Block data distribution optimization method and system in unmanned aerial vehicle cluster network

Abstract

The invention discloses a block data distribution optimization method and system in an unmanned aerial vehicle cluster network, and the method comprises the steps: dividing unmanned aerial vehicles into a plurality of clusters, and enabling each cluster to comprise a storage unmanned aerial vehicle and a plurality of task unmanned aerial vehicles; secondly, making a block data distribution scheme of the storage unmanned aerial vehicles in each cluster based on the storage demand degree of the task unmanned aerial vehicles in the cluster for the block data; and finally, iteratively optimizing a cluster division result and a block data distribution scheme by repeating the operations of cluster division and block data distribution, and obtaining an optimal unmanned aerial vehicle cluster network block data distribution scheme according to the performance of each iteration of the cluster division result and the block data distribution scheme. According to the method, the unmanned aerial vehicles are divided into the task unmanned aerial vehicles and the storage unmanned aerial vehicles, along with task execution of the task unmanned aerial vehicles, the key data needing to be acquired can directly request data from the storage unmanned aerial vehicles in the air with lower time delay and energy consumption than those from the air to the ground, and rapid and reliable data access service is provided for the task unmanned aerial vehicles; and the communication cost in the unmanned aerial vehicle network is efficiently reduced.

Description

Technical Field

[0001] The present invention belongs to the field of drone and blockchain data storage, and specifically relates to a method and system for optimizing block data distribution in a drone cluster network. Background Art

[0002] Unmanned aerial vehicle (UAV) networks, with their unique flexibility and cost-effectiveness, have found widespread application in areas such as collaborative search, agricultural monitoring, and communication relay. However, as these applications expand, the data collected and stored by drones becomes increasingly important, and data privacy and security become key issues. In long-distance UAV communications, traditional relay-based data forwarding mechanisms are not only complex and carry a high risk of data privacy leakage, but also increase communication latency and energy consumption, severely impacting the real-time and energy efficiency of UAV networks. Therefore, new UAV network architecture strategies aimed at optimizing UAV network energy consumption and security are necessary and challenging. Blockchain technology, introduced as a technology with the potential to protect data privacy, offers new possibilities for information security in UAV networks.

[0003] From a communications security perspective, the high-frequency nature of drone networks presents potential security risks such as malicious attacks, false data, and unauthorized intrusions. The physical fragility of drone networks also contributes to the vulnerability of communication systems. Therefore, how to leverage blockchain technology to ensure the security of important communications data and access security is a key issue currently underway in drone networks.

[0004] From the perspective of communication efficiency, the exchange of information between drone networks and the ground often requires long-distance communication through relay nodes, resulting in low communication efficiency. Therefore, how to design a new drone network architecture and corresponding communication strategy based on blockchain technology to improve the communication efficiency between drone networks and ground data is a significant and challenging issue. Summary of the Invention

[0005] The purpose of the present invention is to provide a method and system for optimizing block data allocation in a drone cluster network in response to the above-mentioned problems existing in the prior art.

[0006] To achieve the above objectives, the technical solutions of the present invention are as follows:

[0007] In a first aspect, the present invention proposes a method for optimizing block data allocation in a drone cluster network, comprising:

[0008] S1. Divide the drones into multiple clusters, each of which contains a storage drone and multiple mission drones;

[0009] S2. Based on the storage requirements of the task drones in the cluster for block data, formulate a block data allocation plan for the storage drones in each cluster;

[0010] S3. Repeat the operations of steps S1-S2 to iteratively optimize the cluster division results and block data allocation scheme, and obtain the optimal drone cluster network block data allocation scheme based on the performance of the cluster division results and block data allocation scheme of each iteration.

[0011] Said S1 comprises:

[0012] S11, presetting a mixed similarity threshold between task drones, calculating the mixed similarity between all unassigned task drones, selecting two task drones with a mixed similarity greater than the preset mixed similarity threshold and the highest mixed similarity, and selecting one of them as the cluster center;

[0013] S12, calculating the mixed similarity between the remaining mission drones and the cluster center, and adding mission drones whose mixed similarity with the cluster center is higher than a preset mixed similarity threshold to the cluster;

[0014] S13, repeating steps S11-S12 until only task drones with mixed similarity below the threshold remain. Then, for each of these unassigned task drones, calculate their mixed similarity with all cluster centers and assign them to the cluster with the highest mixed similarity. At this point, the initialization of all clusters is completed.

[0015] S14, based on the initialized clusters, presetting a cluster similarity threshold between two clusters, calculating cluster similarities between all clusters, and merging two clusters whose cluster similarities are higher than the threshold;

[0016] S15. For each merged cluster, preset a hybrid similarity variance threshold and hybrid similarity quantile within each cluster, count the hybrid similarities between all mission drones within each cluster and its cluster center, calculate the hybrid similarity variance within the cluster, and for clusters with hybrid similarity variance greater than the threshold, split out the mission drones in the cluster whose hybrid similarity with the cluster center is lower than the preset hybrid similarity quantile, calculate the hybrid similarity between each split mission drone and other mission drones, and reallocate these mission drones to the cluster containing the other mission drones with the highest hybrid similarity.

[0017] S16. After the task drones are redistributed, storage drones equal in number to the number of clusters are allocated to each cluster, so that the task drones in the cluster can make data requests to the storage drones first, thus completing the division of all drone clusters.

[0018] The S2 includes:

[0019] S21. Calculate the storage requirement of each mission drone in all clusters for the first block of data. Add up the storage requirements of all mission drones in a single cluster to obtain the storage requirement of the current cluster for the first block of data.

[0020] The storage requirement of each mission drone for block data is calculated using the following formula:

[0021]

[0022] In the above formula, d jb For cluster c p Mission UAV in j The storage requirement for block b, I jb For mission drone u j Interest value for block b, α h is the weight of the adjacent drone cluster, u ph is the mission drone in the adjacent drone cluster, C ph For cluster c p The adjacent drone cluster, x bph is the storage allocation of block b by the adjacent drone cluster, ω bh is the weight of block b to the task drone node in the adjacent drone cluster, is the mission UAV u in the τ-1 time slot j The number of times block b is requested, is the mission UAV u in the τ-1 time slot j The total number of block requests initiated, B is the total number of block data, R jb For mission drone u j In the relevant block, u j The proportion of uploaded transactions to the total transactions in block b, T b is the current time, The time when block b was generated;

[0023] S22. Sort the storage requirements of all clusters for the first block of data in descending order. If the available storage capacity of the storage drone in the cluster with the highest storage requirement is exhausted, the first block of data is stored in the cluster with the next highest storage requirement, and so on, until the storage of the first block of data is completed.

[0024] S23, repeating steps S21-S22 to complete the storage of the second block of data;

[0025] S24. Repeat S23 in a loop to deploy each block data in the blockchain consensus network to the storage drones in turn until the available storage capacity of all storage drones is exhausted. At this time, the block data allocation plan for the storage drones in each cluster is obtained.

