A smart contract driven inspection formation dynamic reconstruction method and system

Abstract

This invention discloses a smart contract-driven dynamic reconfiguration method and system for inspection convoys, relating to the field of blockchain smart contract technology. The method includes: a smart contract that, based on the physical state vectors uploaded by each inspection vehicle node to the blockchain sidechain, generates a comprehensive resource availability score through a preset resource evaluation function to monitor the operational status of the lead node; a smart contract-driven node election mechanism triggers dynamic convoy reconfiguration: selecting the optimal node as the new lead node, calculating the target positions of subordinate nodes according to preset formation parameters, and updating the geometric formation of subordinate nodes using spherical coordinate transformation and distance locking mechanisms; identifying detection failure points through task status determination during the inspection process, encapsulating the failure points and publishing them to the on-chain task pool, and comprehensively evaluating the spatial distance between nodes and task points and the node resource redundancy capability to select the optimal node to execute the supplementary testing task. The method described in this invention achieves optimal scheduling of inspection resources.

Description

Technical Field

[0001] This invention relates to the field of blockchain smart contract technology, specifically to a smart contract-driven method and system for dynamic reconstruction of inspection formations. Background Technology

[0002] With the development of industrial automation and intelligent inspection technologies, collaborative inspection modes based on mobile robot formations are gradually being applied to complex scenarios such as high-speed rail machine rooms, substations, and large industrial parks. In existing technologies, multi-node inspection systems typically employ a leader-follower control architecture. A leader node plans the path and guides the movement, while the other subordinate nodes follow in a pre-defined formation to automate the inspection task. Simultaneously, with the development of sensor technology and visual recognition algorithms, inspection systems are gradually acquiring environmental perception and status recognition capabilities, improving inspection efficiency and intelligence to some extent. However, under complex dynamic environments and long-term operation conditions, traditional formation inspection systems still have significant limitations in terms of system robustness, resource scheduling capabilities, and task closed-loop mechanisms.

[0003] Specifically, existing patrol formation technologies generally rely heavily on a single lead node as the path reference and control core. If this node runs out of power, experiences communication failures, or mechanical malfunctions, the entire formation loses its motion constraints, creating a single point of failure and making it difficult to guarantee the continuity of patrol tasks. Existing systems mostly employ static task allocation mechanisms, lacking unified modeling and dynamic evaluation capabilities for multi-dimensional resource information such as node remaining power, task load, computing resources, and communication status. This results in some nodes operating at high loads for extended periods while other nodes remain idle, hindering global resource optimization and preventing the achievement of optimal scheduling based on real-time resource status. During patrols, when environmental occlusion or insufficient recognition confidence leads to detection failures, existing methods typically rely on manual verification or offline retesting, lacking an automatic retesting mechanism based on internal formation collaboration. This prevents closed-loop processing of patrol tasks and hinders adaptive redistribution and efficient retesting of failed detection points. Furthermore, existing formation control methods are mostly based on local coordinates or simple distance constraints, lacking a formation reconstruction mechanism that combines spherical coordinate transformation and stable distance locking, making it difficult to maintain formation consistency and spatial accuracy in complex geographical environments. Therefore, existing technologies are unable to achieve collaborative control effects based on dynamic decision-making, adaptive switching of navigation nodes, and closed-loop optimization of inspection tasks, all of which are based on multi-dimensional resource perception. Summary of the Invention

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

[0005] Therefore, the technical problem solved by this invention is: the existing patrol formation collaborative control method has a high risk of single-point failure, insufficient resource scheduling capability, and lack of closed-loop mechanism for task execution, and how to achieve dynamic reconstruction and adaptive allocation of supplementary testing tasks based on multi-dimensional resource perception.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: a smart contract-driven dynamic reconfiguration method for inspection convoys, comprising: a smart contract generating a comprehensive resource availability score based on the physical state vectors uploaded to the blockchain sidechain by each inspection vehicle node, and monitoring the operational status of the inspection vehicle's lead node; triggering dynamic convoy reconfiguration through a smart contract execution node election mechanism: selecting the optimal node as the new lead node, calculating the target positions of subordinate nodes according to preset formation parameters, and updating the geometric formation of subordinate nodes by combining spherical coordinate transformation and distance locking mechanisms; identifying detection failure points through task status determination during the inspection process, encapsulating the failure points and publishing them to the on-chain task pool, and comprehensively evaluating the spatial distance between nodes and task points and the node's resource redundancy capabilities to select the optimal node to execute the supplementary testing task.

[0007] As a preferred embodiment of the smart contract-driven dynamic reconfiguration method for inspection formations described in this invention, the generation of a comprehensive resource availability score includes: the smart contract constructing standardized resource description data based on the physical state vectors uploaded by each inspection vehicle node to the blockchain sidechain; inputting the multi-dimensional resource attributes in the physical state vectors into a preset resource evaluation function; representing the multi-dimensional resource attributes by weighted summation according to the weight coefficients corresponding to each multi-dimensional resource attribute in the current task stage, and outputting a comprehensive resource availability score for the node's comprehensive execution capability.

[0008] As a preferred embodiment of the smart contract-driven dynamic reconfiguration method for inspection formations described in this invention, the method of monitoring the operational status of the navigation node using a comprehensive resource availability score includes, when the inspection formation is in operation, comparing the comprehensive resource availability score of the navigation node with a preset safety threshold in real time, based on the comprehensive resource availability scores of each node; when... If the leader node is determined to be unable to continue executing tasks stably, a soft refactoring process is triggered. This represents the overall resource availability score of the current navigation node. This indicates a preset safety threshold; when LiDAR detects that the formation inspection path is blocked by an obstacle, the duration of the obstacle blockage must meet the following conditions. If the current path is deemed impassable, a hard refactoring process is triggered. Indicates the duration of the obstruction. This indicates a preset time threshold; when path congestion exists, the path congestion factor is set to an active state. The smart contract broadcasts the fleet reconstruction request on-chain.

