entity can first retrieve the communication link status data reported by each UAV node, which has undergone integrity and format verification. This data may include at least one of the following: link connectivity matrix, rate of change of communication signal strength over time, round-trip delay statistics between nodes, packet loss rate, and number of retransmissions. All data are associated with the corresponding sampling period timestamp and the node's unique identifier.

[0094] Next, the topology nodes of the communication topology subgraph are determined. Specifically, each drone node in the drone swarm can be treated as an independent topology node, and each topology node is assigned a node identifier that is consistent with the unique identifier of the drone node itself, so that there is a one-to-one correspondence between the topology nodes and the actual drone nodes.

[0095] Next, the topological edges of the communication topology subgraph are constructed. The link connectivity matrix can be used as a criterion to identify node pairs in the UAV swarm with valid wireless communication connections. That is, if the elements corresponding to two UAV nodes in the link connectivity matrix indicate a valid communication connection, then a topological edge is constructed between the corresponding topological nodes. If the elements indicate no valid communication connection, then no topological edge is constructed between the corresponding topological nodes.

[0096] Finally, configure attribute information for the topology edges. Based on non-matrix parameters in the communication link state, add corresponding attribute values ​​to each constructed topology edge. These attribute values ​​can be consistent with the communication link state parameters for the corresponding sampling period. Specifically, the rate of change of communication signal strength over time can be used as the signal stability attribute of the topology edge; the mean of the round-trip delay statistics can be used as the transmission delay mean attribute; variance can be used as the transmission delay fluctuation attribute; packet loss rate can be used as the transmission packet loss attribute; and the number of retransmissions can be used as the transmission retransmission attribute. After completing the topology node definition, topology edge construction, and edge attribute configuration, a communication topology subgraph can be formed.

[0097] Step S144: Construct a location collaboration subgraph based on location information that reflects the spatial collaboration relationship between UAV nodes.

[0098] In this embodiment, the location collaboration subgraph can be represented as a topological structure with UAV nodes as topological nodes and the relative spatial relationships between nodes as topological edges, which is used to characterize the spatial collaboration relationships between UAV swarm nodes.

[0099] In one possible and specific implementation, a communication topology can be adopted first. Figure 1 The consistent node identification rules treat each drone node in the drone swarm as an independent topology node, assigning it the same node identifier as the communication topology subgraph, ensuring that the topology nodes of the two subgraphs can be accurately matched.

[0100] Then, the topological edges of the location collaboration subgraph can be constructed based on the relative distance matrix. Topological edges can be constructed for all pairs of UAV nodes in the UAV swarm (the spatial collaboration relationship needs to cover the relative positions between all nodes). That is, a topological edge is constructed between the topological nodes corresponding to any two UAV nodes, so as to comprehensively represent the relative spatial relationship between nodes.

[0101] Next, based on the non-matrix parameters in the location information, corresponding attribute values ​​can be added to each topological edge. These attribute values ​​can be consistent with the location information parameters for the corresponding sampling period. Specifically, the values ​​in the relative distance matrix can be used as the relative distance attribute of the topological edge, the relative azimuth angle as the relative azimuth attribute, and the rate of change of the relative azimuth angle as the azimuth change rate attribute. To reflect the absolute positional characteristics of the nodes, the corresponding 3D spatial coordinates associated with the topological nodes can also be added as node attributes to the location collaboration subgraph. Finally, after completing the topological node definition, topological edge construction, and edge attribute configuration, the location collaboration subgraph is formed.

[0102] Step S146: Jointly model the communication topology subgraph and the location collaboration subgraph under a unified topology representation framework to obtain a collaborative topology graph that represents the communication connection relationship and spatial collaboration relationship between nodes of the UAV swarm.

[0103] In this embodiment, the joint modeling can be represented as merging the communication topology subgraph and the location collaboration subgraph into a single graph structure, so that the graph can simultaneously express the communication connection relationship and spatial collaboration relationship between nodes.

[0104] In one possible and specific implementation, the implementing entity first retrieves the already constructed communication topology subgraph and location coordination subgraph. Since the two subgraphs use the same node identification rules, that is, each topology node corresponds one-to-one with a UAV node and the identification is consistent, the node sets of the two subgraphs are naturally aligned, and there is no need for node matching or remapping.

[0105] Then, the executing entity determines the node set of the collaborative topology graph. Specifically, the node set in the communication topology subgraph (or location collaborative subgraph) is directly used, and each UAV node in the UAV swarm is treated as a collaborative topology node. Each collaborative topology node retains its node identifier in the subgraph and optionally associates the UAV's three-dimensional spatial coordinates as a node attribute.

[0106] Next, the implementing entity constructs the edge set of the collaborative topology graph.

[0107] Specifically, the implementing entity can construct the edge set of the collaborative topology graph using a composite edge structure. More specifically, the edges in the collaborative topology graph inherit edges from both the communication topology subgraph and the location collaborative subgraph. For any pair of nodes, if a communication edge exists in the communication topology subgraph, a communication edge is added to the collaborative topology graph, along with all its attributes (signal stability, average transmission delay, packet loss, etc.). If a spatial collaborative edge exists in the location collaborative subgraph, a spatial collaborative edge is added to the collaborative topology graph, along with all its attributes (relative distance, relative orientation, rate of orientation change). A pair of nodes may have both communication and spatial collaborative edges, or only one type; the two types of edges remain independent but coexist in the collaborative topology graph.

[0108] Specifically, the implementing entity can also construct the edge set of the collaborative topology graph in other ways. More specifically, for any pair of nodes, only one edge is constructed in the collaborative topology graph, but this edge carries two attribute fields: a communication attribute field and a spatial attribute field. If a communication edge exists in the communication topology subgraph, the attributes of the communication edge are filled into the communication attribute field. If no communication edge exists in the communication topology subgraph, the communication attribute field is left empty or filled with a default value (no connection). Similarly, if a spatial collaborative edge exists in the location collaborative subgraph, the attributes of the spatial collaborative edge are filled into the spatial attribute field. If no such edge exists, the spatial attribute field is also filled with a default value. More specifically, first, an N×N adjacency matrix (N is the total number of UAV nodes) is initialized. Each element of the matrix is ​​designed as a structure or dictionary to store the communication and spatial attributes. Then, traverse each communication edge in the communication topology subgraph. For node pair (i, j), write all attributes of the communication edge (signal stability attribute, average transmission delay attribute, transmission delay fluctuation attribute, packet loss attribute, transmission retransmission attribute, etc.) into the communication attribute field of element (i, j) in the adjacency matrix of the collaborative topology graph. Since the edges in the communication topology subgraph may be undirected (communication links are usually bidirectional and peer-to-peer), both elements (i, j) and (j, i) are written simultaneously.

[0109] Next, each spatial collaboration edge in the location collaboration subgraph is traversed. For node pair (i, j), all attributes of the spatial collaboration edge (relative distance, relative orientation, rate of orientation change, etc.) are written into the spatial attribute domain of element (i, j) in the adjacency matrix of the collaboration topology graph. The location collaboration subgraph can construct edges for all node pairs, and the spatial relationship is bidirectional (distance and orientation angles are opposites or complementary angles), so the information is written symmetrically as well. Finally, the three-dimensional spatial coordinates of each node are used as node attributes and associated with the node set of the collaboration topology graph for subsequent verification requiring absolute position information.