[0026] The S3 includes:

[0027] S31. Record the cost value of the cluster partition result and the block data allocation plan. The size of the cost value represents the quality of performance. The smaller the cost value, the better the performance.

[0028] The cost value is calculated using the following formula:

[0029]

[0030] In the above formula, is the cost value, The access cost generated by the task drone u reading the block b stored by the storage drone s, I ub is the interest value of task drone u in block b, U is the total number of task drones, B is the total number of block data, S is the total number of storage drones, Transmit block data D for mission drone u b The transmission power at D is the fixed energy consumption power used by the mission UAV u to maintain basic capabilities, b is the block data of block b, R u Transmit block data D for mission drone u b The data transmission rate at this time, λ is the probability of packet loss, d us is the physical distance between the mission UAV u and the storage UAV s, f is the signal frequency, and c is the speed of light;

[0031] S32, entering the first iteration, adjusting the preset hybrid similarity threshold, cluster similarity threshold, cluster internal hybrid similarity variance threshold, and hybrid similarity quantile, and obtaining a new cluster division result through the operation steps of S1;

[0032] S33: Adjust the weights of adjacent drone clusters, obtain a new block data allocation plan through the operation steps of S2, and calculate the cost value of the current iteration based on the new cluster division result and block data allocation plan;

[0033] S34. Compare the cost values before and after the iteration. If the performance after the iteration is better than the performance before the iteration, update the cluster division result and block data allocation plan, and record the performance after the iteration. Otherwise, revert to the cluster division result and block data allocation plan before the iteration.

[0034] S35. Determine whether the iteration termination condition is met. If not, return to step S32 to iteratively adjust the cluster division results and block data allocation plan again. If so, the algorithm terminates and outputs the optimal drone cluster network block data allocation plan.

[0035] In S11, the mixed similarity between the two mission drones is calculated using the following formula:

[0036] DS(u i ,u j )=S i，j +βD i,j ;

[0037]

[0038] In the above formula, DS(u i ,u j ) is the mission drone u i and mission drone u j The mixed similarity, S i，j For mission drone u i and mission drone u j The similarity of access patterns to blocks, β is the balance factor, D i，j For mission drone u i and mission drone u j The positional relationship between i，j For mission drone u i and mission drone u j Cosine similarity of block access frequency, τ k is the time slot of the access history, τ is the current time slot, is the parameter that affects the access history to the future time, δ is the attenuation parameter of the impact parameter, B is the total number of block data, P i，b 、P j，b They are respectively mission drone u i , mission drone u j The access frequency to block b, K is the number of all known time slots, They are respectively mission drone u i , mission drone u j In time slot τ k Geographical location within the

[0039] In S14, the cluster similarity between clusters is calculated using the following formula:

[0040]

[0041] In the above formula, E(c p , c q ) is cluster c p and c q The cluster similarity of |c q |、|c p | respectively cluster c p Cluster c qThe number of mission drones, For cluster c p Cluster Center With cluster c q Mission UAV in i The mixed similarity of For cluster c q Cluster Center With cluster c p Mission UAV in j Mixed similarity of

[0042] In S15, the internal mixed similarity variance of the cluster is calculated using the following formula:

[0043]

[0044] In the above formula, σ 2 is the variance of mixed similarity within the cluster, G is the total number of mission drones in the cluster, DS g (u a ,u j ) is the mission drone u j For cluster center u a Mixed similarity, DS G is the mean of the mixed similarity of all mission drones to the cluster center.

[0045] In a second aspect, the present invention proposes a block data allocation optimization system in a UAV cluster network, comprising a cluster partitioning module, a block data allocation module, and an iterative optimization module;

[0046] The cluster division module is used to divide the drones into multiple clusters, each cluster containing a storage drone and multiple mission drones;

[0047] The block data allocation module is used to formulate a block data allocation plan for the storage drones in each cluster based on the storage demand of the task drones in the cluster for block data;

[0048] The iterative optimization module is used to repeat the operations of the cluster division module and the block data allocation module, iteratively optimize the cluster division results and the block data allocation plan, and obtain the optimal drone cluster network block data allocation plan based on the performance of the cluster division results and the block data allocation plan of each iteration.

[0049] The cluster division module includes a cluster center selection unit, a drone joining unit, a cluster initialization unit, a cluster merging unit, a drone reallocation unit, and a storage drone allocation unit;

[0050] The cluster center selection unit is used to preset a mixed similarity threshold between task drones, calculate the mixed similarity between all unassigned task drones, select two task drones with a mixed similarity greater than the preset mixed similarity threshold and the highest mixed similarity, and use one of them as the cluster center;

[0051] The drone adding unit is used to calculate the mixed similarity between the remaining mission drones and the cluster center, and add the mission drones whose mixed similarity with the cluster center is higher than a preset mixed similarity threshold to the cluster;

[0052] The cluster initialization unit is used to cyclically repeat the operations of the cluster center selection unit and the drone joining unit until only task drones with a mixed similarity lower than a threshold value remain. Then, for each of these unassigned task drones, the mixed similarity between them and all cluster centers is calculated, and the drones are assigned to the cluster with the highest mixed similarity. At this point, the initialization of all clusters is completed.

[0053] The cluster merging unit is used to preset a cluster similarity threshold between two clusters based on the initialized clusters, calculate the cluster similarity between all clusters, and merge two clusters whose cluster similarity is higher than the threshold;

[0054] The drone reallocation unit is used to preset a mixed similarity variance threshold and a mixed similarity quantile within each cluster after the merger, count the mixed similarities between all task drones within each cluster and its cluster center, calculate the mixed similarity variance within the cluster, and for clusters with mixed similarity variance greater than the threshold, split out the task drones in the cluster whose mixed similarity with the cluster center is lower than the preset mixed similarity quantile, calculate the mixed similarity between each split task drone and other task drones, and reallocate these task drones to the cluster where other task drones with the highest mixed similarity are located;

[0055] The storage drone allocation unit is used to allocate storage drones equal to the number of clusters to each cluster after the task drones are redistributed, so that the task drones in the cluster can preferentially request data from the storage drone, thereby completing the division of all drone clusters.

[0056] The block data allocation module includes a storage demand calculation unit, a block data storage unit, a repetition unit, and a circulation unit;

[0057] The storage requirement calculation unit is used to calculate the storage requirement of each task drone in all clusters for the first block data, and the storage requirement of all task drones in a single cluster is accumulated to obtain the storage requirement of the current cluster for the first block data;

[0058] The storage requirement of each mission drone for block data is calculated using the following formula:

[0059]

[0060] In the above formula, d jb For cluster c p Mission UAV in j The storage requirement for block b, I jb For mission drone u j Interest value for block b, α h is the weight of the adjacent drone cluster, u ph is the mission drone in the adjacent drone cluster, C ph For cluster c p The adjacent drone cluster, x bph is the storage allocation of block b by the adjacent drone cluster, ω bh is the weight of block b to the task drone node in the adjacent drone cluster, is the mission UAV u in the τ-1 time slot j The number of times block b is requested, is the mission UAV u in the τ-1 time slot j The total number of block requests initiated, B is the total number of block data, R jb For mission drone u j In the relevant block, u j The proportion of uploaded transactions to the total transactions in block b, T b is the current time, The time when block b was generated;

[0061] The block data storage unit is used to sort the storage requirements of all clusters for the first block data in descending order. If the available storage capacity of the storage drone in the cluster with the highest storage requirement is exhausted, the first block data is stored in the cluster with the second highest storage requirement, and so on, until the storage of the first block data is completed.