[0009] As a preferred embodiment of the smart contract-driven dynamic reconfiguration method for inspection formations described in this invention, the step of selecting the optimal node as the new leader node includes: after receiving a reconfiguration request on the chain, the formation node enters a temporary hovering mode and synchronizes the latest resource snapshot to the sidechain; the smart contract selects a set of feasible nodes from the currently active subordinate nodes based on the latest resource snapshot synchronized on the chain. , Indicates the index of feasible nodes. In the set of feasible candidate nodes A maximum value search is performed, and the node with the highest overall resource availability score is selected as the new leader node, represented as:

[0010]

[0011] in, Represents a node Based on the comprehensive resource availability score, the candidate node with the highest score is selected. As a new leading node.

[0012] As a preferred embodiment of the smart contract-driven dynamic reconfiguration method for inspection formations described in this invention, the step of updating the geometric formation of subordinate nodes by combining spherical coordinate transformation and distance locking mechanism includes, after a new leader node is determined, the subordinate nodes update the geometric formation based on the geographical location of the new leader node. The new target pose is calculated using the Haversine spherical coordinate transformation. , is represented as:

[0013]

[0014]

[0015] in, This represents the azimuth offset of the subordinate node relative to the new leader node. This indicates the target distance between the subordinate node and the new leader node. This represents the Earth's average radius. Indicates the latitude of the subordinate node. Indicates the longitude of the subordinate node. Indicates the latitude of the new navigation node. Represent the longitude of the new navigation node; abstract each preset formation as a set. , is represented as:

[0016]

[0017] in, This represents the set of preset spatial position parameters of all subordinate nodes in the preset formation relative to the leader node. Indicates the first The array parameters of each subordinate node, Indicates the first The target distance of each subordinate node relative to the leader node Indicates the first The azimuth offset of each subordinate node relative to the current heading of the leader node. This indicates the number of subordinate nodes; a distance locking mechanism based on Euclidean distance is introduced, where each node calculates the distance to the new leader node and adjacent nodes in real time during movement; the node's attitude and displacement are continuously adjusted through a feedback control law, and when an environmental obstacle is detected that interferes with the node's movement, causing the distance deviation to exceed the preset locking distance... The attitude is fine-tuned based on the offset vector matrix synchronized by the blockchain. After each node completes the target position recalculation and distance locking adjustment, the reconstructed array coordinate data is written into the blockchain sidechain ledger. The data is confirmed through the sidechain consensus mechanism and broadcast and synchronized among the nodes.

[0018] As a preferred embodiment of the smart contract-driven dynamic reconfiguration method for inspection formations described in this invention, the following steps are included: Encapsulating failure points and publishing them to the on-chain task pool; real-time monitoring of task status using both physical layer blocking and algorithmic layer low-confidence criteria; determining whether the current data collection action is physically blocked when the inspection vehicle reaches a preset inspection point; if the LiDAR detects a temporary obstacle within the target device's range, preventing the vehicle from entering the preset optimal shooting position, the inspection point is determined to be a detection failure point; if the lifting mechanism or gimbal attempts to align with the target point and detects abnormal motor torque, physical obstruction, or the target point exceeding the effective envelope of the current joint space, the inspection point is determined to be a detection failure point; if the physical layer determines no blocking occurs, a deep learning model is used to identify status lights or meters in real time and output the confidence score of the target category. When the recognition result meets If the algorithm fails to extract valid edge features within 3 seconds, the recognition is deemed a failure, and an environmental snapshot is automatically extracted and marked as a detection failure point. This indicates a pre-set confidence threshold; when an inspection point is determined to be a detection failure point, the detection failure point is packaged into a standardized task transaction and published to the on-chain task pool; through on-chain, all active nodes and smart contracts share the failure point status.

[0019] As a preferred embodiment of the smart contract-driven dynamic reconfiguration method for inspection formations described in this invention, the step of selecting the optimal node to execute the supplementary testing task includes, after the on-chain task pool is formed, calculating the geographical proximity between the tasks in the smart contract computation pool and each node in the formation, using the Haversine spherical distance model to calculate the geographical distance between the nodes and the task points, and detecting failure points. The coordinates are ,node The coordinates are The difference between latitude and longitude is expressed as:

[0020]

[0021] in, Represents a node With detection failure points The difference in latitude, Represents a node With detection failure points The difference in longitude, Represents a node latitude, Indicates the detection failure point latitude, Represents a node longitude, Indicates the detection failure point The longitude; after obtaining the difference between latitude and longitude, construct intermediate variables. , is represented as:

[0022]

[0023] Based on intermediate variables Calculate the spherical distance between the failure point and the node. , is represented as:

[0024]

[0025] The actual geographical distance from each candidate node to the failure point is calculated using spherical distance calculation. The proximity score is then obtained by normalizing the actual geographical distance. , is represented as:

[0026]

[0027] in, This represents the maximum distance between the current set of candidate nodes and the failure point; the closer a node is to the failure point, the greater the distance. The smaller the value, the higher the proximity score; the node with the highest proximity score is selected as the primary candidate. A resource redundancy vector scoring mechanism is introduced. The capacity of a node to handle unexpected supplementary testing tasks while fulfilling its primary tasks is denoted as:

[0028]

[0029] in, This represents the weighting coefficient corresponding to the difference between the node's current battery level and the recharge warning threshold. This indicates the weighting coefficient corresponding to the proportion of nodes that are not currently scheduled for inspection. This represents the weighting coefficient corresponding to the instantaneous remaining CPU / GPU ratio of the in-vehicle edge computing unit. This represents the difference between the node's current battery level and the recharge warning threshold. This indicates the proportion of nodes currently scheduled for inspection. This represents the instantaneous CPU / GPU remaining ratio of the in-vehicle edge computing unit; the smart contract performs a maximum value search in the set of feasible nodes, selecting the node with the highest comprehensive redundancy score as the node to execute the supplementary test task, and the determination form is expressed as:

[0030]

[0031] After the smart contract determines the supplementary testing node, the failure point task is officially assigned to the target node; after completing the predetermined main trunk inspection path, the supplementary testing node automatically switches to supplementary testing mode and goes to the failure point location to perform a second inspection.