[0110] In some implementations, the cooperative topology operation constraint model includes a static baseline constraint sub-model and a scenario dynamic adaptation sub-model;

[0111] The static baseline constraint sub-model is a constraint sub-model established based on historical normal cooperative flight data; wherein, the static baseline constraint sub-model includes at least one of the following: allowable range of node degree value change, allowable range of relative distance change between nodes, and allowable range of node communication status change.

[0112] The scenario dynamic adaptation sub-model is a sub-model with a pre-set power inspection scenario feature tag library; wherein, the power inspection scenario feature tag library includes at least one of terrain feature tags, electromagnetic environment feature tags, and formation operation feature tags; wherein, the scenario dynamic adaptation sub-model is used to dynamically adjust the constraint threshold in the static benchmark constraint sub-model according to the real-time identified current operation scenario feature tags.

[0113] In this embodiment, the static baseline constraint sub-model can be a constraint sub-model established based on historical normal cooperative flight data. Historical normal cooperative flight data can be samples of communication link status and position information collected under conditions of stable UAV swarm formation, no abnormal events, and no external interference. These samples cover various conventional flight states, such as straight-line cruise, constant-speed flight, and small-amplitude attitude adjustments.

[0114] In this embodiment, the process of establishing the static baseline constraint sub-model is as follows: First, a sufficient number of historical normal cooperative flight data samples are collected. Each sample includes a cooperative topology map and its corresponding node features within a complete time window. Then, statistical analysis is performed on the various topological features of each node to obtain the statistical distribution under normal conditions. The statistical analysis can use statistical measures such as mean, variance, and percentiles, or more complex probability distribution fitting methods. Finally, based on the statistical results, allowable ranges for each feature are set as constraints.

[0115] In this embodiment, the node degree value can be the number of neighboring nodes with which the node has a communication connection or spatial cooperative relationship. For a communication topology subgraph, the node degree value can reflect the number of communication connections of the node. For a location cooperative subgraph, the node degree value can reflect the number of spatial adjacencies of the node in the formation.

[0116] In this embodiment, the allowable range of node degree value variation is used to characterize the upper and lower limits of the fluctuation of the node degree value relative to the historical average under normal cooperative flight conditions. For example, if historical statistics show that the degree value of a certain type of node is between 4 and 6, the allowable range can be set to 3 to 7, and anything outside this range is considered abnormal.

[0117] In this embodiment, regarding the allowable range of relative distance changes between nodes, it is understood that for node pairs with spatial cooperative relationships, their relative distance should be within the range specified by the formation strategy, and the changes should not be too rapid. The allowable range of relative distance changes between nodes is used to characterize the reasonable fluctuation range of relative distance under normal conditions. This range can be a fixed range, for example, ±10 meters centered on the desired distance. It can also be a dynamic range, for example, adaptively adjusted according to flight speed or altitude. Exceeding this range may indicate that the node has deviated from its formation position.

[0118] In this embodiment, the allowable range of node communication status changes includes indicators such as signal strength, round-trip time, and packet loss rate. The allowable range of node communication status changes characterizes the reasonable fluctuation range of these indicators under normal conditions. For example, signal strength should be between -70dBm and -50dBm, latency should be between 10ms and 50ms, and packet loss rate should be less than 3%. Exceeding these ranges may indicate that the node is in a communication interference area or that the communication module is malfunctioning.

[0119] In this embodiment, the scene dynamic adaptation sub-model is a sub-model with a pre-built power inspection scene feature tag library. The power inspection scene feature tag library may include scene tags closely related to the power inspection operation environment. These tags are used to identify the specific operation scene in which the UAV swarm is currently located.

[0120] In this embodiment, the power inspection scene feature tag library includes at least one of the following tags:

[0121] Terrain feature tags are used to identify the type of terrain the drone swarm is currently flying over, such as mountainous terrain, plains, and terrain crossing rivers. Understandably, different terrains have varying impacts on communication links and flight attitude; mountainous areas may cause communication blockages, and areas crossing rivers may experience sudden changes in wind speed.

[0122] Electromagnetic environment feature tags are used to identify the type of electromagnetic environment in which the drone swarm is currently located, such as high-voltage line sections, ordinary sections, or near substations. It is understandable that high-voltage line sections experience strong electromagnetic interference, which may affect communication signal quality and positioning accuracy.

[0123] Formation operation feature tags are used to identify the current formation operation mode of the UAV swarm, such as regular patrol, pole-circling flight, and cross-line maneuvering. Understandably, the formation shape and the relative positions of the nodes will change systematically under different operation modes.

[0124] In one possible and specific implementation, during the flight of a drone swarm, the executing entity can identify the characteristic labels of the current operational scene in real time. The identification method can be based on preset flight path information (e.g., known high-voltage line coordinates and tower locations), or on real-time data analysis (e.g., detecting periodic fluctuations in signal strength as electromagnetic interference, or detecting formation shrinkage as tower maneuvering). The identified scene characteristic labels are then passed as input to the scene dynamic adaptation sub-model.

[0125] The scenario-dynamic adaptation sub-model can dynamically adjust the constraint thresholds in the static baseline constraint sub-model based on the identified feature labels of the current work scenario. The adjustment strategy is pre-set in the model and corresponds one-to-one with various scenario labels.

[0126] Specifically, when the current operational scenario is identified to include high-voltage electromagnetic environment characteristics, the executing entity determines that the drone swarm is flying over a high-voltage power line segment. It is understandable that in this scenario, electromagnetic interference may cause increased fluctuations in communication signal strength and a higher packet loss rate. It is also understandable that these changes are normal physical phenomena rather than node anomalies. Therefore, the scenario dynamic adaptation sub-model can appropriately relax the allowable range of communication state changes in the static baseline constraint sub-model. For example, the lower limit of signal strength can be adjusted from -70dBm to -80dBm, and the upper limit of packet loss rate from 3% to 8%. Meanwhile, location-related constraints can remain unchanged or undergo only minor adjustments.

[0127] Specifically, when the current operational scenario feature label includes the pole-and-tower formation operation feature label, the executing entity can determine that the UAV swarm is performing a pole-and-tower maneuver. Understandably, in this scenario, the formation needs to temporarily contract or expand, and the relative distances and relative azimuths between nodes will change systematically. Node degree values ​​may also temporarily change due to adjustments in formation. It is also understandable that these changes are normal operational actions rather than node anomalies. Therefore, the scenario dynamic adaptation sub-model can appropriately relax the allowable range of relative distance changes between nodes and the allowable range of node degree value changes in the static baseline constraint sub-model. For example, the allowable range of relative distance can be expanded from ±10 meters of the desired distance to ±20 meters, and the allowable range of degree values ​​can be expanded from 3 to 7 to 2 to 8. Meanwhile, the constraints related to communication status can remain unchanged.