[0062] The repeating unit is used to repeatedly store the demand calculation unit and the block data storage unit to complete the storage of the second block data;

[0063] The loop unit is used to cyclically repeat the operation of the repeating unit to sequentially deploy each block data in the blockchain consensus network to the storage drones until the available storage capacity of all storage drones is exhausted, at which time a block data allocation plan for the storage drones in each cluster is obtained.

[0064] The iterative optimization module includes a cost value recording unit, a new cluster division unit, a new block data allocation unit, a comparison unit, and a termination unit;

[0065] The initial state recording unit is used to record the cost value of the cluster division result and the block data allocation plan. The size of the cost value represents the quality of the performance. The smaller the cost value, the better the performance.

[0066] The cost value is calculated using the following formula:

[0067]

[0068] In the above formula, is the cost value, The access cost generated by the task drone u reading the block b stored by the storage drone s, I ub is the interest value of task drone u in block b, U is the total number of task drones, B is the total number of block data, S is the total number of storage drones, Transmit block data D for mission drone u b The transmission power at D is the fixed energy consumption power used by the mission UAV u to maintain basic capabilities, b is the block data of block b, R u Transmit block data D for mission drone u b The data transmission rate at this time, λ is the probability of packet loss, d us is the physical distance between the mission UAV u and the storage UAV s, f is the signal frequency, and c is the speed of light;

[0069] The new cluster division unit is used to enter the first iteration, adjust the preset hybrid similarity threshold, cluster similarity threshold, cluster internal hybrid similarity variance threshold, and hybrid similarity quantile, and obtain a new cluster division result through the operation steps of the cluster division module;

[0070] The new block data allocation unit is used to adjust the weights of adjacent drone clusters, obtain a new block data allocation plan through the operation steps of the block data allocation module, and calculate the cost value of the current iteration based on the new cluster division result and the block data allocation plan;

[0071] The optimization calculation unit is used to optimize the cluster division result and the block data allocation plan in each iteration, and calculate the cost value of the current iteration after completing the cluster division and allocation plan optimization;

[0072] The comparison unit is used to compare the cost values before and after the iteration. If the performance after the iteration is better than the performance before the iteration, the cluster division result and the block data allocation plan are updated and the performance after the iteration is recorded. Otherwise, the cluster division result and the block data allocation plan before the iteration are rolled back.

[0073] The termination unit is used to determine whether the iteration termination condition is met. If not, it returns to the new cluster division unit to iteratively adjust the cluster division result and block data allocation plan again. If so, the algorithm terminates and outputs the optimal drone cluster network block data allocation plan.

[0074] In the cluster center selection unit, the mixed similarity between two mission drones is calculated using the following formula:

[0075] DS(u i ,u j )=S i，j +βD i,j ;

[0076]

[0077] In the above formula, DS(u i ,u j ) is the mission drone u i and mission drone u j The mixed similarity, S i，j For mission drone u i and mission drone u j The similarity of access patterns to blocks, β is the balance factor, D i，j For mission drone u i and mission drone u j The positional relationship between i，j For mission drone u i and mission drone u j Cosine similarity of block access frequency, τ k is the time slot of the access history, τ is the current time slot, is the parameter that affects the access history to the future time, δ is the attenuation parameter of the impact parameter, B is the total number of block data, P i，b 、P j，b They are respectively mission drone u i , mission drone u j The access frequency to block b, K is the number of all known time slots, They are respectively mission drone u i , mission drone u j In time slot τ k Geographical location within the

[0078] In the cluster merging unit, the cluster similarity between clusters is calculated using the following formula:

[0079]

[0080] In the above formula, E(c p , cq ) is cluster c p and c q The cluster similarity of |c q |、|c p | respectively cluster c p Cluster c q The number of mission drones, For cluster c p Cluster Center With cluster c q Mission UAV in i The mixed similarity of For cluster c q Cluster Center With cluster c p Mission UAV in j Mixed similarity of

[0081] In the UAV reallocation unit, the variance of the mixed similarity within the cluster is calculated using the following formula:

[0082]

[0083] In the above formula, σ 2 is the variance of mixed similarity within the cluster, G is the total number of mission drones in the cluster, DS g (u a ,u j ) is the mission drone u j For cluster center u a Mixed similarity, DS G is the mean of the mixed similarity of all mission drones to the cluster center.

[0084] Compared with the prior art, the present invention has the following beneficial effects:

[0085] The present invention proposes a method and system for optimizing block data allocation in a drone swarm network. The method divides drones into multiple clusters, each of which contains a storage drone and multiple mission drones. Then, based on the storage demand for block data of the mission drones in the cluster, a block data allocation plan for the storage drones in each cluster is formulated. Finally, by repeating the cluster division and block data allocation operations, the cluster division results and block data allocation plan are iteratively optimized, and the optimal drone swarm network block data allocation plan is obtained based on the performance of the cluster division results and block data allocation plan in each iteration. On the one hand, this method designs a drone network architecture under the background of blockchain storage requirements. The architecture divides drones into mission drones and storage drones, and formulates a block data allocation plan for storage drones in each cluster according to the storage requirements of mission drones for block data. As the mission drones execute their tasks, the key data that needs to be obtained can be requested directly from the storage drones in the air with lower latency and energy consumption than from air to ground, providing fast and reliable data access services for mission drones; on the other hand, this method iteratively optimizes the cluster division results and block data allocation plan to obtain a nearly optimal drone cluster network block data allocation plan, realizing block data allocation within storage drones guided by storage and access efficiency, effectively reducing the communication cost in the drone network, and promoting the application of blockchain systems in drone networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] Figure 1 The figure is an overall flow chart of the method of the present invention.

[0087] Figure 2 This is a schematic diagram of the structure of the drone cluster network described in Example 1.

[0088] Figure 3 1 is a structural diagram of the system of the present invention. DETAILED DESCRIPTION

[0089] The present invention will be further described in detail below with reference to specific embodiments and the accompanying drawings.

[0090] The present invention proposes a method and system for optimizing block data allocation in a drone cluster network. Under the background of blockchain storage requirements, drones are divided into two roles: mission drones and storage drones, which jointly participate in the blockchain network and share perception data. A drone clustering algorithm based on hybrid similarity is used to divide drones into multiple clusters, each of which contains one storage drone and multiple mission drones. Based on the storage requirements of the mission drones in the cluster for block data and guided by storage and access efficiency, a storage allocation algorithm based on proximity perception optimization is used to formulate a block data allocation plan for the storage drones in each cluster. A cluster and block allocation algorithm based on iterative optimization is used to iteratively optimize the cluster division results and the block data allocation plan, and the optimal drone cluster network block data allocation plan is obtained based on the performance of each iterative cluster division result and the block data allocation plan, thereby effectively reducing the communication cost in the drone network and promoting the application of blockchain systems in drone networks.