[0032] As a preferred embodiment of the smart contract-driven dynamic reconfiguration system for inspection convoys described in this invention, the system includes: a resource assessment module, a convoy reconfiguration module, and a supplementary testing scheduling module. The resource assessment module is used by the smart contract to generate a comprehensive resource availability score based on the physical state vectors uploaded to the blockchain sidechain by each inspection vehicle node, and to monitor the operating status of the inspection vehicle's lead node. The convoy reconfiguration module is used to trigger dynamic convoy reconfiguration through a node election mechanism executed by the smart contract: selecting the optimal node as the new lead node, calculating the target positions of subordinate nodes according to preset formation parameters, and updating the geometric formation of subordinate nodes using spherical coordinate transformation and distance locking mechanisms. The supplementary testing scheduling module is used to identify detection failure points during the inspection process through task status determination, encapsulate the failure points, publish them to the on-chain task pool, and comprehensively evaluate the spatial distance between nodes and task points and the node's resource redundancy capabilities to select the optimal node to execute the supplementary testing task.

[0033] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program as steps to implement a smart contract-driven method for dynamic reconfiguration of inspection formations.

[0034] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of a smart contract-driven method for dynamic reconfiguration of an inspection formation.

[0035] The beneficial effects of this invention are as follows: The smart contract-driven dynamic reconstruction method for inspection formations provided by this invention introduces a dynamic weighted resource evaluation method oriented towards task stages, achieving adaptive fusion and unified quantification of multi-dimensional heterogeneous resources of nodes. This allows node capabilities to change in real time according to task requirements, improving the accuracy and flexibility of resource scheduling. By combining resource scoring and environmental awareness to construct a dual-trigger reconstruction mechanism, predictive switching of the lead node and continuous reconstruction of the formation are achieved, effectively avoiding formation loss of control caused by single-point failures and enhancing the system's stable operation in dynamic environments. By constructing an on-chain task pool and a supplementary testing scheduling mechanism, detection failure points are transformed into shareable and schedulable task objects. Optimal node allocation is achieved by combining spatial proximity and resource redundancy, forming a closed-loop execution path for inspection tasks. This invention solves the problems of high single-point failure risk and rigid task allocation in traditional inspection formations, achieving distributed optimal scheduling of heterogeneous inspection resources. Attached Figure Description

[0036] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 The above is an overall flowchart of a smart contract-driven dynamic reconfiguration method for inspection formations provided in Embodiment 1 of the present invention.

[0038] Figure 2 This is a supplementary test assignment diagram for a smart contract-driven dynamic reconfiguration method for inspection formations provided in Embodiment 1 of the present invention.

[0039] Figure 3 This is a schematic diagram of a computer device for a smart contract-driven dynamic reconfiguration method for inspection formations, as provided in Embodiment 4 of the present invention. Detailed Implementation

[0040] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0041] Example 1, referring to Figures 1-2 As an embodiment of the present invention, a smart contract-driven method for dynamic reconfiguration of inspection formations is provided, comprising:

[0042] S1: The smart contract generates a comprehensive resource availability score based on the physical state vectors uploaded to the blockchain sidechain by each inspection vehicle node, and monitors the operating status of the inspection vehicle leader node through a preset resource evaluation function.

[0043] Furthermore, generating a comprehensive resource availability score involves the smart contract constructing standardized resource description data based on the physical state vectors uploaded to the blockchain sidechain by each inspection vehicle node; inputting the multi-dimensional resource attributes in the physical state vectors into a preset resource evaluation function; representing the multi-dimensional resource attributes by weighted summation according to the weight coefficients corresponding to each multi-dimensional resource attribute in the current task stage; and outputting a comprehensive resource availability score that reflects the node's overall execution capability.

[0044] It should also be noted that a preferred scheme for generating a comprehensive resource availability score includes: constructing a dynamic weight allocation strategy to dynamically allocate the importance of each resource dimension; adaptively adjusting the importance of various resource attributes by analyzing the current task's different stages, such as path exploration, data collection, or data feedback, so that under different operating conditions, energy factors, computing power factors, and communication performance factors can dynamically adjust their influence according to task requirements; each node of the inspection formation periodically broadcasts its own physical state vector to the sidechain; and the smart contract uses a resource evaluation function to calculate the node's status. The overall resource availability score is expressed as:

[0045]

[0046] in, Represents a node The overall resource availability score, Indicates the node index. Represents a node The normalized remaining power is used to characterize the node's endurance. Represents a node The normalized task load is used to reflect the current task occupancy of a node. Represents a node The normalized computing power reserve is used to reflect the node's ability to execute computing tasks; Represents a node The normalized bandwidth is used to characterize the data transmission carrying capacity of a node in the current network environment; Represents a node The normalized communication delay factor is used to characterize the communication response quality of nodes under the current network topology and electromagnetic environment. Represents a node Weighting coefficients corresponding to normalized remaining power Represents a node Weighting coefficients corresponding to normalized task load Represents a node Weighting coefficients corresponding to normalized computing power reserves Represents a node Weighting coefficients corresponding to normalized bandwidth Represents a node The weighting coefficients corresponding to the normalized communication delay factor are dynamically adjusted according to the current task stage; during the path search stage, the computing power weight is increased to improve environmental awareness and path planning capabilities. During the centralized data transmission phase of the inspection, bandwidth weighting is increased to ensure data transmission efficiency. ; Used for real-time quantification of inspection vehicle nodes Communication quality under the current electromagnetic environment and network topology.

[0047] It should be noted that monitoring the operational status of the navigation node using comprehensive resource availability scoring includes dynamically monitoring the continuous performance capability of the current navigation node while the inspection formation is in operation, based on the comprehensive resource availability scores of each node. When the navigation node is no longer suitable to assume navigation responsibilities, or the formation's operational path is physically impassable, a reconfiguration process is automatically initiated to achieve navigation rights switching, formation recalculation, and position locking, thereby ensuring the continuous execution of the inspection mission.