[0128] Specifically, when the current operational scenario is identified as having plain terrain features and no special electromagnetic environment or formation operation features, the executing entity can determine that the drone swarm is in a normal patrol phase. In this scenario, there is no need to adjust the constraint thresholds; the original constraint conditions of the static baseline constraint sub-model can be directly used to maintain high anomaly detection sensitivity.

[0129] In this embodiment, the scene dynamic adaptation sub-model can adjust the constraint threshold through linear scaling, discrete adjustment based on a preset mapping table, or predictive adjustment based on a machine learning model. The adjustment magnitude and method are preset according to the actual scene requirements and can be optimized and updated based on empirical data during system operation.

[0130] In some implementations, the step of verifying whether the operating state of each UAV node satisfies the corresponding cooperative topology operating constraints based on the cooperative topology graph and a preset cooperative topology operating constraint model includes:

[0131] Step S162: Identify the operational scenario feature tags of the current flight of the drone swarm.

[0132] In this embodiment, before executing the inspection task, the inspection route can be pre-planned and loaded into the UAV swarm or ground control station. Key locations and their corresponding scene attributes are marked on the route; for example, when passing a high-voltage line section, the high-voltage electromagnetic environment is marked; when passing a tower location, tower maneuvering is marked; and when passing through mountainous areas, mountainous terrain is marked. During flight, the executing entity can query the scene attributes of the corresponding location in the route information based on the real-time position of the UAV swarm, thereby identifying the current operational scene feature tags.

[0133] In this embodiment, the identified operation scenario feature label can be a single label or a combination of multiple labels (e.g., simultaneously being in a high-voltage electromagnetic environment and a tower-wrapping operation scenario).

[0134] Step S164: Based on the identified job scenario feature labels, call the scenario dynamic adaptation sub-model to dynamically adjust the constraint thresholds in the static baseline constraint sub-model to obtain the currently applicable collaborative topology operation constraints.

[0135] In this embodiment, the executing entity can use the job scenario feature tags identified in step S162 as an index to query the preset adjustment strategies in the scenario dynamic adaptation sub-model. The scenario dynamic adaptation sub-model can maintain a preset mapping table, which records the constraint threshold adjustment rules corresponding to various scenario feature tags.

[0136] Specifically, the constraint threshold adjustment rule can be a linear scaling rule. More specifically, for a given constraint threshold, a corresponding scaling factor can be defined. For example, in a high-voltage electromagnetic environment, the absolute value of the lower limit of communication signal strength can be increased by 20%, and the upper limit of packet loss rate can be increased by 50%. The executing entity can read the original threshold in the static baseline constraint sub-model, multiply it by the scaling factor, and obtain the adjusted threshold.

[0137] Specifically, the constraint threshold adjustment rules can also be discrete mapping rules. More specifically, a complete set of constraint thresholds can be pre-defined for different scenarios. For example, in a high-voltage electromagnetic environment, the lower limit of signal strength is set to -80dBm and the upper limit of packet loss rate is set to 8%. In a normal environment, the lower limit of signal strength is set to -70dBm and the upper limit of packet loss rate is set to 3%. The executing entity directly loads the corresponding threshold set based on the scenario label.

[0138] In this embodiment, the adjusted constraint threshold, together with the other unadjusted parts of the static baseline constraint sub-model, constitutes the currently applicable cooperative topology operation constraints. It is understood that this constraint is a dynamically changing set, updated in real time as the operational scenario changes.

[0139] Step S166: For each UAV node, extract the real-time topological features of the UAV node in the collaborative topology graph; wherein, the real-time topological features include at least one of the following: node degree value, relative distance to neighboring nodes, and communication status.

[0140] In this embodiment, the executing entity can traverse each node in the collaborative topology graph and extract at least one of the following features from node attributes and edge attributes:

[0141] Node degree, specifically, can be calculated by counting the number of neighboring nodes with which a node has an edge in the collaborative topology graph. Since the collaborative topology graph includes two types of edges (communication edges and spatial collaboration edges), the node degree can be distinguished as a communication degree (based on communication edge statistics) and a spatial degree (based on spatial collaboration edge statistics), or a combined degree (the union of the two types of edges) can be calculated. The node degree reflects the degree of connectivity of the node within the group.

[0142] The relative distance to neighboring nodes can be obtained by extracting the spatial collaborative edge attributes between a node and all its neighbors from the collaborative topology graph, and then obtaining the relative distance attribute value. It can calculate the average, minimum, and maximum distances to all neighbors, and can also focus on the distance to key nodes (leader) in the formation.

[0143] Specifically, regarding communication status, the communication edge attributes between the node and all its neighboring nodes can be extracted from the collaborative topology graph, yielding attributes such as signal stability, average transmission delay, and packet loss. The average value or statistical distribution of each indicator can be calculated.

[0144] In this embodiment, the extracted real-time topology features can be organized in the form of feature vectors, with each node corresponding to a feature vector, including the current values ​​of the aforementioned indicators.

[0145] Step S168: Compare the real-time topology features with the currently applicable cooperative topology operation constraints.

[0146] In this embodiment, for node i, the execution entity can perform the following operations:

[0147] The real-time degree value of node i is compared with the allowable range of node degree value changes in the constraints. If the degree value is lower than the lower limit or higher than the upper limit, it is marked as an abnormal degree value.

[0148] The relative distances between node i and its neighboring nodes are compared with the allowable range of relative distance changes between nodes in the constraints. If a neighbor's distance exceeds the allowable range, it is marked as a distance anomaly.

[0149] The various communication status indicators of node i are compared with the allowable range of node communication status changes in the constraints. If the signal strength is below the lower limit, the latency is above the upper limit, or the packet loss rate is above the upper limit, it is marked as an abnormal communication status.

[0150] Step S1610: If the real-time topology features exceed the range defined by the currently applicable cooperative topology operation constraints, then it is determined that the operating state of the UAV node does not meet the corresponding cooperative topology operation constraints.

[0151] In this implementation, a single-index trigger rule can be used (if any real-time topology feature exceeds the corresponding constraint range, the node's operating state is determined to not meet the cooperative topology operating constraints), or a multi-index comprehensive rule can be used (only when multiple indicators exceed the constraint range simultaneously, or when the deviation of a certain indicator exceeds the severe threshold, is it determined to be unsatisfactory).

[0152] In some implementations, the step of identifying drone nodes that do not meet the cooperative topology operation constraints as abnormal nodes includes:

[0153] Step S182: Mark drone nodes that do not meet the cooperative topology operation constraints as suspected abnormal nodes.

[0154] In this embodiment, the marking action can be represented as the executing entity assigning a temporary state identifier, i.e., a suspected abnormal node, to nodes whose operating state does not meet the cooperative topology operating constraints based on the verification result of step S16. This identifier differs from the finally determined abnormal node; it indicates that the node has triggered the constraint violation condition but has not yet undergone further verification.

[0155] Step S184: Obtain the operating status information of each UAV node in the neighboring node set of the suspected abnormal node; wherein, the neighboring node set includes at least one UAV node that has a direct or indirect communication connection with the suspected abnormal node.