[0091] Example 1:

[0092] like Figure 1 As shown in FIG, a method for optimizing block data allocation in a drone cluster network is performed in the following steps:

[0093] 1. In the context of blockchain storage requirements, drones are divided into two categories: mission drones and storage drones, which jointly participate in the blockchain network and share perception data;

[0094] The network structure of drone cluster is as follows Figure 2 As shown in Figure 1, it includes three layers: ground data center, perceived environment, and drone network. The drone network divides drones into mission drones and storage drones according to their own functions and storage resource size. Mission drones are responsible for performing various tasks and requesting data to support decision-making when needed. The mission drone set is defined as The storage drone is responsible for storing and managing data, providing fast and reliable data access services for mission drones. The storage drone collection is defined as The corresponding storage resource collection is The ground data center participates in the update and maintenance of the blockchain status as a consensus node, and together with the mission drone and storage drone, it forms the consensus network of the blockchain. The key data perceived by the mission drone is processed and stored safely and reliably on the ground. The block data set is defined as

[0095] The process of sharing perception data in a drone swarm network includes:

[0096] Raw data generation: In a drone network, mission drones perform environmental perception in clusters. When each mission drone senses key data, it encrypts and transmits it to a ground terminal for further processing.

[0097] Data processing: After receiving key data information, ground terminal service facilities such as base stations and servers encapsulate it into a transaction. When the cumulative transaction volume reaches a certain threshold, multiple transactions are packaged into a block. The block is uploaded to the blockchain through the consensus mechanism among multiple ground nodes, updating the global data information status in the blockchain consensus network.

[0098] Data storage: The complete data information in the blockchain consensus network is strategically distributed to multiple storage drones in the air;

[0099] Data access: When a mission drone needs to access certain data information during a mission, it obtains block data through an aerial storage drone and verifies the validity and legality of the data according to the block data characteristics, without the need for frequent communication with ground equipment, thereby reducing the latency and energy consumption caused by the mission drone's data request.

[0100] 2. Use the hybrid similarity-based drone clustering algorithm to divide the drones into multiple clusters, each of which contains a storage drone and multiple mission drones;

[0101] Preset the mixed similarity threshold between task drones based on experience and the expected number of clusters, calculate the mixed similarity between all unassigned task drones, select the two task drones with the highest mixed similarity that is greater than the preset mixed similarity threshold, and use one of them as the cluster center;

[0102] The mixed similarity between task drones is obtained by calculating the cosine similarity of the physical distance and visit frequency between task drones, which is used to guide cluster division:

[0103] DS(u i ,u j )=S i，j +βD i，j ;

[0104] Specifically, since drones’ access to data changes dynamically, it is not accurate to directly use the cosine similarity of access frequency within a certain period of time to evaluate the logical relationship between drones. Therefore, by introducing the Hawkes model to model the timeliness and self-excitation effect of access frequency, we consider that the occurrence of an event will affect the probability of the event occurring in the future, and more accurately analyze the logical relationship between drones. Therefore, the access pattern similarity is expressed as:

[0105]

[0106] The position relationship between the mission drones is represented by the average distance between the two mission drones in all known time slots:

[0107]

[0108] In the above formula, DS(u i ,u j ) is the mission drone u i and mission drone u j The mixed similarity, S i，j For mission drone u i and mission drone u j The similarity of access patterns to blocks, β is the balance factor, D i，j For mission drone u i and mission drone u j The positional relationship between i，j For mission drone u i and mission drone u j Cosine similarity of block access frequency, τ k is the time slot of the access history, τ is the current time slot, is the parameter that affects the access history to the future time, δ is the attenuation parameter of the impact parameter, B is the total number of block data, P i，b 、P j，b They are respectively mission drone u i , mission drone u j The access frequency to block b, K is the number of all known time slots, They are respectively mission drone u i , mission drone u j In time slot τ k Geographical location within the

[0109] Calculate the mixed similarity between the remaining mission drones and the cluster center, and add the mission drones whose mixed similarity with the cluster center is higher than the preset mixed similarity threshold to the cluster;

[0110] Repeat the above operation until only task drones with mixed similarity below the threshold remain. Then, for these unassigned task drones, calculate their mixed similarity with all cluster centers and assign them to the cluster with the highest mixed similarity. At this point, the initialization of all clusters is completed.

[0111] Based on the initialized clusters, a cluster similarity threshold between two clusters is preset, the cluster similarity between all clusters is calculated, and two clusters with cluster similarity higher than the threshold are merged to improve cluster cohesion;

[0112] The cluster similarity between clusters is calculated using the following formula:

[0113]

[0114] In the above formula, E(c p , c q ) is cluster c p and c q The cluster similarity of |c q |、|c p | respectively cluster c p Cluster c q The number of mission drones, For cluster c p Cluster Center With cluster c q Mission UAV in i The mixed similarity of For cluster c q Cluster Center With cluster c p Mission UAV in j Mixed similarity of

[0115] To improve the granularity of the clustering solution, for each merged cluster, a mixed similarity variance threshold and a 25% quantile of mixed similarity are preset within each cluster. The mixed similarity of all mission drones within each cluster to the cluster center is counted, and the mixed similarity variance within the cluster is calculated. For clusters with mixed similarity variance greater than the threshold, the mission drones in the cluster whose mixed similarity with the cluster center is lower than the preset mixed similarity quantile are split out. The mixed similarity of each split mission drone with other mission drones is calculated, and these mission drones are reallocated to the cluster where the other mission drones with the highest mixed similarity are located.

[0116] The variance of the mixed similarity within the cluster is calculated using the following formula:

[0117]

[0118] In the above formula, σ 2 is the variance of mixed similarity within the cluster, G is the total number of mission drones in the cluster, DS g (u a ,u j ) is the mission drone u j For cluster center u a Mixed similarity, DS G is the mean of the mixed similarity of all mission UAVs to the cluster center;

[0119] After the task drones are redistributed, storage drones equal to the number of clusters are allocated to each cluster, so that the task drones in the cluster can make data requests to the storage drone first, completing the division of all drone clusters.

[0120] 3. Based on the storage requirements of the task drones in the cluster for block data, and guided by storage and access efficiency, a block data allocation plan for the storage drones in each cluster is formulated using a storage allocation algorithm based on proximity perception optimization;

[0121] Calculate the storage requirement of the first block of data for each mission drone in all clusters. Add up the storage requirements of all mission drones in a single cluster to get the storage requirement of the first block of data for the current cluster.

