Unmanned aerial vehicle cluster dynamic ad hoc network method and system

Abstract

The invention discloses an unmanned aerial vehicle cluster dynamic ad hoc network method and system. The method comprises the following steps: acquiring key parameters of each unmanned aerial vehicle; dynamically electing a main unmanned aerial vehicle and member unmanned aerial vehicles through a comprehensive weight model; the main unmanned aerial vehicle performs dynamic clustering on an unmanned aerial vehicle cluster based on a Voronoi graph algorithm to obtain a secondary topological structure; according to the second-level topological structure, the main unmanned aerial vehicle coordinates with multi-path routing and path measurement, and an unmanned aerial vehicle cluster ad hoc network is constructed and optimized; whether the member unmanned aerial vehicles are maliciously attacked or not is detected through the main unmanned aerial vehicle, and if the member unmanned aerial vehicles are attacked, the corresponding unmanned aerial vehicles are isolated and the main unmanned aerial vehicle is selected again. According to the method, the main unmanned aerial vehicle is dynamically elected, dynamic clustering is performed, the link stability prediction model is combined, the unmanned aerial vehicle cluster structure is automatically adjusted according to the unmanned aerial vehicle position, the link problem is repaired in advance, and the network robustness is enhanced. A multi-path routing protocol is matched with a path measurement model, and an optimal path is selected, so that the data transmission efficiency is improved, and the energy consumption and delay are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) and, in particular, to a method and system for dynamically self-organizing a UAV cluster network. Background Art

[0002] Drones are widely used in areas such as regional reconnaissance and disaster relief. When there is a mission, swarms of drones are often deployed. Scanning by multiple drones can achieve a comprehensive scan of the disaster-stricken area or reconnaissance area with high scanning efficiency. However, multiple drones in dynamic environments often face challenges such as frequent changes in topological relationships, energy constraints, and security threats.

[0003] Traditionally, a fixed master drone architecture is generally used. However, in a dynamic environment, this approach has problems such as single point failure and uneven load. When the drone formation changes, the distance between the master drone and some drones may be far, resulting in a weakened signal transmission capability between drones. Patent publication number CN117042083A discloses that in an unmanned cluster network, when a drone is damaged and a network blind spot appears, the invention can quickly calculate a mobile strategy for drones in the entire network to ensure network connectivity; and can calculate reliable paths and backup paths that balance load and minimize energy consumption in the network, thereby improving the efficiency of information transmission. However, in this method, the behavior of the master drone changing is passive, and the change of the master drone is determined by the backup path. However, due to too much uncertainty in a dynamic environment, this method still cannot solve the problem of unstable signal transmission between the changed master drone and other drones. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a method and system for dynamically self-organizing a cluster of drones to solve the problem of unstable signal transmission capability between clustered drones.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a method for dynamically organizing a drone cluster network, comprising:

[0008] Get key parameters of each drone;

[0009] Based on the key parameters of drones, the master drone and member drones are dynamically elected through a comprehensive weight model;

[0010] The master UAV dynamically clusters the UAV cluster based on the Voronoi diagram algorithm to obtain the secondary topological structure;

[0011] Based on the secondary topology, the master UAV coordinates multi-path routing and path metrics to build and optimize the UAV cluster self-organizing network;

[0012] The master drone is used to detect whether the member drones are under malicious attack. If an attack is detected, the corresponding drone is isolated and a new master drone is selected.

[0013] As a preferred solution of the method for the dynamic self-organizing network of drone clusters described in the present invention, the master drone is dynamically elected through a comprehensive weight model based on the key parameters of the drones, including:

[0014] Based on the comprehensive weight model, calculate the weight value of each drone;

[0015] All drones broadcast their status information periodically. When a drone's weight exceeds the weight threshold, it recommends itself as a candidate master drone.

[0016] The candidate master drones reach a consensus through the Byzantine fault-tolerant algorithm and elect the drone with the highest weight as the new master drone;

[0017] The old master drone hands over the routing table, and the new master drone completes the switch and broadcasts the update within the set time.