[0048] Specifically, the comprehensive resource availability score of the lead node is read and compared with a preset safety threshold. When the comprehensive resource availability score of the lead node is lower than the preset safety threshold, it indicates that the node lacks one or more basic capabilities in multiple dimensions such as remaining power, task load, computing power, and communication capabilities, and is not suitable to continue to lead the system. In this case, a soft reconfiguration is triggered. The judgment can be expressed as follows:

[0049]

[0050] in, This represents the overall resource availability score of the current navigation node. This represents a preset safety threshold used to determine whether the lead node has the ability to continuously perform its duties. It is set based on operational experience or historical data statistics, with a preferred range of 0.3 to 0.5. Based on the normalized distribution characteristics of the node's comprehensive resource availability score, a score higher than 0.5 indicates that the node has sufficient reserves in multiple dimensions such as energy, computing power, and communication, and can stably undertake the lead task. A score lower than 0.3 indicates that the node is approaching a bottleneck in at least one critical resource dimension, and continuing to perform lead duties will increase the risk of failure. When the score is in the range of 0.3 to 0.5, the node is in a resource critical state, which can prevent premature reconfiguration and avoid formation loss of control due to resource depletion. The preset safety threshold can be adaptively adjusted based on historical inspection data statistics or operating environment characteristics.

[0051] When LiDAR detects that the formation inspection path is blocked by an obstacle, the duration of the obstacle blockage meets the following condition. When this is triggered, a hard refactoring is performed. This indicates the duration of obstacle blockage, obtained by accumulating the time from multiple consecutive frames of detection results. This represents the time threshold for determining congestion, used to distinguish between transient occlusion and persistent path blocking. Preferably, The value range is 2s to 3s. Due to the presence of short-term dynamic interferences such as personnel passing by and equipment swinging in the inspection environment, the duration is usually less than 2s. If the blocking judgment time threshold is set too small, it is easy to misjudge the instantaneous occlusion as path blocking, which will lead to unnecessary formation reconfiguration. When the obstacle continues to block for more than 3s, it usually indicates that the path is in a stable blocking state and path reconfiguration needs to be triggered to ensure the continuity of inspection. When the time of the obstacle's continuous existence exceeds the threshold, the path blocking is determined to be established.

[0052] To characterize the path congestion state, a path congestion state flag variable is introduced. When path congestion exists, the path congestion factor is set to an active state. The array reconstruction request is broadcast on-chain by the smart contract; the path blocking factor is obtained by LiDAR jointly judging the distance to the obstacle and the duration of the blockage.

[0053] It should also be noted that in a multi-node inspection system, a unified and dynamically adjustable evaluation mechanism needs to be established among multi-dimensional heterogeneous resources (energy, computing power, communication, etc.) to support subsequent scheduling and decision-making. Existing technologies typically use a single indicator (such as power consumption or distance) or a fixed-weight model for evaluation, which is difficult to reflect the true capability status of nodes at different task stages, especially in scenarios with frequent switching of working conditions such as path exploration, data collection, and data backhaul, where the importance of resources exhibits obvious stage-specific changes. Therefore, this invention elevates "resource evaluation" from a static weighted model to a "task-stage-driven dynamic weight model." By introducing a task stage identification mechanism, the weights of key resources such as computing power and bandwidth are adaptively adjusted, and a unified comprehensive resource availability scoring system is constructed by combining normalization processing. Smart contracts are used to achieve on-chain synchronization and consistent computation of node states, ensuring global consistency and credibility of the evaluation results. A dynamic mapping relationship from multi-dimensional resource vectors to single decision indicators is constructed, enabling different resource dimensions to play a dominant role in different task contexts. This achieves a fine-grained characterization and comparable expression of node capabilities.

[0054] S2: Trigger dynamic reconfiguration of the formation through the smart contract execution node election mechanism: Select the best node as the new leader node, calculate the target position of the subordinate nodes according to the preset formation parameters, and update the geometric formation of the subordinate nodes by combining spherical coordinate transformation and distance locking mechanism.

[0055] Furthermore, selecting the optimal node as the new leader node involves performing a full node state synchronization after the reconstruction request is issued. Specifically, after receiving the reconstruction request on the chain, all swarm nodes enter a temporary hovering mode and synchronize the latest resource snapshot to the sidechain; the resource snapshot includes the node's remaining power, current task occupancy status, remaining computing power, and current geographical coordinates.

[0056] After completing the full node state synchronization, the process enters the leader node election phase. Specifically, the smart contract selects a set of feasible nodes from the currently active subordinate nodes based on the latest resource snapshot synchronized on the chain. , Indicates the index of feasible nodes. The selection criteria are: the node has more than 20% remaining power and more than 15% remaining computing power.

[0057] After obtaining a set of feasible candidate nodes, the optimal candidate is further selected as the new leader node from the candidate set. A maximum value search is performed, and the node with the highest overall resource availability score is selected as the new leader node, represented as:

[0058]

[0059] in, Represents a node Based on the comprehensive resource availability score, the candidate node with the highest score is selected. As a new leading node.

[0060] It should be noted that updating the geometric formation of subordinate nodes by combining spherical coordinate transformation and distance locking mechanisms includes, after the new leader node is determined, the remaining subordinate nodes are based on the geographical location of the new leader node. The new target pose is calculated using the Haversine spherical coordinate transformation. , is represented as:

[0061]

[0062]

[0063] in, This represents the azimuth offset of the subordinate node relative to the new leader node. This indicates the target distance between the subordinate node and the new leader node. This represents the Earth's average radius. Indicates the latitude of the subordinate node. Indicates the longitude of the subordinate node. Indicates the latitude of the new navigation node. This indicates the longitude of the new lead node; by recalculating the target latitude and longitude, it ensures that even during leadership transitions in large-scale industrial parks or long-distance route inspections, each node can still obtain a high-precision global positioning reference; furthermore, to uniformly manage different formations, this invention abstracts each preset formation (V-shape, rhombus, straight line) into a set of "angle-distance" pairs. , is represented as:

[0064]

[0065] in, This represents the set of preset spatial position parameters of all subordinate nodes relative to the leader node in a given preset formation. Let be any pair of parameters in the set, representing the th The array parameters of each subordinate node, Indicates the first The target distance of each subordinate node relative to the leader node Indicates the first The azimuth offset of each subordinate node relative to the current heading of the leader node. This indicates the number of subordinate nodes. Different formations can be defined using different sets of parameters. After recalculating the target latitude and longitude, it is also necessary to ensure that each node can stably maintain the preset formation during movement and adjustment; furthermore, a distance locking mechanism based on Euclidean distance is introduced to constrain the relative spatial relationships between nodes in real time.