[0156] In this embodiment, the set of adjacent nodes can be a set of drone nodes that have direct or indirect communication connections with the suspected abnormal node. Specifically, the set of adjacent nodes can be a first-order set of adjacent nodes, a second-order set of adjacent nodes, or a dynamic set of adjacent nodes.

[0157] More specifically, the first-order adjacency set includes all neighboring nodes that have a direct communication connection with the suspected anomalous node (i.e., a communication edge exists in the cooperative topology graph). These first-order adjacency nodes are the group of nodes most closely related to the suspected anomalous node, and their operational state is most likely to be affected by the same environmental factors or formation actions. The second-order adjacency set is an extension of the first-order adjacency set, further including nodes that have a direct communication connection with first-order neighboring nodes (i.e., nodes within two hops of the suspected anomalous node).

[0158] In this embodiment, the operational status information of each UAV node in the adjacent node set may specifically include:

[0159] Constraint deviation information (constraint deviation values ​​calculated for each adjacent node, and whether it is marked as a suspected abnormal node), real-time topology features (node ​​degree value, relative distance, communication status, and other real-time topology feature values ​​for each adjacent node), and scene feature information (feature labels of the current work scene for each adjacent node). This operational status information may also include historical status information (the trend of status changes of adjacent nodes over a recent period).

[0160] Step S186: Based on the running status information of the adjacent node set and the characteristics of the current work scenario, verify whether the suspected abnormal node is a true abnormal node.

[0161] In this embodiment, the executing entity can verify whether the suspected abnormal node is a genuine abnormal node based on preset verification rules. Specifically, the verification rules may include: if the abnormal behavior of the suspected abnormal node (communication status degradation, positional deviation) is prevalent in its neighboring node set, that is, multiple neighboring nodes exhibit the same or similar types of abnormalities, and these nodes are in the same operating scenario as the suspected abnormal node, then it is likely to be judged as a false anomaly. This is because environmental factors (electromagnetic interference) or formation maneuvers often affect multiple nodes in an area, rather than a single node. Correspondingly, if the abnormal behavior of the suspected abnormal node only occurs within itself, while the operating status of other nodes in the neighboring node set is normal, and the current operating scenario feature label does not indicate the presence of a special environment or formation maneuver, then it is likely to be judged as a genuine anomaly. This is because the anomaly of an isolated node is more likely caused by its own failure or attack.

[0162] In one possible and specific implementation, step S186 may further include:

[0163] Step S1862: If the current operation scenario feature label is identified to include a high-voltage electromagnetic environment feature label, then execute the high-voltage electromagnetic interference scenario verification sub-process; wherein, the high-voltage electromagnetic interference scenario verification sub-process includes: determining whether the communication status of the suspected abnormal node exceeds the range limited by the currently applicable cooperative topology operation constraints; if it exceeds, then determining whether there are more than a preset proportion of neighboring nodes in the neighboring node set whose communication status has synchronously fluctuated; if so, then determining that the deviation of the communication status of the suspected abnormal node is caused by high-voltage electromagnetic environment interference, and the verification result is a false anomaly; if not, the verification result is a true anomaly.

[0164] Step S1864: If the current operation scenario feature label is identified as including the tower-around formation operation feature label, then execute the formation maneuver scenario verification sub-process; wherein, the formation maneuver scenario verification sub-process includes: determining whether the relative distance between the suspected abnormal node and its adjacent nodes exceeds the range limited by the currently applicable cooperative topology operation constraints; if it exceeds, determining whether the adjacent node set has a consistent positional offset along a specific spatial axis direction that matches the formation maneuver command; if consistent, determining that the positional deviation of the suspected abnormal node is caused by the overall formation maneuver, and the verification result is a false anomaly; if inconsistent, the verification result is a true anomaly.

[0165] In this embodiment, the verification result can be a binary result (true anomaly / false anomaly) or a quantitative result with confidence level.

[0166] Step S188: If the verification result is a true anomaly, then the suspected anomaly node is determined as an anomaly node.

[0167] In some implementations, the step of verifying whether a suspected abnormal node is a genuine abnormal node based on the running status information of the adjacent node set and the characteristics of the current work scenario includes:

[0168] Step S1862: If the current operation scenario feature label is identified to include a high-voltage electromagnetic environment feature label, then execute the high-voltage electromagnetic interference scenario verification sub-process; wherein, the high-voltage electromagnetic interference scenario verification sub-process includes: determining whether the communication status of the suspected abnormal node exceeds the range limited by the currently applicable cooperative topology operation constraints; if it exceeds, then determining whether there are more than a preset proportion of neighboring nodes in the neighboring node set whose communication status has synchronously fluctuated; if so, then determining that the deviation of the communication status of the suspected abnormal node is caused by high-voltage electromagnetic environment interference, and the verification result is a false anomaly; if not, the verification result is a true anomaly.

[0169] In this embodiment, it is understood that when the executing entity identifies that the current operation scenario feature tag includes a high-voltage electromagnetic environment feature tag, it indicates that the drone swarm is flying over a high-voltage line area and may be subject to strong electromagnetic interference. In such scenarios, the degradation of communication status is often regional, affecting multiple neighboring nodes, rather than an isolated phenomenon affecting a single node.

[0170] Therefore, in one possible and specific implementation, the high-voltage electromagnetic interference scenario verification sub-process may specifically include:

[0171] First, determine whether the communication status of the suspected abnormal node exceeds the range defined by the currently applicable cooperative topology operation constraints. Communication status can include indicators sensitive to electromagnetic interference, such as communication signal strength, packet loss rate, and round-trip delay. The executing entity can read the communication status indicator value of the suspected abnormal node at the current moment and compare it with the communication status constraint threshold adjusted by scenario dynamic adaptation in step S164. If the communication status indicator is within the constraint range, it indicates that the node does not exhibit communication abnormalities, and there is no need to proceed to the subsequent synchronization fluctuation judgment; a comprehensive judgment can be made directly in conjunction with other verification rules. If the communication status indicator exceeds the constraint range, proceed to the next step.

[0172] Then, it can be determined whether a preset proportion or more of the neighboring nodes in the neighboring node set experience synchronous fluctuations in their communication states. The determination of synchronous fluctuations can include two dimensions: first, the temporal consistency of the fluctuations, meaning that the communication state indicators of neighboring nodes change in the same direction as the suspected abnormal nodes within a similar time window (signal strength decreases simultaneously, packet loss rate increases simultaneously); second, the correlation of the fluctuation amplitudes, meaning that the amplitudes of the changes are statistically correlated. The preset proportion can be set according to the formation size and scenario characteristics, for example, 50%, 70%, or 80%, etc.

[0173] If a preset proportion or higher of neighboring nodes experience synchronous fluctuations, it can be determined that the communication status deviation of the suspected abnormal node is caused by interference from the high-voltage electromagnetic environment, and the verification result is a false anomaly. It is understandable that the physical characteristic of electromagnetic interference is that it covers a continuous area, causing the communication quality of multiple nodes within that area to deteriorate synchronously. Therefore, if the abnormal behavior of a suspected abnormal node is accompanied by a large number of surrounding nodes exhibiting the same abnormal characteristics, it conforms to the collective pattern of electromagnetic interference and should not be attributed to a fault or attack on a single node.