[0018] As a preferred solution of the UAV cluster dynamic self-organizing network method described in the present invention, the calculation formula of the comprehensive election weight model is:

[0019]

[0020] Among them, W i is the comprehensive election weight, p i is the current position of the drone, is the geometric center of the UAV cluster, R is the communication radius of the UAV cluster, and E i is the remaining battery power of the drone, E max D is the maximum power of the drone. i is the total number of other drones adjacent to the drone, D max is the maximum number of drones, α, β and γ are weight coefficients, and α+β+γ=1.

[0021] As a preferred solution of the method for dynamic self-organizing networking of drone clusters described in the present invention, the master drone dynamically clusters the drone cluster based on the Voronoi diagram algorithm to obtain a secondary topological structure, including:

[0022] The master UAV divides the UAV swarm into multiple clusters by obtaining Voronoi cells using the Voronoi diagram algorithm based on the positions of the member UAVs.

[0023] A link stability prediction model is used to calculate the link quality between drones and predict the link. If the predicted link reaches the failure threshold, the link is repaired.

[0024] When the density of drones in a cluster is higher than the density threshold, the cluster is split; when the density of drones in a cluster is lower than the density threshold, the cluster is merged with the adjacent cluster.

[0025] As a preferred solution of the UAV cluster dynamic self-organizing network method described in the present invention, the calculation formula of the link stability prediction model is:

[0026]

[0027] Among them, P link (t+Δt) is the predicted link quality at the future time point t+Δt, P Rx is the signal strength currently received by the drone, η is the path loss index, v i and v j are the velocity vectors of drones i and j respectively, d0 is the reference distance, ||v i -v j || is the relative speed between drones i and j.

[0028] As a preferred solution of the method for the dynamic self-organizing network of drone clusters described in the present invention, wherein: according to the secondary topology structure, the master drone coordinates multi-path routing and path metrics to build and optimize the drone cluster self-organizing network, including:

[0029] The master drone maintains the global routing table, and the member drones save the local routes:

[0030] An improved AODV protocol is used for on-demand route discovery, and a routing table is generated based on link stability and energy consumption indicators.

[0031] Based on the path metric model, the optimal route between the master UAV and the member UAVs is determined, and the ad hoc network between the master UAV and the member UAVs is obtained.

[0032] As a preferred solution of the method for the dynamic self-organizing network of drone clusters described in the present invention, the method of determining whether a member drone is under malicious attack by detecting the master drone includes:

[0033] Traffic anomaly detection is used to determine whether a member drone is under malicious attack, and the abnormal traffic detection is obtained through an LSTM network.

[0034] In a second aspect, the present invention provides a drone cluster dynamic self-organizing network system, comprising:

[0035] Data acquisition module, used to obtain key parameters of each drone;

[0036] An election module is used to dynamically elect the master drone and member drones through a comprehensive weight model based on the key parameters of the drones;

[0037] The partitioning and networking module is used by the master UAV to dynamically cluster the UAV cluster based on the Voronoi diagram algorithm to obtain a secondary topology structure. Based on the secondary topology structure, the master UAV coordinates multi-path routing and path metrics to build and optimize the UAV cluster self-organizing network;

[0038] The secondary election module is used to determine whether the member drones are attacked maliciously through the master drone detection. If an attack is detected, the corresponding drones are isolated and the master drone is re-elected.

[0039] In a third aspect, the present invention provides a computing device, comprising:

[0040] memory and processor;

[0041] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method for dynamic self-organizing networking of drone clusters are implemented.

[0042] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for dynamic self-organizing networking of drone clusters.