[0066] Specifically, each node calculates its distance to the new leader node and neighboring nodes in real time during the movement. , is represented as:

[0067]

[0068] in, Represents a node Global geographic coordinates Represents a node The system uses global geographic coordinates and a feedback control law to continuously adjust node attitude and displacement. When an environmental obstacle is detected interfering with the node's movement, causing a distance deviation exceeding a preset locking distance, the system will lock the node. The attitude is fine-tuned based on the offset vector matrix synchronized with the blockchain. Environmental obstacles include external factors that affect node movement, such as fixed obstacles, dynamic obstacles, and path space limitations. Preferably, the preset locking distance is between 0.5m and 2m. During actual operation, the inspection vehicle is affected by global positioning errors, sensor measurement errors, and motion control errors, inevitably causing deviations in the distance between nodes. When the deviation is less than 0.5m, the impact on formation stability and inspection effect is small, and frequent adjustments are not required. When the deviation exceeds 2m, the formation structure deviates from the preset geometric relationship, causing interference or collision risks between nodes. Setting the preset locking distance within the range of 0.5m to 2m can ensure formation stability. If the distance deviation exceeds the preset locking distance due to environmental obstacles, the local controller will fine-tune the attitude based on the offset vector matrix synchronized with the blockchain, thereby maintaining the geometric consistency of the formation during reconstruction and preventing node collisions.

[0069] After each node completes the target position recalculation and distance locking adjustment, the reconstructed formation coordinate data is written into the blockchain sidechain ledger. The data is confirmed through the sidechain consensus mechanism and broadcast synchronously among the nodes, serving as a shared reference benchmark for the inspection phase. The sidechain ledger is a blockchain branch ledger deployed in the inspection formation. Each inspection vehicle node participates in data recording and maintenance as a sidechain node, used to store node status information, formation coordinate data, and task-related information, and confirms the data through the sidechain consensus mechanism.

[0070] It should also be noted that in the inspection formation, the stability of the lead node directly determines the overall operational reliability. The traditional Leader-Follower mode faces the passive control problem of "responding only after failure." When the lead node completely fails, the system only switches, leading to short-term loss of control of the formation or even mission interruption. Existing technologies lack a unified handling mechanism for two different anomalies: "path impassable" and "node capability degradation," making it difficult to achieve stable transitions in complex environments. Therefore, this invention constructs a threshold judgment mechanism based on comprehensive resource scoring, which can trigger soft reconstruction in advance when resources approach a bottleneck. At the same time, it introduces path blocking judgment logic based on LiDAR and time accumulation to achieve hard-trigger identification of physically impassable states. Through on-chain state synchronization, candidate node screening, and maximum value search mechanisms, the optimal selection of a new lead node is achieved, and the formation is continuously rebuilt by combining spherical coordinate transformation and distance locking mechanisms. The process of "node capability assessment—environmental perception—reconstruction triggering—formation recovery" is linked into a closed-loop control process. It enables smooth reconfiguration and seamless transition of formations in dynamic environments, eliminates the risk of single point of failure, improves the continuous operation capability of formations, and enhances system robustness.

[0071] S3: During the inspection process, the failure points are identified by judging the task status. After the failure points are encapsulated, they are published to the on-chain task pool. Based on the spatial distance between the node and the task point and the node's resource redundancy capability, a comprehensive evaluation is performed, and the optimal node is selected to execute the supplementary test task.

[0072] Furthermore, the failure points are encapsulated and published to the on-chain task pool, including real-time monitoring of task status using both physical layer blocking and low-confidence algorithmic criteria. When the inspection vehicle reaches the preset inspection point, it is determined whether the current acquisition action is physically blocked. If the LiDAR detects temporary obstacles within 1 meter of the target device, such as maintenance tools, on-site personnel, or temporary piles of objects, preventing the vehicle from entering the preset optimal shooting position, the inspection point is determined to be a detection failure point. When the lifting mechanism or gimbal attempts to align with the target point, if abnormal motor torque is detected, indicating physical obstruction, or if the target point exceeds the effective envelope of the current joint space, the inspection point is determined to be a detection failure point.

[0073] If the physical layer determines that no obstruction has occurred, a deep learning model is used to identify the status lights or meters in real time and output the confidence score of the target category. When the recognition result meets If the algorithm fails to extract valid edge features within 3 seconds, the recognition is deemed a failure, and an environmental snapshot is automatically extracted and marked as a detection failure point. This indicates a preset confidence threshold, preferably set between 0.8 and 0.9. Based on the output characteristics of target detection and recognition algorithms in actual industrial inspection scenarios, when the confidence level is higher than 0.9, the recognition results are generally highly reliable, but some valid targets may be missed. When the confidence level is lower than 0.8, the recall rate increases, but the false detection probability increases, making it easy to misclassify abnormal recognition results as normal. Therefore, setting the confidence threshold in the range of 0.8 to 0.9 ensures recognition accuracy while also considering detection coverage.

[0074] Once an inspection point is identified as a detection failure, it is packaged into a standardized task transaction and published to the on-chain task pool. The packaged information includes the task identifier corresponding to the failure point, the global coordinates of the failure point, the current environment snapshot, the failure reason tag, and the generation timestamp. Through on-chain processing, all active nodes and smart contracts can share the failure point's status.