[0122] The storage demand of the task drone node for block data can be expressed as the drone's interest value in the block as the block allocation scheme changes. The storage demand of each task drone for block data is calculated using the following formula:

[0123]

[0124] The interest value of a drone in a block is determined by the frequency with which the mission drone visits the block and the correlation between the drone and the block. That is, the higher the frequency with which the mission drone visits the block and the greater the correlation between the block and the drone, the greater the interest value, while the longer the block exists, the lower the interest value:

[0125]

[0126] In the above formula, d jb For cluster c p Mission UAV in j The storage requirement for block b, I jb For mission drone u j Interest value for block b, α h is the weight of the adjacent UAV cluster. By increasing this weight, the influence of the adjacent cluster on the current block demand can be strengthened and the cost can be reduced. u ph is the mission drone in the adjacent drone cluster, c ph For cluster c p The adjacent drone cluster, x bph is the storage allocation of block b by the adjacent drone cluster, ω bh is the weight of block b to the task drone node in the adjacent drone cluster, x bp For cluster c p For block b, the binary variable storing the allocation status is x. bp =1 means block b is assigned to cluster c pThe storage drone is used for storage, whereas x bp =0 means cluster c p There is no storage block b, is the mission UAV u in the τ-1 time slot j The number of times block b is requested, is the mission UAV u in the τ-1 time slot j The total number of block requests initiated, B is the total number of block data, R jb For mission drone u j In the relevant block, u j The proportion of uploaded transactions to the total transactions in block b, T b is the current time, The time when block b was generated; for cluster c p Mission UAV in j For example, if the storage drone in its cluster stores block b, the drone can access the block data at the minimum cost, so its storage demand for block b is x. bp is 0; and if the mission drone u j The cluster where the block is located does not store block b. Its storage demand is not only related to the weight of the block data on it, but also to the distribution of block data in neighboring clusters.

[0127] Sort all clusters by their storage requirements for the first block of data in descending order. If the available storage capacity of the storage drone in the cluster with the highest storage requirement is exhausted, the first block of data will be stored in the cluster with the next highest storage requirement, and so on, until the storage of the first block of data is completed.

[0128] Repeat the above operations to complete the storage of the second block data, the third block data, and so on, and deploy each block data in the blockchain consensus network to the storage drones in turn until the available storage capacity of all storage drones is exhausted. At this time, the block data allocation plan for the storage drones in each cluster is obtained.

[0129] 4. Repeat steps 2 and 3, and use the cluster and block allocation algorithm based on iterative optimization to iteratively optimize the cluster division results and block data allocation scheme. Based on the performance of each iterative cluster division result and block data allocation scheme, the optimal UAV cluster network block data allocation scheme is obtained to achieve global optimization.

[0130] Initialize the number of iterations, performance improvement threshold, and stagnation limit to control the execution of the algorithm. Record the initial cluster division results and block data allocation plan, and calculate the cost value of the initial state. The size of the cost value represents the performance quality. The smaller the cost value, the better the performance.

[0131] The cost value is calculated using the following formula:

[0132]

[0133] In the above formula, is the cost value, The access cost generated by the task drone u reading the block b stored by the storage drone s, I ub is the interest value of task drone u in block b, U is the total number of task drones, B is the total number of block data, S is the total number of storage drones, Transmit block data D for mission drone u b The transmission power at D is the fixed energy consumption power used by the mission UAV u to maintain basic capabilities, b is the block data of block b, R u Transmit block data D for mission drone u b The data transmission rate at this time, λ is the probability of packet loss, d us is the physical distance between the mission UAV u and the storage UAV s, f is the signal frequency, and c is the speed of light;

[0134] Enter the first iteration, adjust the preset hybrid similarity threshold, cluster similarity threshold, cluster internal hybrid similarity variance threshold, and hybrid similarity quantile, and obtain a new cluster division result through the operation in step 2;

[0135] Adjust the weight α of adjacent drone clusters h , obtain the new block data allocation plan through the operation of step 3, and calculate the cost value of the current iteration based on the new cluster division result and block data allocation plan;

[0136] Compare the cost values before and after the iteration. If the performance after the iteration is better than the performance before the iteration, update the cluster division result and block data allocation plan, record the performance after the iteration, and reset the stagnation count. Otherwise, roll back to the cluster division result and block data allocation plan before the iteration and increase the stagnation count to ensure that performance does not degrade due to optimization drift and optimize towards the lowest cost.

[0137] The cluster division results and block data allocation scheme are iteratively adjusted again. When the number of iterations reaches the upper limit, the performance improvement is lower than the threshold, or the number of stagnations exceeds the limit, the algorithm terminates and outputs the optimal UAV cluster network block data allocation scheme.

[0138] Example 2:

[0139] like Figure 3 As shown, a block data allocation optimization system in a UAV cluster network includes a cluster division module, a block data allocation module, and an iterative optimization module;

[0140] The cluster division module is used to divide the drones into multiple clusters, each cluster containing a storage drone and multiple mission drones;

[0141] The block data allocation module is used to formulate a block data allocation plan for the storage drones in each cluster based on the storage demand of the task drones in the cluster for block data;

[0142] The iterative optimization module is used to repeat the operations of the cluster division module and the block data allocation module, iteratively optimize the cluster division results and the block data allocation plan, and obtain the optimal drone cluster network block data allocation plan based on the performance of the cluster division results and the block data allocation plan of each iteration.

[0143] The cluster division module includes a cluster center selection unit, a drone joining unit, a cluster initialization unit, a cluster merging unit, a drone reallocation unit, and a storage drone allocation unit;

[0144] The cluster center selection unit is used to preset a mixed similarity threshold between task drones, calculate the mixed similarity between all unassigned task drones, select two task drones with a mixed similarity greater than the preset mixed similarity threshold and the highest mixed similarity, and use one of them as the cluster center;

[0145] The drone adding unit is used to calculate the mixed similarity between the remaining mission drones and the cluster center, and add the mission drones whose mixed similarity with the cluster center is higher than a preset mixed similarity threshold to the cluster;

[0146] The cluster initialization unit is used to cyclically repeat the operations of the cluster center selection unit and the drone joining unit until only task drones with a mixed similarity lower than a threshold value remain. Then, for each of these unassigned task drones, the mixed similarity between them and all cluster centers is calculated, and the drones are assigned to the cluster with the highest mixed similarity. At this point, the initialization of all clusters is completed.

[0147] The cluster merging unit is used to preset a cluster similarity threshold between two clusters based on the initialized clusters, calculate the cluster similarity between all clusters, and merge two clusters whose cluster similarity is higher than the threshold;

[0148] The drone reallocation unit is used to preset a mixed similarity variance threshold and a mixed similarity quantile within each cluster after the merger, count the mixed similarities between all task drones within each cluster and its cluster center, calculate the mixed similarity variance within the cluster, and for clusters with mixed similarity variance greater than the threshold, split out the task drones in the cluster whose mixed similarity with the cluster center is lower than the preset mixed similarity quantile, calculate the mixed similarity between each split task drone and other task drones, and reallocate these task drones to the cluster where other task drones with the highest mixed similarity are located;

[0149] The storage drone allocation unit is used to allocate storage drones equal to the number of clusters to each cluster after the task drones are redistributed, so that the task drones in the cluster can preferentially request data from the storage drone, thereby completing the division of all drone clusters.