[0043] Compared with existing technologies, the present invention offers the following advantages: It dynamically selects a master drone based on various factors, performs dynamic clustering, and integrates a link stability prediction model to automatically adjust the drone cluster structure based on drone location, proactively remediating link issues and enhancing network robustness. The present invention's multipath routing protocol, combined with a path metric model, selects the optimal path, improves data transmission efficiency, and reduces energy consumption and latency. When a master drone fails or requires switching, candidate master drones are rapidly elected using a Byzantine fault-tolerant algorithm, ensuring rapid response and efficient switching when network topology changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 This is a logical diagram of the overall process of the UAV cluster dynamic self-organizing network method according to one embodiment of the present invention. DETAILED DESCRIPTION

[0046] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0047] Example 1

[0048] Reference Figure 1 , as one embodiment of the present invention, provides a method for a dynamic self-organizing network of a drone cluster, comprising:

[0049] S100: Obtain key parameters of each drone; based on the key parameters of the drone, dynamically elect the master drone and member drones through a comprehensive weight model;

[0050] In a possible embodiment, the key parameters of the drone include drone battery level, network connection status, drone location, routing table, etc.

[0051] A hierarchical dynamic election architecture is constructed, including a master drone, candidate master drones, and member drones; the master drone is the networking center, used for networking with candidate master drones and member drones, as well as signal transmission; and the candidate master drones are the group that competes when electing the master drone. After the master drone is selected, the candidate master drones also become member drones.

[0052] Preferably, based on the comprehensive weight model, the weight value of each drone is calculated; all drones periodically broadcast status information, and when the weight value of a drone exceeds the weight threshold, it recommends itself as a candidate master drone;

[0053] Preferably, the candidate master drones reach a consensus through the Byzantine fault-tolerant algorithm and elect the drone with the highest weight as the new master drone; the old master drone transfers the routing table, and the new master drone completes the switch and broadcasts the update within the set time.

[0054] In the embodiment of the present invention, the calculation formula of the comprehensive election weight model is:

[0055]

[0056] Among them, W i is the comprehensive election weight, p i is the current position of the drone, is the geometric center of the UAV cluster, R is the communication radius of the UAV cluster, and E i is the remaining battery power of the drone, E maxD is the maximum power of the drone. i is the total number of other drones adjacent to the drone, D max is the maximum number of drones, α, β and γ are weight coefficients, and α+β+γ=1.

[0057] In one possible embodiment, the weight threshold may be set to 0.6, 0.7, 0.8, etc.;

[0058] The calculation formula of the Byzantine fault tolerance algorithm is:

[0059] Leader=argmax i∈C (W i );

[0060] Among them, Leader is the elected master drone, C is the set of candidate master drones, and the drone with the largest weight value is selected from the candidate master drones, that is, the master drone.

[0061] In an embodiment of the present invention, the old master drone hands over the routing table, and the new master drone completes the switch and broadcasts the update within 90-100ms.

[0062] It should be noted that the comprehensive election weight model is calculated based on the location, remaining battery power and connectivity of the drone. This metric is used to assess the drone's position within the swarm. The remaining battery life reflects the drone's endurance and is particularly important for the position of a master drone, which requires long-term stable operation. If a drone's battery reserves are low, even if it performs well in terms of location and connectivity, it may not be selected as the new master drone. This can prevent network partitioning or frequent master switching due to energy depletion. Therefore, considering remaining battery life can effectively improve the stability and efficiency of the entire drone swarm network.

[0063] S102: The master UAV dynamically clusters the UAV cluster based on the Voronoi diagram algorithm to obtain a secondary topological structure;

[0064] Preferably, the master UAV divides the UAV group into multiple clusters by obtaining Voronoi cells through a Voronoi diagram algorithm according to the positions of the member UAVs;

[0065] In the embodiment of the present invention, the calculation process of obtaining the Voronoi unit by dividing the Voronoi diagram algorithm is as follows:

[0066]

[0067] Among them, V iis the Voronoi cell with UAV i as the cluster head, p is any point in space, p i is the current position of drone i, j is all other drones except drone i, ‖pp i ‖ represents the Euclidean distance from a point p in space to drone i, ||pp j ||Similarly, it is the Euclidean distance from a point p in space to drone j.

[0068] Preferably, a link stability prediction model is used to calculate the link quality between drones and predict the link. If the predicted link reaches a failure threshold, the link is repaired.