[0075] It should be noted that, referring to Figure 2 The process of selecting the optimal node to perform the supplementary testing task includes: after the on-chain task pool is formed, automatically assigning the task to the nearest node with the highest redundancy based on the geographical proximity between the task in the smart contract computing pool and each node in the formation; determining the spatial relationship between each node in the formation and the failure point; since the inspection scenarios addressed by this invention may cover large industrial parks, long-distance routes, and complex environments with terrain changes, to avoid errors caused by simple planar distance calculations, this invention uses the Haversine spherical distance model to calculate the geographical distance between nodes and task points to detect failure points. The coordinates are ,node The coordinates are The difference between latitude and longitude is expressed as:

[0076]

[0077] in, Represents a node With detection failure points The difference in latitude, Represents a node With detection failure points The difference in longitude, Represents a node latitude, Indicates the detection failure point latitude, Represents a node longitude, Indicates the detection failure point The longitude; after obtaining the difference between latitude and longitude, further construct intermediate variables. , is represented as:

[0078]

[0079] Based on intermediate variables Calculate the spherical distance between the failure point and the node. , is represented as:

[0080]

[0081] The actual geographical distance from each candidate node to the failure point is calculated using spherical distance calculation. The proximity score is then obtained by normalizing the actual geographical distance. , is represented as:

[0082]

[0083] in, This represents the maximum distance from the failure point within the current set of candidate nodes. The closer a node is to the failure point... The smaller the value, the higher the proximity. The node with the highest score is selected as the first candidate to ensure the shortest possible supplementary test path and reduce the additional energy consumption of the formation.

[0084] Selecting supplementary testing nodes solely based on geographical proximity is insufficient. From a location perspective, this could lead to frequent selection of a particular nearby node, resulting in overload, premature retirement, or disruption of main path inspections. Therefore, this invention further introduces a resource redundancy vector scoring mechanism. The capacity of a node to handle unexpected supplementary testing tasks while fulfilling its primary tasks is denoted as:

[0085]

[0086] in, This represents the weighting coefficient corresponding to the difference between the node's current battery level and the recharge warning threshold. This indicates the weighting coefficient corresponding to the proportion of nodes that are not currently scheduled for inspection. This represents the weighting coefficient corresponding to the instantaneous CPU / GPU remaining ratio of the in-vehicle edge computing unit. The weighting coefficient satisfies the normalization constraint. This represents the difference between the node's current battery level and the recharge warning threshold. The larger the difference, the more energy the node has to perform supplementary testing tasks that require it to leave the main path. This indicates the proportion of nodes currently scheduled for inspection. The lower the value, the lighter the current task load of the node, and the more immediately it can respond to the supplementary test command. This represents the instantaneous CPU / GPU remaining ratio of the vehicle-mounted edge computing unit. For supplementary testing points requiring complex image processing, nodes with high computing power redundancy are prioritized. Furthermore, the smart contract performs a maximum value search within the feasible node set, selecting the node with the highest overall redundancy score as the execution node for the supplementary testing task. The determination can be expressed as follows:

[0087]

[0088] Dynamic load balancing is achieved by selecting the node with the highest overall redundancy score, thus avoiding the long-term concentration of supplementary testing tasks on a single node. After the smart contract determines the supplementary testing node, the failure point task is formally assigned to the target node. After completing the predetermined backbone inspection path, the supplementary testing node automatically switches to supplementary testing mode and proceeds to the failure point location to perform a secondary inspection, achieving 100% coverage and closed-loop inspection tasks.

[0089] It should also be noted that this invention achieves reliable identification of failure points through a dual criterion of physical blockage and low confidence; it encapsulates failure points into standardized on-chain tasks, possessing shareable and schedulable attributes; in the task allocation phase, it introduces a geographical proximity assessment based on spherical distance to reduce path costs, and constructs a resource redundancy vector scoring model to comprehensively evaluate the node's ability to undertake supplementary testing tasks from three dimensions: energy, task load, and computing power, thus determining the optimal execution node; through a dual-dimensional decision-making mechanism of "spatial factors + resource capabilities," it breaks through the traditional method of task allocation based on distance or a single resource indicator. This achieves a fully automated closed loop from anomaly detection to secondary supplementary testing, eliminating blind spots in detection.

[0090] Example 2 is an embodiment of the present invention, which provides a smart contract-driven method for dynamic reconfiguration of inspection formations. To verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculations and simulation experiments.

[0091] This invention underwent 50 comparative experiments in a high-speed rail data center with dense server racks, including 40 inspection points, 3 heterogeneous inspection vehicles, and simulated signal interference. The key performance indicators of the proposed solution (Group C) were quantitatively compared with those of traditional centralized static programming (Group A) and a heuristic reconstruction scheme based on the Artificial Potential Field (APF) method (Group B). The comparison results are shown in Table 1.

[0092] Table 1. Performance Comparison Results of Dynamic Reconfiguration and Load Balancing in Formation

[0093]

[0094] In a scenario simulating a leader node suddenly going offline due to battery depletion, Solution A, relying on centralized heartbeat monitoring, immediately becomes blocked after the leader fails, unable to issue reorganization commands. Solution B, while possessing heuristic obstacle avoidance, lacks global identity consensus, requiring 12.5 seconds for the reorganization process and is prone to formation disarray. In contrast, this invention utilizes the real-time state table of the blockchain sidechain, triggering an on-chain election contract when the leader's score falls below a preset threshold, completing the selection and authorization of a new leader in 3.2 seconds, representing a 74.4% improvement in response speed.

[0095] Inspection task coverage is a core indicator for measuring closed-loop capability. In 50 sets of experiments, both Solution A and Solution B could only record the fault and wait for manual intervention after the task was completed when detection failures were caused by temporary personnel movement or obstruction of the robotic arm's field of vision. The failure point task pool mechanism introduced in this invention automatically assigns the failure coordinates to the node with the highest resource redundancy for retesting through smart contracts, ensuring 100% closed-loop coverage of inspection tasks and realizing unmanned auditing of industrial operations and maintenance.