[0174] Correspondingly, it's understandable that if no more than a predetermined proportion of neighboring nodes experience synchronous fluctuations—that is, only the suspected abnormal node itself or a very small number of nodes experience communication anomalies, while the majority of neighboring nodes maintain normal communication—then the verification result is a genuine anomaly. Therefore, if the anomaly is regional, it should affect multiple nodes, but if only a single node is abnormal, it's more likely due to a failure in the node's own communication module, firmware tampering, or protocol hijacking.

[0175] Step S1864: If the current operation scenario feature label is identified as including the pole-and-tower formation operation feature label, then execute the formation maneuver scenario verification sub-process; wherein, the formation maneuver scenario verification sub-process includes: determining whether the relative distance between the suspected abnormal node and its adjacent nodes exceeds the range defined by the currently applicable cooperative topology operation constraints; if it exceeds, determining whether the adjacent node set has undergone a consistent positional offset along the spatial axis direction associated with the formation maneuver command; wherein, the spatial axis direction is determined according to the maneuver type and maneuver parameters included in the formation maneuver command; if consistent, determining that the positional deviation of the suspected abnormal node is caused by the overall formation maneuver, and the verification result is a false anomaly; if inconsistent, the verification result is a true anomaly.

[0176] In one possible and specific implementation, the formation maneuver scenario verification sub-process may specifically include:

[0177] First, determine whether the relative distance between the suspected abnormal node and its neighboring nodes exceeds the range defined by the currently applicable cooperative topology operation constraints. The executing entity can read the relative distance value between the suspected abnormal node and its neighboring nodes and compare it with the relative distance constraint range adjusted by scene dynamic adaptation in step S164. If the relative distance is within the constraint range, it means that the node does not show any positional abnormality, and there is no need to proceed to the subsequent consistent offset judgment; if the relative distance exceeds the constraint range, proceed to the next step.

[0178] If the number of drones exceeds the limit, the current formation maneuver command issued to the drone swarm is retrieved. The formation maneuver command may include the maneuver type (e.g., turning right around the tower, formation shrinking, crossing the line and lifting, etc.) and maneuver parameters (e.g., turning angle, shrinking ratio, lifting height, etc.).

[0179] Then, the spatial axis direction corresponding to this maneuver can be determined according to the maneuver type and maneuver parameters; for example, for a right turn maneuver around a tower, the spatial axis direction is perpendicular to the tower and points towards the formation rotation center; for a formation shrinking maneuver, the spatial axis direction is the direction pointing towards the geometric center of the formation.

[0180] Finally, the executing entity can extract the real-time position data of each node in the adjacent node set and the position data of the previous sampling period, calculate the position offset vector of each node, and then determine whether the position offset of all adjacent nodes is consistent along the spatial axis direction determined above. More specifically, it can verify whether the position offset direction of each node is in the same direction as the spatial axis direction, and whether the offset magnitude of each node matches the maneuver parameters in the formation maneuver command (for example, in the formation contraction maneuver, the closer the node is to the geometric center of the formation, the smaller the offset magnitude, and the offset magnitude meets the contraction ratio in the command; in the right turn maneuver around the tower, the offset direction of all adjacent nodes is consistent with the direction of the rotation center, and the offset angle meets the turning angle requirement in the command). If all nodes in the adjacent node set meet the above conditions, it is determined that the adjacent node set has undergone a consistent position offset along the spatial axis direction associated with the formation maneuver command, and thus it is determined that the position deviation of the suspected abnormal node is caused by the overall formation maneuver, and the verification result is a false anomaly. If some or all of the adjacent nodes in the cluster have offset directions that are not in the same direction as the spatial axis, or if the offset magnitude does not match the maneuver parameters, it is determined that no consistent positional offset has occurred, and the verification result is a true anomaly.

[0181] Understandably, if the adjacent node set experiences a consistent positional shift along the spatial axis direction that matches the formation maneuver command, then the positional deviation of the suspected anomalous node is determined to be caused by the overall formation maneuver, and the verification result is a false anomaly. This determination is based on the fact that formation maneuvering is a controlled group behavior, and its positional shifts exhibit regularity, consistency, and traceability of commands. If the shift of the suspected anomalous node is coordinated with the shifts of surrounding nodes and conforms to the current maneuver command, then it falls within the scope of normal operation.

[0182] If the adjacent node set does not show a consistent positional offset, meaning the suspected abnormal node's positional deviation direction is inconsistent with that of surrounding nodes, or surrounding nodes show no obvious offset, or the offset pattern does not match the maneuver commands, then the verification result is a genuine anomaly. This judgment is based on the fact that if the formation as a whole does not exhibit consistent maneuvers, but a single node shows a positional deviation, it is more likely due to a malfunction in that node's flight control system, the injection of abnormal data into its navigation information, or hijacking.

[0183] In some implementations, the step of determining whether there are more than a preset proportion of neighboring nodes in the neighboring node set whose communication states experience synchronous fluctuations includes:

[0184] Step S186202: Determine the formation type of the current power line inspection operation of the UAV swarm; wherein, the formation type includes linear formation along the power line and planar formation around the tower.

[0185] In this embodiment, the linear formation along the power line can specifically refer to the use of linear formation when a swarm of drones is conducting patrol inspections along a high-voltage transmission line. This means that each drone node is arranged sequentially along the line direction, forming an approximately straight line. In this formation, the nodes are narrowly distributed perpendicular to the line direction, but have a large span along the line direction.

[0186] In this embodiment, the planar formation around the pole can be used when the drone swarm arrives at the pole location and performs detailed inspections. This planar formation involves each drone node being dispersed around the pole, observing it from different angles. In this formation, the nodes are distributed planarly in the horizontal plane, resulting in a high density.

[0187] In this embodiment, the executing entity can identify whether it is currently in a patrol segment or a tower inspection segment based on the preset inspection route information (the route indicates whether it is currently in a patrol segment or a tower inspection segment), or it can determine the formation arrangement type of the current power inspection operation of the UAV swarm based on real-time location data analysis.

[0188] Step S186204: Based on the formation arrangement type and the interference coverage of the high-voltage electromagnetic environment, dynamically determine the judgment ratio of communication state synchronization fluctuation as the preset ratio.

[0189] In this embodiment, the interference coverage area of ​​the high-voltage electromagnetic environment can be the radius of a spatial region centered on the high-voltage line, where the electromagnetic interference intensity is sufficient to affect the communication status of the UAV. This coverage area can be preset through theoretical calculations (based on voltage level, line type, etc.) or calibrated through field testing.

[0190] In this embodiment, when linear formation along the power line is used, the number of nodes falling within the interference coverage area is small due to the sparse arrangement of nodes along the line. Therefore, the judgment ratio can be set to a low value, for example, 30%-50%, to avoid missing synchronization fluctuations that should be judged due to an excessively high ratio setting.