[0069] Preferably, the cluster is split when the density of drones in the cluster is higher than the density threshold, and is merged with the adjacent cluster when the density of drones in the cluster is lower than the density threshold.

[0070] In the embodiment of the present invention, the calculation formula of the link stability prediction model is:

[0071]

[0072] Among them, P link (t+Δt) is the predicted link quality at the future time point t+Δt, P Rx is the signal strength currently received by the drone, η is the path loss index, v i and v j are the velocity vectors of drones i and j respectively, d0 is the reference distance, ||v i -v j || is the relative speed between drones i and j.

[0073] As drones move or some fail, the original grouping method may no longer be applicable. Therefore, it is necessary to be able to automatically adjust its structure to cope with these changes, that is, topology adaptive adjustment. By sending and receiving heartbeat packets, which are small data packets similar to health checks, and analyzing changes in signal strength, it is possible to predict which connections may have problems. If a connection is predicted to deteriorate, the system will attempt to repair the problem in advance. If it is found that there are too many drones in a group, the group will be split into two or more groups. Conversely, if there are too few drones in a group, it will be considered to be merged with other adjacent groups. This is done to maintain workload balance between groups and avoid waste or over-concentration of resources. Link repair can be achieved by adjusting the transmission power, switching to an alternative frequency, or re-planning the route.

[0074] It should be noted that based on the Voronoi diagram algorithm, the drone cluster is dynamically clustered to form a two-level topological structure; the communication structure within the drone cluster is effectively organized to ensure that information can be transmitted efficiently and stably within the cluster; and the Voronoi diagram clustering algorithm specifically divides the entire drone cluster into several small groups, that is, clusters, and each group has a leader, that is, the cluster head; the advantage of this is that the distance between the drones in each group is closer, which facilitates fast and effective communication, while also reducing the need for cross-group communication and improving overall efficiency.

[0075] S104: Based on the secondary topology, the master UAV coordinates multi-path routing and path metrics to build and optimize the UAV cluster self-organizing network;

[0076] Preferably, the master drone maintains the global routing table and the member drones save the local routes.

[0077] Preferably, an improved AODV protocol is used for on-demand route discovery, and a routing table is generated in combination with link stability and energy consumption indicators.

[0078] Preferably, based on the path metric model, the optimal route between the master UAV and the member UAVs is determined to obtain an ad hoc network between the master UAV and the member UAVs;

[0079] In the embodiment of the present invention, the calculation formula of the path metric model is:

[0080]

[0081] Among them, T k is the delay on path k, T max is the maximum delay in the ad hoc network, E k is the energy consumption of the kth link or UAV on the path, E max is the maximum energy consumption value of any link or UAV in the network, S k is the link stability factor, λ is the weight of the stability factor; It is used to measure the relative energy consumption of each segment k in the path p, that is, to standardize the energy consumption. Similarly, It is also to standardize the delay.

[0082] When the energy level of some drones drops below a preset threshold, the energy balancing scheduling strategy will be activated to try to reduce the pressure on these drones by reallocating tasks or choosing other more energy-efficient communication paths;

[0083] The process of energy balance scheduling is as follows:

[0084]

[0085] Select Next Hop represents the objective function or strategy for selecting the next-hop drone. The goal is to find the optimal next-hop drone j so that the energy consumption of the entire network is more balanced. j represents one of the candidate next-hop drones. There may be multiple candidate drones in the network as the next hop for data forwarding, and the most suitable one needs to be selected from them. Refers to the energy used by UAV j, which usually refers to the total amount of energy consumed by the UAV to perform various tasks such as data transmission and processing from the start of the network operation to the current moment; Refers to the current remaining energy of UAV j;

[0086] Assume that there are three candidate drones A, B and C in a drone swarm, and their energy states are as follows:

[0087]

[0088] It can be seen that according to the minimization principle, UAV A will be selected as the next hop because it has the lowest energy consumption ratio, indicating that it has more unused energy than other UAVs and is suitable for taking on more data forwarding tasks.