[0150] In the cluster center selection unit, the mixed similarity between two mission drones is calculated using the following formula:

[0151] DS(u i ,u j )=S i，j +βD i,j ,

[0152]

[0153] In the above formula, DS(u i ,u j ) is the mission drone u i and mission drone u j The mixed similarity, S i，j For mission drone u i and mission drone u j The similarity of access patterns to blocks, β is the balance factor, D i，j For mission drone u i and mission drone u j The positional relationship between i，j For mission drone u i and mission drone u j Cosine similarity of block access frequency, τ k is the time slot of the access history, τ is the current time slot, is the parameter that affects the access history to the future time, δ is the attenuation parameter of the impact parameter, B is the total number of block data, P i，b 、P j，b They are respectively mission drone u i , mission drone u jThe access frequency to block b, K is the number of all known time slots, They are respectively mission drone u i , mission drone u j In time slot τ k Geographical location within the

[0154] In the cluster merging unit, the cluster similarity between clusters is calculated using the following formula:

[0155]

[0156] In the above formula, E(c p , c q ) is cluster c p and c q The cluster similarity of |c q |、|c p | respectively cluster c p Cluster c q The number of mission drones, For cluster c p Cluster Center With cluster c q Mission UAV in i The mixed similarity of For cluster c q Cluster Center With cluster c p Mission UAV in j Mixed similarity of

[0157] In the UAV reallocation unit, the variance of the mixed similarity within the cluster is calculated using the following formula:

[0158]

[0159] In the above formula, σ 2 is the variance of mixed similarity within the cluster, G is the total number of mission drones in the cluster, DS g (u a ,u j ) is the mission drone u j For cluster center u a Mixed similarity, DS G is the mean of the mixed similarity of all mission drones to the cluster center.

[0160] The block data allocation module includes a storage demand calculation unit, a block data storage unit, a repetition unit, and a circulation unit;

[0161] The storage requirement calculation unit is used to calculate the storage requirement of each task drone in all clusters for the first block data, and the storage requirement of all task drones in a single cluster is accumulated to obtain the storage requirement of the current cluster for the first block data;

[0162] The storage requirement of each mission drone for block data is calculated using the following formula:

[0163]

[0164] In the above formula, d jb For cluster c p Mission UAV in j The storage requirement for block b, I jb For mission drone u j Interest value for block b, α h is the weight of the adjacent drone cluster, u ph is the mission drone in the adjacent drone cluster, C ph For cluster c p The adjacent drone cluster, x bph is the storage allocation of block b by the adjacent drone cluster, ω bh is the weight of block b to the task drone node in the adjacent drone cluster, is the mission UAV u in the τ-1 time slot j The number of times block b is requested, is the mission UAV u in the τ-1 time slot j The total number of block requests initiated, B is the total number of block data, R jb For mission drone u j In the relevant block, u j The proportion of uploaded transactions to the total transactions in block b, T b is the current time, The time when block b was generated;

[0165] The block data storage unit is used to sort the storage requirements of all clusters for the first block data in descending order. If the available storage capacity of the storage drone in the cluster with the highest storage requirement is exhausted, the first block data is stored in the cluster with the second highest storage requirement, and so on, until the storage of the first block data is completed.

[0166] The repeating unit is used to repeatedly store the demand calculation unit and the block data storage unit to complete the storage of the second block data;

[0167] The loop unit is used to cyclically repeat the operation of the repeating unit to sequentially deploy each block data in the blockchain consensus network to the storage drones until the available storage capacity of all storage drones is exhausted, at which time a block data allocation plan for the storage drones in each cluster is obtained.

[0168] The iterative optimization module includes a cost value recording unit, a new cluster division unit, a new block data allocation unit, a comparison unit, and a termination unit;

[0169] The initial state recording unit is used to record the cost value of the cluster division result and the block data allocation plan. The size of the cost value represents the quality of the performance. The smaller the cost value, the better the performance.

[0170] The cost value is calculated using the following formula:

[0171]

[0172] In the above formula, is the cost value, The access cost generated by the task drone u reading the block b stored by the storage drone s, I ub is the interest value of task drone u in block b, U is the total number of task drones, B is the total number of block data, S is the total number of storage drones, Transmit block data D for mission drone u b The transmission power at D is the fixed energy consumption power used by the mission UAV u to maintain basic capabilities, b is the block data of block b, R u Transmit block data D for mission drone u b The data transmission rate at this time, λ is the probability of packet loss, d us is the physical distance between the mission UAV u and the storage UAV s, f is the signal frequency, and c is the speed of light;

[0173] The new cluster division unit is used to enter the first iteration, adjust the preset hybrid similarity threshold, cluster similarity threshold, cluster internal hybrid similarity variance threshold, and hybrid similarity quantile, and obtain a new cluster division result through the operation steps of the cluster division module;

[0174] The new block data allocation unit is used to adjust the weights of adjacent drone clusters, obtain a new block data allocation plan through the operation steps of the block data allocation module, and calculate the cost value of the current iteration based on the new cluster division result and the block data allocation plan;

[0175] The optimization calculation unit is used to optimize the cluster division result and the block data allocation plan in each iteration, and calculate the cost value of the current iteration after completing the cluster division and allocation plan optimization;

[0176] The comparison unit is used to compare the cost values before and after the iteration. If the performance after the iteration is better than the performance before the iteration, the cluster division result and the block data allocation plan are updated and the performance after the iteration is recorded. Otherwise, the cluster division result and the block data allocation plan before the iteration are rolled back.

[0177] The termination unit is used to determine whether the iteration termination condition is met. If not, it returns to the new cluster division unit to iteratively adjust the cluster division result and block data allocation plan again. If so, the algorithm terminates and outputs the optimal drone cluster network block data allocation plan.

Claims

1. A method for optimizing block data allocation in a drone cluster network, characterized in that: The method comprises: S1. Divide the drones into multiple clusters, each of which contains a storage drone and multiple mission drones; S2. Based on the storage requirements of the task drones in the cluster for block data, formulate a block data allocation plan for the storage drones in each cluster; S3. Repeat the operations of steps S1-S2 to iteratively optimize the cluster division results and block data allocation scheme, and obtain the optimal drone cluster network block data allocation scheme based on the performance of the cluster division results and block data allocation scheme of each iteration.

2. The method for optimizing block data allocation in a drone cluster network according to claim 1, characterized in that: Said S1 comprises: S11, presetting a mixed similarity threshold between task drones, calculating the mixed similarity between all unassigned task drones, selecting two task drones with a mixed similarity greater than the preset mixed similarity threshold and the highest mixed similarity, and selecting one of them as the cluster center; S12, calculating the mixed similarity between the remaining mission drones and the cluster center, and adding mission drones whose mixed similarity with the cluster center is higher than a preset mixed similarity threshold to the cluster; S13, repeating steps S11-S12 until only task drones with mixed similarity below the threshold remain. Then, for each of these unassigned task drones, calculate their mixed similarity with all cluster centers and assign them to the cluster with the highest mixed similarity. At this point, the initialization of all clusters is completed. S14, based on the initialized clusters, presetting a cluster similarity threshold between two clusters, calculating cluster similarities between all clusters, and merging two clusters whose cluster similarities are higher than the threshold; S15. For each merged cluster, preset a hybrid similarity variance threshold and hybrid similarity quantile within each cluster, count the hybrid similarities between all mission drones within each cluster and its cluster center, calculate the hybrid similarity variance within the cluster, and for clusters with hybrid similarity variance greater than the threshold, split out the mission drones in the cluster whose hybrid similarity with the cluster center is lower than the preset hybrid similarity quantile, calculate the hybrid similarity between each split mission drone and other mission drones, and reallocate these mission drones to the cluster containing the other mission drones with the highest hybrid similarity. S16. After the task drones are redistributed, storage drones equal in number to the number of clusters are allocated to each cluster, so that the task drones in the cluster can make data requests to the storage drones first, thus completing the division of all drone clusters.