[0191] In this embodiment, when a planar formation around the tower is used, the nodes are densely distributed around the tower, resulting in a large number of nodes falling within the interference coverage area. Therefore, the judgment ratio can be set to a relatively high value, for example, 70%-90%, to avoid random fluctuations being misjudged as synchronous fluctuations due to an excessively low ratio setting.

[0192] Step S186206: Extract indicators strongly correlated with the high-voltage electromagnetic environment from the communication status of the suspected abnormal node and its neighboring nodes as judgment indicators; wherein, the judgment indicators include communication signal strength and packet loss rate.

[0193] In this embodiment, it is understood that the impact of a high-voltage electromagnetic environment on UAV communication is selective; not all communication indicators will deteriorate to the same degree. In this embodiment, the judgment indicator includes at least one of the following:

[0194] Communication signal strength is a crucial indicator. Electromagnetic interference directly reduces the received signal strength, manifesting as a significant decrease in the RSSI value. Signal strength is one of the most sensitive indicators to interference and is easily measured in real time.

[0195] Packet loss rate, as we understand it, is caused by electromagnetic interference, which leads to an increase in the error rate of data packet transmission, and thus an increase in packet loss rate. Packet loss rate reflects the actual availability of the communication link and is an indicator directly related to interference.

[0196] Step S186208: Based on the judgment index, determine whether the number of nodes in the adjacent node set whose judgment index fluctuation period matches the high-voltage transmission frequency and whose fluctuation trend is consistent with the suspected abnormal node reaches the dynamically determined judgment ratio.

[0197] In this embodiment, the matching of the fluctuation period with the high-voltage transmission frequency can be interpreted as the electromagnetic interference generated by the high-voltage transmission line being strongly correlated with the transmission frequency (power frequency 50Hz or 60Hz) and its harmonics, which may cause communication indicators to exhibit fluctuations with a specific period. Therefore, the executing entity can perform spectral analysis or period detection on the time series of the judgment indicators of suspected abnormal nodes and their adjacent nodes to extract their main fluctuation periods. If the fluctuation period of a node matches the high-voltage transmission frequency (e.g., a 20-millisecond period corresponding to 50Hz) or its harmonics (e.g., allowing an error range of ±5%), then the fluctuation of that node is considered to conform to the time-domain characteristics of electromagnetic interference.

[0198] In this embodiment, the consistency between the fluctuation trend and the suspected abnormal node can be expressed as, based on periodic matching, further determining whether the phase and amplitude change trends of the fluctuation are consistent with those of the suspected abnormal node. For example, whether the signal strength decreases synchronously within the same time window, and whether the packet loss rate increases synchronously. Consistency can be measured by calculating the correlation coefficient or dynamic time warping distance between the two time series; if it exceeds a preset threshold, it is considered consistent.

[0199] In this embodiment, the executing entity can count the number of nodes in the adjacent node set that simultaneously meet the above two conditions, and record it as the number of synchronous fluctuation nodes. This number can be compared with the judgment ratio dynamically determined in step S186204 to determine whether the ratio has been reached (i.e., whether the number of synchronous fluctuation nodes is greater than or equal to the total number of adjacent nodes multiplied by the judgment ratio).

[0200] Step S1862010: If yes, it is determined that the communication status of adjacent nodes exceeding the preset proportion has experienced synchronous fluctuations; if no, it is determined that no such fluctuations have occurred.

[0201] This embodiment provides a safety control method based on cooperative flight of unmanned aerial vehicle swarms, the method comprising:

[0202] Step S22: Identify abnormal nodes; wherein, an abnormal node identification method based on the above-described UAV swarm cooperative topology operation constraint verification is used to identify abnormal nodes.

[0203] Step S24: Execute a preset security control strategy for the abnormal node to limit its impact on the coordinated flight of the UAV swarm.

[0204] In this embodiment, the preset safety control strategy may include dynamically reducing the weight of abnormal nodes in swarm collaborative control. Specifically, in the UAV swarm collaborative control algorithm, the state data of each node can be weighted and fused to generate swarm control commands. Therefore, when a node is identified as an abnormal node, the executing entity can dynamically reduce the weight coefficient of that node in data fusion, reducing its impact on swarm decision-making. For example, the weight of normal nodes can be set to 1.0, and the weight of abnormal nodes can be gradually reduced to 0.5, 0.2, or even 0.

[0205] In this embodiment, the preset security control strategy may also include: limiting the communication frequency or communication range of abnormal nodes. Specifically, abnormal nodes may send a large amount of invalid data or malicious instructions due to attacks, occupying communication resources. The executing entity can issue instructions to abnormal nodes to limit the transmission frequency of their communication modules (e.g., reduce it from 10Hz to 1Hz) or limit their communication range (e.g., reduce the transmission power to shorten the communication distance).

[0206] In this embodiment, the preset security control strategy may further include: for abnormal nodes confirmed to pose serious security risks (flight control firmware tampered with, executing malicious commands), the executing entity can logically isolate them from the cooperative topology. Specifically, the node and all its edges can be removed from the cooperative topology graph, preventing it from participating in group cooperative control. Other nodes can be notified to stop receiving data or commands from the node. Alternatively, the node's state can be marked as isolated and continuously monitored. After logical isolation, the remaining nodes reconstruct the formation based on the new cooperative topology graph and continue executing the mission.

[0207] In this implementation, the security control strategy can be a single strategy or a combination of multiple strategies. For example, for mildly anomalous nodes, the weight can be reduced and the communication frequency limited initially. If the anomaly continues to worsen, it can be escalated to logical isolation or forced return. The selection and execution timing of the strategy can be automatically triggered according to preset rules, or it can be executed after remote confirmation by ground operators.

[0208] In one specific implementation scheme, an abnormal node identification method based on UAV swarm cooperative topology operation constraint verification is provided, which may specifically include:

[0209] This invention is deployed in the security analysis module of an unmanned aerial vehicle (UAV) swarm control system. By continuously collecting the communication status and spatial coordination status of UAV nodes, it models and verifies the swarm coordination topology operation status, thereby realizing real-time identification and control of abnormal nodes.

[0210] Step 1: Collaborative Data Acquisition and Preprocessing

[0211] During the coordinated flight of a drone swarm, each drone node periodically reports its communication link status and location information to the swarm control center. Assume the drone swarm includes... Each node, at time... Acquire communication status matrix:

[0212] ;

[0213] In the formula, Let N represent the N×N matrix representing the effective communication connection status between all pairs of nodes in the UAV swarm at time t, where N is the total number of nodes in the UAV swarm. This is a status indicator indicating whether there is a valid communication connection between node i and node j at time t.

[0214] Simultaneously collect location information vectors:

[0215] ;

[0216] In the formula, This represents the set of three-dimensional spatial position information of all nodes in the UAV swarm at time t, where each element corresponds to the three-dimensional coordinates of a node. The subscript i indicates the i-th UAV node, corresponding to the i-th node in the position information. This represents the coordinate value of node i in the X-axis direction of the three-dimensional coordinate system at time t. This represents the coordinate value of node i in the Y-axis direction in the three-dimensional spatial coordinate system at time t. This represents the coordinate value of node i in the Z-axis direction in the three-dimensional spatial coordinate system at time t.