[0089] In addition, after forming an ad hoc network, Kalman filtering can be used to predict the location of drones, update the routing table in advance to reduce the possibility of link interruption, and monitor the link quality through a two-way heartbeat packet mechanism combined with a signal strength prediction model, so that link repair or re-routing can be carried out in a timely manner to ensure the connectivity and stability of the network.

[0090] The calculation formula for predicting the position of the UAV by Kalman filter is:

[0091]

[0092] in, is the prediction of the position of UAV i at time t+1, F is the state transfer matrix, B is the control input matrix, u i (t) is the control input to UAV i at time t, ω t is the process noise;

[0093] Because the network topology changes as the drone moves, technologies such as Kalman filtering are used to predict the drone's location and update the routing table in advance to reduce the possibility of link interruption, allowing the system to respond more proactively to upcoming changes rather than passively.

[0094] It should be noted that the purpose of calculating multipath routing and path metrics is to enable drone swarms to communicate efficiently and stably in complex and ever-changing environments. Through master node coordination, on-demand path search, comprehensive evaluation of path quality, and real-time strategy adjustment, good communication performance can be maintained even in the face of changes in network topology or failure of individual drones.

[0095] Because drones are flying in the air and their positions are constantly changing, their routes will also change accordingly. Therefore, the system needs to continuously monitor the network status and dynamically adjust the routing table based on actual conditions. For example, if a path suddenly becomes unstable or consumes too much energy, the system will automatically switch to another more optimal path.

[0096] By adopting an on-demand routing mechanism coordinated by the master drone and combining link stability and energy consumption indicators to select the optimal path, it is ensured that even if the network topology changes or some nodes fail, an effective communication link can still be maintained, avoiding network interruption caused by single point failure; the path cost is calculated based on comprehensive factors such as delay, energy consumption and stability, so that data packets can be transmitted through the most appropriate path, avoiding excessive energy consumption or excessive concentration on certain nodes, thereby extending the working life of the entire network; when the master drone is detected to fail, the new master drone is elected and the network-wide routing table is updated in an extremely short time (such as no more than 100 milliseconds), ensuring that the network can quickly adapt to the new state and reduce delays and service interruptions caused by leadership changes.

[0097] S106: Determine whether a member drone is under malicious attack through the master drone detection. If an attack is detected, isolate the corresponding drone and reselect the master drone;

[0098] In an embodiment of the present invention, the master UAV uses dynamic identity authentication to verify the identity of the member node. If the member identity is abnormal, the member node is isolated;

[0099] Dynamic identity authentication uses a lightweight ECC signature algorithm. The signature generation process includes:

[0100] Sign(m)=(r,s)

[0101] r=(kG) x

[0102] s=k -1 (H(m)+dr)mod n

[0103] Where Sign(m) is the signature result of message m, s and r are parts of the signature, r is the x-coordinate of a point obtained by the elliptic curve point multiplication operation kG, k is the random number used in the signing process, H(m) is the hash value of message m, which is usually used to ensure data integrity, d is the private key, and mod n means modular operation, and n is the order of the elliptic curve, that is, the order of the base point G.

[0104] Preferably, traffic anomaly detection is used to determine whether a member drone is under malicious attack, and abnormal traffic detection is obtained through an LSTM network;

[0105] The flow detection formula is expressed as:

[0106] y t =Softmax(W t h t +b h )

[0107] h t =LSTM(x t ,h t-1 )

[0108] Among them, y t is the output of the output layer, W t and b h are the weight matrix and bias term, h t is the hidden state at the current moment, h t-1 is the state at the previous moment, x t is the current input; and y t It is the output after the Softmax function conversion, which represents the probability that the behavior belongs to a certain category at the current time step; h t It is the hidden state inside the LSTM unit, which carries the historical information in the sequence but is not directly used for anomaly judgment.