3. The method for optimizing block data allocation in a drone cluster network according to claim 1, characterized in that: The S2 includes: S21. Calculate the storage requirement of each mission drone in all clusters for the first block of data. Add up the storage requirements of all mission drones in a single cluster to obtain the storage requirement of the current cluster for the first block of data. The storage requirement of each mission drone for block data is calculated using the following formula: In the above formula, d jb For cluster c p Mission UAV in j The storage requirement for block b, I jb For mission drone u j Interest value for block b, α h is the weight of the adjacent drone cluster, u ph is the mission drone in the adjacent drone cluster, C ph For cluster c p The adjacent drone cluster, x bph is the storage allocation of block b by the adjacent drone cluster, ω bh is the weight of block b to the mission drone in the adjacent drone cluster, is the mission UAV u in the τ-1 time slot j The number of times block b is requested, is the mission UAV u in the τ-1 time slot j The total number of block requests initiated, B is the total number of block data, R jb For mission drone u j In the relevant block, u j The proportion of uploaded transactions to the total transactions in block b, T b is the current time, The time when block b was generated; S22. Sort the storage requirements of all clusters for the first block of data in descending order. If the available storage capacity of the storage drone in the cluster with the highest storage requirement is exhausted, the first block of data is stored in the cluster with the next highest storage requirement, and so on, until the storage of the first block of data is completed. S23, repeating steps S21-S22 to complete the storage of the second block of data; S24. Repeat S23 in a loop to deploy each block data in the blockchain consensus network to the storage drones in turn until the available storage capacity of all storage drones is exhausted. At this time, the block data allocation plan for the storage drones in each cluster is obtained.

4. The method for optimizing block data allocation in a drone cluster network according to claim 1, characterized in that: The S3 includes: S31. Record the cost value of the cluster partition result and the block data allocation plan. The size of the cost value represents the quality of performance. The smaller the cost value, the better the performance. The cost value is calculated using the following formula: In the above formula, is the cost value, The access cost generated by the task drone u reading the block b stored by the storage drone s, I ub is the interest value of task drone u in block b, U is the total number of task drones, B is the total number of block data, S is the total number of storage drones, Transmit block data D for mission drone u b The transmission power at D is the fixed energy consumption power used by the mission UAV u to maintain basic capabilities, b is the block data of block b, R u Transmit block data D for mission drone u b The data transmission rate at this time, λ is the probability of packet loss, d us is the physical distance between the mission UAV u and the storage UAV s, f is the signal frequency, and c is the speed of light; S32, entering the first iteration, adjusting the preset hybrid similarity threshold, cluster similarity threshold, cluster internal hybrid similarity variance threshold, and hybrid similarity quantile, and obtaining a new cluster division result through the operation steps of S1; S33: Adjust the weights of adjacent drone clusters, obtain a new block data allocation plan through the operation steps of S2, and calculate the cost value of the current iteration based on the new cluster division result and block data allocation plan; S34. Compare the cost values before and after the iteration. If the performance after the iteration is better than the performance before the iteration, update the cluster division result and block data allocation plan, and record the performance after the iteration. Otherwise, revert to the cluster division result and block data allocation plan before the iteration. S35. Determine whether the iteration termination condition is met. If not, return to step S32 to iteratively adjust the cluster division results and block data allocation plan again. If so, the algorithm terminates and outputs the optimal drone cluster network block data allocation plan.

5. The method for optimizing block data allocation in a drone cluster network according to claim 2, characterized in that: In S11, the mixed similarity between the two mission drones is calculated using the following formula: DS(u i ,u j )=S i，j +βD i,j ; In the above formula, DS(u i ,u j ) is the mission drone u i and mission drone u j The mixed similarity, S i，j For mission drone u i and mission drone u j The similarity of access patterns to blocks, β is the balance factor, D i，j For mission drone u i and mission drone u j The positional relationship between i，j For mission drone u i and mission drone u j Cosine similarity of block access frequency, τ k is the time slot of the access history, τ is the current time slot, is the parameter that affects the access history to the future time, δ is the attenuation parameter of the impact parameter, B is the total number of block data, P i，b 、P j，b They are respectively mission drone u i , mission drone u j The access frequency to block b, K is the number of all known time slots, They are respectively mission drone u i , mission drone u j In time slot τ k Geographical location within the In S14, the cluster similarity between clusters is calculated using the following formula: In the above formula, E(c p , c q ) is cluster c p and c q The cluster similarity of |c q |、|c p | respectively cluster c p Cluster c q The number of mission drones, For cluster c p Cluster Center With cluster c q Mission UAV in i The mixed similarity of For cluster c q Cluster Center With cluster c p Mission UAV in j Mixed similarity of In S15, the internal mixed similarity variance of the cluster is calculated using the following formula: In the above formula, σ 2 is the variance of mixed similarity within the cluster, G is the total number of mission drones in the cluster, DS g (u a ,u j ) is the mission drone u j For cluster center u a Mixed similarity, DS G is the mean of the mixed similarity of all mission drones to the cluster center.

6. A block data allocation optimization system in a drone cluster network, characterized by: The system includes a cluster partitioning module, a block data allocation module, and an iterative optimization module; The cluster division module is used to divide the drones into multiple clusters, each cluster containing a storage drone and multiple mission drones; The block data allocation module is used to formulate a block data allocation plan for the storage drones in each cluster based on the storage demand of the task drones in the cluster for block data; The iterative optimization module is used to repeat the operations of the cluster division module and the block data allocation module, iteratively optimize the cluster division results and the block data allocation plan, and obtain the optimal drone cluster network block data allocation plan based on the performance of the cluster division results and the block data allocation plan of each iteration.