[0217] The system performs time synchronization, missing value repair, and outlier filtering on the collected data to ensure that communication status and location information remain consistent within the same time window, providing accurate input for subsequent topology analysis.

[0218] Step 2: Construction of UAV Swarm Collaborative Topology

[0219] Construct a communication adjacency matrix based on the communication state matrix:

[0220] ;

[0221] In the formula, It is based on the communication state matrix The elements in the constructed communication adjacency matrix are used to indicate whether there is a communication adjacency relationship between UAV node i and node j at time t.

[0222] Simultaneously, the spatial distance matrix between nodes is calculated based on the location information vector:

[0223] ;

[0224] In the formula, It is based on the location information vector The elements in the calculated spatial distance matrix between nodes are used to represent the three-dimensional straight-line distance (Euclidean distance) between UAV node i and node j at time t. , as well as It is the three-dimensional spatial coordinate components (X, Y, Z axis coordinates) of node i at time t. , as well as It is the three-dimensional spatial coordinate component of node j at time t.

[0225] The communication adjacency matrix and spatial distance matrix are jointly modeled under a unified topological representation framework to construct a cooperative topology graph for UAV swarms. It is used to characterize the collaborative operation structure of current drone swarms.

[0226] Step 3: Construction of Topological Stability Model

[0227] When the system is in normal collaborative state, multiple rounds of topology data are collected, and statistical analysis is performed on the distribution of node degree values ​​and the changes in distance between nodes to calculate the historical mean of node degree values. With variance : ;

[0228] And statistically analyze the stable intervals of the distance changes between nodes. To form a topological constraint model for the collaborative operation of unmanned aerial vehicle swarms This is used to characterize the operational constraints that topological changes must satisfy under normal cooperative flight conditions. Among them, This represents the lower threshold for the spatial distance between drone nodes. This represents the upper limit threshold for the spatial distance between drone nodes.

[0229] Step 4: Real-time collaborative topology offset calculation

[0230] During the real-time operation phase, the system computing nodes The degree value at the current moment:

[0231] ;

[0232] In the formula, Let be the real-time degree value of the i-th drone node at time t.

[0233] And calculate the spatial cooperative change of the nodes. Based on the above characteristics, a constraint deviation function is constructed to characterize the degree of deviation between the current running state of a node and the cooperative topology running constraints:

[0234] ;

[0235] In the formula, This represents the constraint deviation of the i-th drone node at time t; and These are the weighting coefficients, where, The weights for communication coordination constraints are used to determine the contribution of deviations in the adjustment degree value to the overall deviation degree. The weights of spatial coordination constraints are used to adjust the contribution of spatial coordination variation deviation to the overall deviation. For the degree deviation term, Let be the real-time degree value of the i-th node at time t. This represents the historical average node degree value under normal collaborative conditions. Let be the spatial collaborative change of the i-th node at time t.

[0236] Step 5: Determine abnormal nodes

[0237] When node constraint deviation Exceeding the allowable deviation range defined in the cooperative topology operation constraint model ,

[0238] ;

[0239] In the formula, This is the threshold for the allowed deviation range defined in the cooperative topology operation constraint model.

[0240] Furthermore, if this deviation persists across multiple consecutive sampling periods, the node is determined to be an abnormal node. Through a continuous verification mechanism based on operational constraints, normal coordinated fluctuations and abnormal operating states can be effectively distinguished, reducing the probability of false alarms caused by short-term communication jitter or formation adjustment processes.

[0241] Step 6: Security Control and Collaborative Topology Reconstruction

[0242] When an abnormal node is detected, the system automatically executes security control strategies, including reducing the decision weight of the abnormal node in the swarm collaborative control, restricting its communication capabilities, or logically isolating it from the collaborative topology, and reconstructing the collaborative topology based on the remaining nodes to ensure that the constraints of the UAV swarm collaborative operation are continuously met after the abnormality is handled.

[0243] Step 7: Output of Execution Results and Audit Log

[0244] The system uniformly stores abnormal node identifiers, comprehensive offset scores, security control strategy execution processes, and topology reconstruction results, and outputs a UAV swarm collaborative security assessment report, providing a basis for subsequent system security audits and optimizations.

[0245] Through the above method, the present invention can jointly model and verify the communication relationship and spatial coordination status of UAV swarms during collaborative flight, achieve accurate identification of abnormal nodes, and trigger security control strategies in a timely manner without relying on known attack characteristics, effectively preventing abnormal nodes from adversely affecting the collaborative operation of the swarm control system.

[0246] This invention takes the topological constraints of UAV swarm collaborative operation as the core analysis object, and breaks through the limitations of existing technologies that only perform security detection on a single node or a single dimension. It can detect covert attack behaviors such as firmware tampering and protocol hijacking from the level of swarm collaborative operation. It has good engineering feasibility and strong general adaptability, and significantly improves the security and robustness of UAV swarm control system in complex environments.

[0247] 1. This invention is no longer limited to detecting the identity or communication content of a single UAV node, but starts from the overall collaborative topology of the UAV swarm, and jointly models the communication and spatial collaborative relationships between nodes to achieve overall security analysis of the collaborative operation status of the swarm.

[0248] 2. This invention analyzes historical normal swarm flight data to form a cooperative topology operation constraint model for UAV swarms. During actual operation, it continuously verifies whether changes in the topology structure meet the cooperative operation constraints. It can identify abnormal nodes based on non-attack characteristics rather than operational constraints.

[0249] 3. The anomaly identification mechanism of the present invention is based on cooperative operation constraint verification rather than known attack characteristics. It also has the ability to identify unknown attacks such as firmware tampering and communication protocol hijacking, which significantly improves the security protection generalization capability of the UAV swarm control system.

[0250] 4. This invention does not rely on a specific hardware platform or fixed formation mode, and can be flexibly integrated into existing UAV swarm control systems as a safety assessment and control module, which has strong engineering practical value.

[0251] According to an embodiment of the present invention, an electronic device is provided; please refer to... Figure 4 The electronic device in this embodiment may include one or more of the following components: a processor, a network interface, memory, non-volatile memory, and one or more application programs, wherein the one or more application programs may be stored in non-volatile memory and configured to be executed by one or more processors, and the one or more programs are configured to perform the methods as described in the foregoing method embodiments.

[0252] According to embodiments of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a computer, causes the computer to perform the method described in any of the above embodiments.

[0253] According to embodiments of the present invention, a computer program product comprising instructions is also provided, which, when executed by a computer, cause the computer to perform a method in any of the above embodiments.

[0254] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for identifying abnormal nodes based on the constraint verification of cooperative topology operation of unmanned aerial vehicle swarms, characterized in that, The method includes: Acquire the communication link status and location information reported by each drone node during the coordinated flight of the drone swarm; Based on the communication link status and the location information, a collaborative topology diagram is constructed to characterize the communication connection relationship and spatial collaboration relationship between the nodes of the UAV swarm; Based on the cooperative topology graph and the preset cooperative topology operation constraint model, it is verified whether the operation state of each UAV node meets the corresponding cooperative topology operation constraints; wherein, the cooperative topology operation constraint model is used to define the constraint conditions that the node connection relationship and spatial cooperation relationship in the cooperative topology graph should meet under normal cooperative flight state; Drone nodes that do not meet the cooperative topology operation constraints will be identified as abnormal nodes.