[0109] The steps of abnormal traffic detection are as follows: First, the master UAV collects data streams from member nodes as input x t ,These data can include various network activity characteristics, such as,transmission rate, request frequency, etc.;

[0110] The LSTM network is based on historical data, that is, the state h at the previous moment t-1 and the current input x t , calculate the hidden state h at the current moment t ;

[0111] Then, a fully connected layer is used to transform the hidden state h t Convert to output y t ,This output is usually a probability distribution, which indicates the probability that the currently observed behavior belongs to normal behavior;

[0112] Finally, compare the output y t Is it consistent with the expected behavior mode? If it displays y t If the probability of the current behavior significantly deviates from the normal range, it is considered that abnormal behavior has occurred.

[0113] In addition, the expected behavior patterns are obtained by relying on statistics and machine learning techniques and learning from a large amount of normal operation data.

[0114] It should be noted that the master drone uses dynamic identity authentication to verify the identities of member nodes, maintaining the security of the entire drone swarm network and preventing network performance degradation or even paralysis due to failures or attacks on individual member nodes. If an attack is detected, the corresponding member node is immediately isolated and a new master drone is selected. Since the isolation of a member node inevitably leads to a change in the overall formation of the drone swarm, reselecting the master drone based on the new formation can maintain optimal communication between drones.

[0115] It is also explained that the dynamic self-organizing network process of the present invention runs through the entire process of the cluster drones performing tasks, and when a member node is attacked, resulting in a change in the drone formation, when the energy of some drones, or even the main drone, is lower than the threshold, or the comprehensive weight of the candidate main drone exceeds that of the main drone, the above situations will trigger a mechanism for re-electing the main drone. Because the occurrence of the above problems means that the networking between the drone clusters will be unstable, re-electing the main drone can keep the self-organizing network of the drone cluster communicating smoothly.

[0116] The above is a schematic diagram of a method for dynamically self-organizing a drone swarm network according to this embodiment. It should be noted that the technical solution of the system for dynamically self-organizing a drone swarm network is identical to the technical solution of the method for dynamically self-organizing a drone swarm network described above. For details not described in detail in the technical solution of the system for dynamically self-organizing a drone swarm network according to this embodiment, please refer to the description of the technical solution of the method for dynamically self-organizing a drone swarm network described above.

[0117] The UAV cluster dynamic self-organizing network system in this embodiment includes:

[0118] Data acquisition module, used to obtain key parameters of each drone;

[0119] An election module is used to dynamically elect the master drone and member drones through a comprehensive weight model based on the key parameters of the drones;

[0120] The partitioning and networking module is used by the master UAV to dynamically cluster the UAV cluster based on the Voronoi diagram algorithm to obtain a secondary topology structure. Based on the secondary topology structure, the master UAV coordinates multi-path routing and path metrics to build and optimize the UAV cluster self-organizing network.

[0121] The secondary election module is used to determine whether the member drones are attacked maliciously through the master drone detection. If an attack is detected, the corresponding drones are isolated and the master drone is re-elected.

[0122] This embodiment further provides a computing device suitable for use in a dynamic self-organizing network of a drone cluster, including:

[0123] Memory and processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the method for realizing dynamic self-organizing networking of drone clusters as proposed in the above embodiment.

[0124] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for realizing a dynamic self-organizing network of a drone cluster proposed in the above embodiment is implemented.

[0125] The storage medium proposed in this embodiment and the method for realizing dynamic self-organizing networking of drone clusters proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0126] Through the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented with the help of software and necessary general hardware, and of course can also be implemented by hardware. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various embodiments of the present invention.

[0127] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for dynamic self-organizing networking of drone clusters, characterized in that: include: Get key parameters of each drone; Based on the key parameters of drones, the master drone and member drones are dynamically elected through a comprehensive weight model; The master UAV dynamically clusters the UAV cluster based on the Voronoi diagram algorithm to obtain the secondary topological structure; Based on the secondary topology, the master UAV coordinates multi-path routing and path metrics to build and optimize the UAV cluster self-organizing network; The master drone is used to detect whether the member drones are under malicious attack. If an attack is detected, the corresponding drone is isolated and a new master drone is selected.