[0096] The standard deviation of node energy consumption reflects the fairness of resource utilization in platooning operations. Scheme A has an energy consumption standard deviation of 45.0%, with the lead vehicle undertaking excessive sensing calculations, leading to rapid battery depletion. This invention dynamically adjusts resource scores through weighted coefficients, allowing nodes to take over leadership or undertake supplementary testing tasks as needed. The node energy consumption standard deviation is reduced to 12.3%, effectively alleviating the problem of single-vehicle battery overload and extending the overall platooning endurance by approximately 18.5%.

[0097] During the dynamic reconstruction process, this invention uses Haversine spherical coordinate transformation for global positioning and alignment, and Euclidean distance locking constraints for local pose fine-tuning, controlling the formation consistency error within 0.08 meters. This ensures that heterogeneous devices (such as ground UGVs and low-altitude UAVs) can maintain a preset perception overlap based on the same geographical base at the moment of leadership switching, increasing the obstacle avoidance success rate to 99.2% and enhancing survivability robustness in high-density dynamic environments.

[0098] In summary, this invention successfully solves three major technical challenges in complex industrial environments—perception blind spots, single-point failures, and uneven task allocation—by deeply coupling the decentralized characteristics of blockchain with the physical resources of inspection vehicles. Experimental data shows that this invention, while ensuring inspection coverage, achieves spontaneous optimal allocation of fleet resources through smart contracts.

[0099] Example 3, an embodiment of the present invention, provides a smart contract-driven dynamic reconfiguration system for inspection convoys, including a resource assessment module, a convoy reconfiguration module, and a supplementary testing scheduling module. The resource assessment module uses a smart contract to generate a comprehensive resource availability score based on the physical state vectors uploaded by each inspection vehicle node to the blockchain sidechain, and monitors the operational status of the lead node. The convoy reconfiguration module triggers dynamic convoy reconfiguration through a node election mechanism executed by the smart contract: selecting the optimal node as the new lead node, calculating the target positions of subordinate nodes according to preset formation parameters, and updating the geometric formation of subordinate nodes using spherical coordinate transformation and distance locking mechanisms. The supplementary testing scheduling module identifies detection failure points during the inspection process through task status determination, encapsulates the failure points, publishes them to the on-chain task pool, and performs a comprehensive evaluation based on the spatial distance between nodes and task points and the node's resource redundancy capability, selecting the optimal node to execute the supplementary testing task.

[0100] Example 4, refer to Figure 3 This embodiment also provides a computer device applicable to the smart contract-driven dynamic reconfiguration method for inspection formations, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the smart contract-driven dynamic reconfiguration method for inspection formations as proposed in the above embodiment.

[0101] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0102] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the smart contract-driven patrol formation dynamic reconfiguration method proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

Claims

1. A method for dynamic reconfiguration of a patrol formation driven by a smart contract, characterized in that, include: The smart contract generates a comprehensive resource availability score based on the physical state vectors uploaded to the blockchain sidechain by each inspection vehicle node, and monitors the operating status of the inspection vehicle navigation node through a preset resource evaluation function. The smart contract execution node election mechanism triggers dynamic reconfiguration of the formation: the optimal node is selected as the new leader node, the target position of the subordinate nodes is calculated according to the preset formation parameters, and the geometric formation of the subordinate nodes is updated by combining spherical coordinate transformation and distance locking mechanism. During the inspection process, failure points are identified by judging the task status. After the failure points are encapsulated, they are published to the on-chain task pool. A comprehensive evaluation is conducted based on the spatial distance between the node and the task point and the node's resource redundancy capability to select the optimal node to execute the supplementary test task. 2.The smart contract driven patrol formation dynamic reconfiguration method of claim 1, wherein: The generated comprehensive resource availability score includes, The smart contract constructs standardized resource description data based on the physical state vectors uploaded by each inspection vehicle node to the blockchain sidechain; Input the multi-dimensional resource attributes in the physical state vector into the preset resource evaluation function; The multidimensional resource attributes are represented by a weighted summation based on the weight coefficients of each multidimensional resource attribute in the current task stage, and the comprehensive resource availability score of the node's overall execution capability is output. 3.The smart contract driven patrol formation dynamic reconfiguration method of claim 2, wherein: The monitoring of the operational status of the inspection vehicle's navigation node includes... When the inspection formation is in operation, the comprehensive resource availability score of each node is combined with the preset safety threshold and the comprehensive resource availability score of the navigation node is compared in real time. when If the leader node is determined to be unable to continue executing tasks stably, a soft refactoring process is triggered. This represents the overall resource availability score of the current navigation node. Indicates the preset safety threshold; When LiDAR detects that the formation inspection path is blocked by an obstacle, the duration of the obstacle blockage meets the following condition. If the current path is deemed impassable, a hard refactoring process is triggered. Indicates the duration of the obstruction. Indicates a preset time threshold; When path congestion exists, set the path congestion factor to an active state. The smart contract broadcasts the fleet reconstruction request on-chain.

4. The smart contract-driven dynamic reconfiguration method for inspection formations as described in claim 3, characterized in that: The selection of the optimal node as the new leader node includes... After receiving a reconstruction request on the chain, the formation node enters a temporary hovering mode and synchronizes the latest resource snapshot to the sidechain; The smart contract selects a set of feasible nodes from the currently active slave nodes based on the latest resource snapshot synchronized on the blockchain. , Indicates the index of feasible nodes. ; In the set of feasible candidate nodes A maximum value search is performed, and the node with the highest overall resource availability score is selected as the new leader node, represented as: in, Represents a node Based on the comprehensive resource availability score, the candidate node with the highest score is selected. As a new leading node.