7. The block data allocation optimization system in a UAV cluster network according to claim 6 is characterized in that: The cluster division module includes a cluster center selection unit, a drone joining unit, a cluster initialization unit, a cluster merging unit, a drone reallocation unit, and a storage drone allocation unit; The cluster center selection unit is used to preset a mixed similarity threshold between task drones, calculate the mixed similarity between all unassigned task drones, select two task drones with a mixed similarity greater than the preset mixed similarity threshold and the highest mixed similarity, and use one of them as the cluster center; The drone adding unit is used to calculate the mixed similarity between the remaining mission drones and the cluster center, and add the mission drones whose mixed similarity with the cluster center is higher than a preset mixed similarity threshold to the cluster; The cluster initialization unit is used to cyclically repeat the operations of the cluster center selection unit and the drone joining unit until only task drones with a mixed similarity lower than a threshold value remain. Then, for each of these unassigned task drones, the mixed similarity between them and all cluster centers is calculated, and the drones are assigned to the cluster with the highest mixed similarity. At this point, the initialization of all clusters is completed. The cluster merging unit is used to preset a cluster similarity threshold between two clusters based on the initialized clusters, calculate the cluster similarity between all clusters, and merge two clusters whose cluster similarity is higher than the threshold; The drone reallocation unit is used to preset a mixed similarity variance threshold and a mixed similarity quantile within each cluster after the merger, count the mixed similarities between all task drones within each cluster and its cluster center, calculate the mixed similarity variance within the cluster, and for clusters with mixed similarity variance greater than the threshold, split out the task drones in the cluster whose mixed similarity with the cluster center is lower than the preset mixed similarity quantile, calculate the mixed similarity between each split task drone and other task drones, and reallocate these task drones to the cluster where other task drones with the highest mixed similarity are located; The storage drone allocation unit is used to allocate storage drones equal to the number of clusters to each cluster after the task drones are redistributed, so that the task drones in the cluster can preferentially request data from the storage drone, thereby completing the division of all drone clusters.

8. The block data allocation optimization system in a UAV cluster network according to claim 6 is characterized in that: The block data allocation module includes a storage demand calculation unit, a block data storage unit, a repetition unit, and a circulation unit; The storage requirement calculation unit is used to calculate the storage requirement of each task drone in all clusters for the first block data, and the storage requirement of all task drones in a single cluster is accumulated to obtain the storage requirement of the current cluster for the first block data; The storage requirement of each mission drone for block data is calculated using the following formula: In the above formula, d jb For cluster c p Mission UAV in j The storage requirement for block b, I jb For mission drone u j Interest value for block b, α h is the weight of the adjacent drone cluster, u ph is the mission drone in the adjacent drone cluster, c ph For cluster c p The adjacent drone cluster, x bph is the storage allocation of block b by the adjacent drone cluster, ω bh is the weight of block b to the task drone node in the adjacent drone cluster, is the mission UAV u in the τ-1 time slot j The number of times block b is requested, is the mission UAV u in the τ-1 time slot j The total number of block requests initiated, B is the total number of block data, R jb For mission drone u j In the relevant block, u j The proportion of uploaded transactions to the total transactions in block b, T b is the current time, The time when block b was generated; The block data storage unit is used to sort the storage requirements of all clusters for the first block data in descending order. If the available storage capacity of the storage drone in the cluster with the highest storage requirement is exhausted, the first block data is stored in the cluster with the second highest storage requirement, and so on, until the storage of the first block data is completed. The repeating unit is used to repeatedly store the demand calculation unit and the block data storage unit to complete the storage of the second block data; The loop unit is used to cyclically repeat the operation of the repeating unit to sequentially deploy each block data in the blockchain consensus network to the storage drones until the available storage capacity of all storage drones is exhausted, at which time a block data allocation plan for the storage drones in each cluster is obtained.

9. The block data allocation optimization system in a UAV cluster network according to claim 6 is characterized in that: The iterative optimization module includes a cost value recording unit, a new cluster division unit, a new block data allocation unit, a comparison unit, and a termination unit; The initial state recording unit is used to record the cost value of the cluster division result and the block data allocation plan. The size of the cost value represents the quality of the performance. The smaller the cost value, the better the performance. The cost value is calculated using the following formula: In the above formula, is the cost value, The access cost generated by the task drone u reading the block b stored by the storage drone s, I ub is the interest value of task drone u in block b, U is the total number of task drones, B is the total number of block data, S is the total number of storage drones, Transmit block data D for mission drone u b The transmission power at D is the fixed energy consumption power used by the mission UAV u to maintain basic capabilities, b is the block data of block b, R u Transmit block data D for mission drone u b The data transmission rate at this time, λ is the probability of packet loss, d us is the physical distance between the mission UAV u and the storage UAV s, f is the signal frequency, and c is the speed of light; The new cluster division unit is used to enter the first iteration, adjust the preset hybrid similarity threshold, cluster similarity threshold, cluster internal hybrid similarity variance threshold, and hybrid similarity quantile, and obtain a new cluster division result through the operation steps of the cluster division module; The new block data allocation unit is used to adjust the weights of adjacent drone clusters, obtain a new block data allocation plan through the operation steps of the block data allocation module, and calculate the cost value of the current iteration based on the new cluster division result and the block data allocation plan; The optimization calculation unit is used to optimize the cluster division result and the block data allocation plan in each iteration, and calculate the cost value of the current iteration after completing the cluster division and allocation plan optimization; The comparison unit is used to compare the cost values before and after the iteration. If the performance after the iteration is better than the performance before the iteration, the cluster division result and the block data allocation plan are updated and the performance after the iteration is recorded. Otherwise, the cluster division result and the block data allocation plan before the iteration are rolled back. The termination unit is used to determine whether the iteration termination condition is met. If not, it returns to the new cluster division unit to iteratively adjust the cluster division result and block data allocation plan again. If so, the algorithm terminates and outputs the optimal drone cluster network block data allocation plan.

10. The block data allocation optimization system in a UAV cluster network according to claim 7, characterized in that: In the cluster center selection unit, the mixed similarity between two mission drones is calculated using the following formula: DS(u i ,u j )=S i，j +βD i,j ; In the above formula, DS(u i ,u j ) is the mission drone u i and mission drone u j The mixed similarity, S i，j For mission drone u i and mission drone u j The similarity of access patterns to blocks, β is the balance factor, D i，j For mission drone u i and mission drone u j The positional relationship between i，j For mission drone u i and mission drone u j Cosine similarity of block access frequency, τ k is the time slot of the access history, τ is the current time slot, is the parameter that affects the access history to the future time, δ is the attenuation parameter of the impact parameter, B is the total number of block data, P i，b 、P j，b They are respectively mission drone u i , mission drone u j The access frequency to block b, K is the number of all known time slots, They are respectively mission drone u i , mission drone u j In time slot τ k Geographical location within the In the cluster merging unit, the cluster similarity between clusters is calculated using the following formula: In the above formula, E(c p , c q ) is cluster c p and c q The cluster similarity of |c q |、|c p | respectively cluster c p Cluster c q The number of mission drones, For cluster c p Cluster Center With cluster c q Mission UAV in i The mixed similarity of For cluster c q Cluster Center With cluster c p Mission UAV in j Mixed similarity of In the UAV reallocation unit, the variance of the mixed similarity within the cluster is calculated using the following formula: In the above formula, σ 2 is the variance of mixed similarity within the cluster, G is the total number of mission drones in the cluster, DS g (u a ,u j ) is the mission drone u j For cluster center u a Mixed similarity, DS G is the mean of the mixed similarity of all mission drones to the cluster center.

Images