2. The method for dynamic self-organizing networking of drone clusters according to claim 1, wherein: Based on the key parameters of the drones, the master drone is dynamically elected through a comprehensive weight model, including: Based on the comprehensive weight model, calculate the weight value of each drone; All drones broadcast their status information periodically. When a drone's weight exceeds the weight threshold, it recommends itself as a candidate master drone. The candidate master drones reach a consensus through the Byzantine fault-tolerant algorithm and elect the drone with the highest weight as the new master drone; The old master drone hands over the routing table, and the new master drone completes the switch and broadcasts the update within the set time.

3. The method for dynamic self-organizing networking of drone clusters according to claim 2, wherein: The calculation formula of the comprehensive election weight model is: Among them, W i is the comprehensive election weight, p i is the current position of the drone, is the geometric center of the UAV cluster, R is the communication radius of the UAV cluster, and E i is the remaining battery power of the drone, E max D is the maximum power of the drone. i is the total number of other drones adjacent to the drone, D max is the maximum number of drones, α, β and γ are weight coefficients, and α+β+γ=1.

4. The method for dynamic self-organizing networking of drone clusters according to claim 3, wherein: The master UAV dynamically clusters the UAV cluster based on the Voronoi diagram algorithm to obtain a secondary topological structure, including: The master UAV divides the UAV swarm into multiple clusters by obtaining Voronoi cells using the Voronoi diagram algorithm based on the positions of the member UAVs. A link stability prediction model is used to calculate the link quality between drones and predict the link. If the predicted link reaches the failure threshold, the link is repaired. When the density of drones in a cluster is higher than the density threshold, the cluster is split; when the density of drones in a cluster is lower than the density threshold, the cluster is merged with the adjacent cluster.

5. The method for dynamic self-organizing networking of drone clusters according to claim 4, wherein: The calculation formula of the link stability prediction model is: Among them, P link (t+Δt) is the predicted link quality at the future time point t+Δt, P Rx is the signal strength currently received by the drone, η is the path loss index, v i and v j are the velocity vectors of drones i and j respectively, d0 is the reference distance, ||v i -v j || is the relative speed between drones i and j.

6. The method for dynamic self-organizing networking of drone clusters according to claim 5, wherein: Based on the secondary topology, the master UAV coordinates multi-path routing and path metrics to build and optimize the UAV cluster self-organizing network, including: The master drone maintains the global routing table, and the member drones save the local routes: An improved AODV protocol is used for on-demand route discovery, and a routing table is generated based on link stability and energy consumption indicators. Based on the path metric model, the optimal route between the master UAV and the member UAVs is determined, and the ad hoc network between the master UAV and the member UAVs is obtained.

7. The method for dynamic self-organizing networking of drone clusters according to claim 1, wherein: The master drone is used to detect and determine whether the member drones are under malicious attack, including: Traffic anomaly detection is used to determine whether a member drone is under malicious attack, and the abnormal traffic detection is obtained through an LSTM network.

8. A system using the method for dynamic self-organizing networking of drone clusters according to any one of claims 1 to 7, characterized in that: include: Data acquisition module, used to obtain key parameters of each drone; An election module is used to dynamically elect the master drone and member drones through a comprehensive weight model based on the key parameters of the drones; The partitioning and networking module is used by the master UAV to dynamically cluster the UAV cluster based on the Voronoi diagram algorithm to obtain the secondary topology structure; Based on the secondary topology, the master UAV coordinates multi-path routing and path metrics to build and optimize the UAV cluster self-organizing network; The secondary election module is used to determine whether the member drones are attacked maliciously through the master drone detection. If an attack is detected, the corresponding drones are isolated and the master drone is re-elected.

9. An electronic device comprising: memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method for dynamic self-organizing networking of drone clusters described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for dynamic self-organizing networking of drone clusters as described in any one of claims 1 to 7.

Images