5. The smart contract-driven dynamic reconfiguration method for inspection formations as described in claim 4, characterized in that: The method of updating the geometric shape of subordinate nodes by combining spherical coordinate transformation and distance locking mechanism includes, Once a new leader node is determined, subordinate nodes are assigned based on the leader node's geographical location. The new target pose is calculated using the Haversine spherical coordinate transformation. , is represented as: in, This represents the azimuth offset of the subordinate node relative to the new leader node. This indicates the target distance between the subordinate node and the new leader node. This represents the Earth's average radius. Indicates the latitude of the subordinate node. Indicates the longitude of the subordinate node. Indicates the latitude of the new navigation node. Indicates the longitude of the new navigation node; Abstract each preset formation into a set. , is represented as: in, This represents the set of preset spatial position parameters of all subordinate nodes in the preset formation relative to the leader node. Indicates the first The array parameters of each subordinate node, Indicates the first The target distance of each subordinate node relative to the leader node Indicates the first The azimuth offset of each subordinate node relative to the current heading of the leader node. Indicates the number of subordinate nodes; A distance locking mechanism based on Euclidean distance is introduced, in which each node calculates the distance between itself and the new leader node and its neighboring nodes in real time during the movement process; The node's attitude and displacement are continuously adjusted via a feedback control law. When an environmental obstacle is detected that interferes with the node's movement, causing a distance deviation exceeding the preset locking distance, the node can be locked in place. The attitude is fine-tuned based on the offset vector matrix synchronized with the blockchain; After each node completes the recalculation of the target position and the adjustment of the distance lock, the reconstructed array coordinate data is written into the blockchain sidechain ledger, the data is confirmed through the sidechain consensus mechanism, and broadcast and synchronized among the nodes.

6. The smart contract-driven dynamic reconfiguration method for inspection formations as described in claim 5, characterized in that: The step of encapsulating the failure point and publishing it to the on-chain task pool includes, Real-time monitoring of task status is achieved through a dual criterion of physical layer blocking and low-confidence algorithm layer. Once the inspection vehicle reaches the preset inspection point, it determines whether the current data collection action is physically blocked. If the lidar detects a temporary obstacle within the target device's range, preventing the vehicle from entering the preset optimal shooting position, the inspection point is determined to be a detection failure point. When the lifting mechanism or gimbal attempts to align with the target point, if abnormal motor torque is detected, physical obstruction is present, or the target point exceeds the effective envelope of the current joint space, the inspection point is determined to be a detection failure point. If the physical layer determines that no obstruction has occurred, a deep learning model is used to identify the status lights or meters in real time and output the confidence score of the target category. ; When the recognition result meets If the algorithm fails to extract valid edge features within 3 seconds, the recognition is deemed a failure, and an environmental snapshot is automatically extracted and marked as a detection failure point. This indicates a preset reliability threshold; Once an inspection point is identified as a detection failure point, the detection failure point is packaged into a standardized task transaction and published to the on-chain task pool. By putting it on the blockchain, all active nodes and smart contracts share the failure point state.

7. The smart contract-driven dynamic reconfiguration method for inspection formations as described in claim 6, characterized in that: The process of selecting the optimal node to perform the supplementary testing task includes... After the on-chain task pool is formed, the geographical proximity of tasks in the smart contract computation pool to each node in the formation is calculated using the Haversine spherical distance model to determine the geographical distance between nodes and task points, and failure points are detected. The coordinates are ,node The coordinates are The difference between latitude and longitude is expressed as: in, Represents a node With detection failure points The difference in latitude, Represents a node With detection failure points The difference in longitude, Represents a node latitude, Indicates the detection failure point latitude, Represents a node longitude, Indicates the detection failure point Longitude; After obtaining the latitude and longitude difference, construct intermediate variables. , is represented as: Based on intermediate variables Calculate the spherical distance between the failure point and the node. , is represented as: The actual geographical distance from each candidate node to the failure point is calculated using spherical distance calculation. The proximity score is then obtained by normalizing the actual geographical distance. , is represented as: in, This represents the maximum distance between the current set of candidate nodes and the failed node; The closer the node is to the failure point The smaller the value, the higher the proximity score. The node with the highest proximity score is selected as the first candidate. Introducing a resource redundancy vector scoring mechanism The capacity of a node to handle unexpected supplementary testing tasks while fulfilling its primary tasks is denoted as: in, This represents the weighting coefficient corresponding to the difference between the node's current battery level and the recharge warning threshold. This indicates the weighting coefficient corresponding to the proportion of nodes that are not currently scheduled for inspection. This represents the weighting coefficient corresponding to the instantaneous remaining CPU / GPU ratio of the in-vehicle edge computing unit. This represents the difference between the node's current battery level and the recharge warning threshold. This indicates the proportion of nodes currently scheduled for inspection. This indicates the instantaneous remaining CPU / GPU ratio of the in-vehicle edge computing unit; The smart contract performs a maximum value search within the set of feasible nodes, selecting the node with the highest overall redundancy score as the node for the supplementary testing task. The determination is expressed as follows: After the smart contract determines the retest node, the failure point task is formally assigned to the target node; After completing the predetermined main inspection path, the supplementary inspection node automatically switches to supplementary inspection mode and proceeds to the location of the failure point to perform a secondary inspection.

8. A smart contract-driven dynamic reconfiguration system for inspection formations, employing the smart contract-driven dynamic reconfiguration method for inspection formations as described in any one of claims 1 to 7, characterized in that: This includes a resource assessment module, a formation reconstruction module, and a supplementary testing and scheduling module; The resource assessment module is used by smart contracts to generate a comprehensive resource availability score based on the physical state vectors uploaded to the blockchain sidechain by each inspection vehicle node, and to monitor the operating status of the inspection vehicle navigation node. The formation reconstruction module is used to trigger dynamic formation reconstruction through the smart contract execution node election mechanism: selecting the optimal node as the new leader node, calculating the target position of the subordinate nodes according to the preset formation parameters, and updating the geometric formation of the subordinate nodes by combining spherical coordinate transformation and distance locking mechanism. The supplementary testing scheduling module is used to identify detection failure points through task status determination during the inspection process, encapsulate the failure points and publish them to the on-chain task pool, and conduct a comprehensive evaluation based on the spatial distance between the node and the task point and the node resource redundancy capability to select the optimal node to execute the supplementary testing task.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the smart contract-driven dynamic reconfiguration method for inspection formations as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the smart contract-driven dynamic reconfiguration method for inspection formations as described in any one of claims 1 to 7.

Images