2. The method according to claim 1, characterized in that, The steps for obtaining the communication link status and location information reported by each UAV node during the cooperative flight of the UAV swarm include: Receive communication link status and location information reported by each UAV node; The communication link status includes at least one of the following: the link connectivity matrix collected by each UAV node through the airborne communication module, the rate of change of communication signal strength over time, the round-trip delay statistics between nodes, and the packet loss rate and retransmission count; the location information includes at least one of the following: the three-dimensional spatial coordinates collected by each UAV node through the airborne positioning module, the relative distance matrix between nodes, and the relative azimuth angle between nodes and its rate of change.

3. The method according to claim 2, characterized in that, The step of constructing a cooperative topology map representing the communication connection relationship and spatial cooperation relationship between UAV swarm nodes based on the communication link status and the location information includes: A communication topology subgraph reflecting the communication connection relationship between UAV nodes is constructed based on the communication link status; Construct a location collaboration subgraph based on location information to reflect the spatial collaboration relationships between UAV nodes; By jointly modeling the communication topology subgraph and the location collaboration subgraph under a unified topology representation framework, a collaborative topology graph representing the communication connection relationship and spatial collaboration relationship between nodes in an unmanned aerial vehicle swarm is obtained.

4. The method according to claim 1, characterized in that, The collaborative topology operation constraint model includes a static baseline constraint sub-model and a scenario dynamic adaptation sub-model. The static baseline constraint sub-model is a constraint sub-model established based on historical normal cooperative flight data; wherein, the static baseline constraint sub-model includes at least one of the following: allowable range of node degree value change, allowable range of relative distance change between nodes, and allowable range of node communication status change; The scenario dynamic adaptation sub-model is a sub-model with a pre-set power inspection scenario feature tag library; wherein, the power inspection scenario feature tag library includes at least one of terrain feature tags, electromagnetic environment feature tags, and formation operation feature tags; wherein, the scenario dynamic adaptation sub-model is used to dynamically adjust the constraint threshold in the static benchmark constraint sub-model according to the real-time identified current operation scenario feature tags.

5. The method according to claim 4, characterized in that, The step of verifying whether the operating state of each UAV node satisfies the corresponding cooperative topology operating constraints based on the cooperative topology graph and the preset cooperative topology operating constraint model includes: Identify the operational scenario feature tags of the current flight of the drone swarm; Based on the identified operation scenario feature labels, the scenario dynamic adaptation sub-model is invoked to dynamically adjust the constraint thresholds in the static baseline constraint sub-model to obtain the currently applicable collaborative topology operation constraints. For each UAV node, extract the real-time topological features of that UAV node in the collaborative topology graph; wherein, the real-time topological features include at least one of the following: node degree value, relative distance to neighboring nodes, and communication status; The real-time topology features are compared with the currently applicable cooperative topology operation constraints; If the real-time topology characteristics exceed the range defined by the currently applicable cooperative topology operation constraints, then the operating state of the UAV node is determined to not meet the corresponding cooperative topology operation constraints.

6. The method according to claim 5, characterized in that, The step of identifying drone nodes that do not meet the cooperative topology operation constraints as abnormal nodes includes: Mark drone nodes that do not meet the cooperative topology operation constraints as suspected abnormal nodes; Obtain the operational status information of each UAV node in the neighboring node set of the suspected abnormal node; wherein, the neighboring node set includes at least one UAV node that has a direct or indirect communication connection with the suspected abnormal node; Based on the running status information of the adjacent node set and the characteristics of the current work scenario, verify whether the suspected abnormal node is a true abnormal node. If the verification result is a true anomaly, then the suspected anomaly node is determined to be an anomaly node.

7. The method according to claim 6, characterized in that, The step of verifying whether a suspected abnormal node is a true abnormal node based on the running status information of the adjacent node set and the characteristics of the current work scenario includes: If the current work scenario feature label is identified to include a high-voltage electromagnetic environment feature label, then the high-voltage electromagnetic interference scenario verification sub-process is executed. This high-voltage electromagnetic interference scenario verification sub-process includes: determining whether the communication status of the suspected abnormal node exceeds the range defined by the currently applicable cooperative topology operation constraints; if it does, determining whether there are any neighboring nodes in the neighboring node set whose communication status fluctuates synchronously; if so, determining that the deviation in the communication status of the suspected abnormal node is caused by high-voltage electromagnetic environment interference, and the verification result is a false anomaly; if not, the verification result is a true anomaly. If the current operation scenario feature label is identified as including the pole-and-tower formation operation feature label, then the formation maneuver scenario verification sub-process is executed. This sub-process includes: determining whether the relative distance between the suspected abnormal node and its adjacent nodes exceeds the range defined by the currently applicable cooperative topology operation constraints; if it does, determining whether the adjacent node set has undergone a consistent positional offset along the spatial axis direction associated with the formation maneuver command; wherein the spatial axis direction is determined based on the maneuver type and maneuver parameters included in the formation maneuver command; if consistent, determining that the positional deviation of the suspected abnormal node is caused by the overall formation maneuver, and the verification result is a false anomaly; if inconsistent, the verification result is a true anomaly.

8. The method according to claim 7, characterized in that, The step of determining whether there are more than a preset proportion of neighboring nodes in the neighboring node set whose communication states are experiencing synchronous fluctuations includes: Determine the formation type of the current power line inspection operation of the drone swarm; wherein, the formation type includes linear formation along the power line and planar formation around the tower; Based on the formation arrangement type and the interference coverage of the high-voltage electromagnetic environment, the judgment ratio of communication state synchronization fluctuation is dynamically determined as a preset ratio. From the communication status of the suspected abnormal nodes and their neighboring nodes, indicators strongly correlated with the high-voltage electromagnetic environment are extracted as judgment indicators; wherein, the judgment indicators include communication signal strength and packet loss rate; Based on the judgment index, determine whether the number of nodes in the adjacent node cluster whose judgment index fluctuation period matches the high-voltage transmission frequency and whose fluctuation trend is consistent with the suspected abnormal node reaches a dynamically determined judgment ratio. If yes, it is determined that the communication status of more than a preset proportion of neighboring nodes has experienced synchronous fluctuations; if no, it is determined that none exist.

9. A safety control method based on the cooperative flight of unmanned aerial vehicle (UAV) swarms, characterized in that, The method includes: Identify abnormal nodes; wherein, the abnormal node identification method based on the verification of cooperative topology operation of UAV swarm as described in any one of claims 1-8 is used to identify abnormal nodes; A preset security control strategy is executed for the abnormal node to limit its impact on the coordinated flight of the drone swarm.

10. An electronic device, characterized in that, include: A memory, and one or more processors communicatively connected to the memory; The memory stores instructions that can be executed by the one or more processors, which are executed by the one or more processors to enable the one or more processors to implement the abnormal node identification method based on the topology operation constraint verification of UAV swarm as described in any one of claims 1 to 8.

Images