Real-time video and data security transmission method between unmanned aerial vehicle nest and centralized control platform

Abstract

This invention discloses a method for secure real-time video and data transmission between a UAV nest and a centralized control platform, relating to the field of data transmission technology. The method deploys a trusted perception-driven edge intelligent agent on the UAV nest side. The steps are as follows: In a trusted execution environment, the data transmitted back by the UAV is processed through a power equipment defect analysis model to generate value level tags; the real-time status of the public network and the dedicated power network is perceived; combining the tags and network status, a hybrid link scheduler dynamically decides the transmission path; relevant information is bound in the trusted execution environment to generate a lightweight evidence header; the data packet containing the evidence header is sent to the centralized control platform according to the scheduling path, and the evidence header is simultaneously stored online; the centralized control platform queries the evidence records to verify and process the data; it also includes anti-interference degradation and operational closed-loop optimization. This invention achieves the integration of data trust and transmission optimization, ensuring end-to-end trust, improving system stability, reducing bandwidth costs, and is applicable to scenarios such as power transmission line operation and maintenance.

Description

Technical Field

[0001] This invention relates to the field of data transmission technology, specifically a method for secure real-time video and data transmission between a drone's nest and a centralized control platform. Background Technology

[0002] With the large-scale application of autonomous drone inspection technology in the operation and maintenance of power transmission lines, automated drone nests deployed on the pole sides serve as critical infrastructure, undertaking the core functions of providing drones with take-off, landing, charging, and data exchange. In this scenario, the nests need to reliably and in real-time transmit high-definition inspection videos, infrared thermal images, and equipment status data collected by the drones back to a remote centralized control platform for monitoring and analysis. Currently, related technologies are evolving mainly in two directions: one is to utilize edge computing for preprocessing at the data source to reduce the backhaul pressure, such as the patent with publication number CN120601614A, which proposes a method for real-time identification and early warning of power grid defects; the other is to build a cloud-edge-device collaborative architecture to optimize computing power allocation and task scheduling, such as the patent with publication number CN116744368B, which describes an intelligent collaborative heterogeneous air-ground unmanned system and its implementation method based on a cloud-edge-device architecture, which improves the overall system efficiency through vertical offloading and horizontal migration of tasks between the cloud, edge, and execution end. These existing technologies have, to some extent, solved the problems of real-time data processing and system collaboration.

[0003] However, in the specific high-reliability and high-security scenario of power line inspection, the aforementioned existing technical solutions still have a pressing technical flaw: the intelligent decision-making of data transmission paths and the reliable assurance of the data itself are isolated from each other, forming a "decision blind spot." Specifically, although existing edge computing solutions can perform local analysis and label data value, the "critical alarm" labels they generate may be tampered with due to edge device intrusion or algorithm interference, lacking a mechanism to prove their innocence. Furthermore, cloud-edge collaborative scheduling solutions mainly rely on external indicators such as network status for task offloading, without using the inherent, reliably verified value attributes of the data as a key input for routing decisions. This results in the system being unable to ensure that a video marked as "critical defect" is both immutable and transmitted through the current optimal network path throughout its generation, value determination, path selection, and upload.

[0004] In summary, existing technologies lack a mechanism that deeply integrates "data trustworthiness" and "transmission optimization" in a closed loop. Therefore, this field needs an innovative data transmission method to address the contradiction between ensuring optimal transmission and end-to-end trustworthiness of critical business data in complex heterogeneous network environments. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a method for secure real-time video and data transmission between UAV nests and centralized control platforms. By deploying trusted perception-driven edge intelligent agents, combined with trusted execution environments, blockchain notarization, and hybrid link scheduling, it achieves deep integration of trusted data value determination and network status perception, dynamically matches transmission paths, and solves the problem that critical data cannot simultaneously guarantee optimal transmission and end-to-end trustworthiness, thus meeting the requirements for high-security and high-reliability transmission.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: a method for secure real-time video and data transmission between a UAV nest and a centralized control platform, comprising deploying a trusted perception-driven edge agent on the UAV nest side, and the method including the following steps: Step 1: The data is processed in the trusted execution environment on the UAV nest side. The processing includes analyzing the original video stream or images transmitted by the UAV through a power equipment defect analysis model, and generating data value level labels based on the analysis results and preset rules. Step 2: Sensing the real-time network link status between the public network link and the private power grid. The network link status includes signal strength, network latency, available bandwidth, and link cost. Step 3: Based on the data value level label and the real-time network link status, the transmission path is dynamically decided and scheduled through the hybrid link scheduler. Step 4: Bind the data value level label, the feature information of the original video stream or image, the timestamp, the device identifier, and the geographical location information in the trusted execution environment and calculate the aggregate hash value to generate a lightweight evidence header; Step 5: Send the data packet containing the lightweight evidence header to the centralized control platform through the scheduled transmission path, and simultaneously submit the lightweight evidence header to the distributed ledger node for evidence storage. Step six: After receiving the data packet, the centralized control platform queries the corresponding evidence storage record from the distributed ledger node for verification, and processes it accordingly based on the verification result and the data value level label.

[0007] Furthermore, in step one, the trusted execution environment is built based on a dedicated security chip, and the power equipment defect analysis model is a lightweight neural network model deployed within the trusted execution environment. The analysis results of the power equipment defect analysis model include normal inspection, suspected defects, and confirmed faults.

[0008] Furthermore, in step one, the data value level label includes at least a first level, a second level, and a third level, wherein the first level corresponds to background stream data, the second level corresponds to routine inspection data, and the third level corresponds to critical alarm data. The critical alarm data is generated from the confirmed fault results output by the power equipment defect analysis model, or generated according to preset equipment importance rules.

[0009] Furthermore, in step two, sensing the real-time network link status between the public network link and the power private network specifically involves: collecting the signal strength, network latency, and available bandwidth of the public network link through the network probe module in the trusted sensing-driven edge agent, while simultaneously obtaining the current load status and dedicated line latency of the power private network from the power communication network management system through a secure channel. The raw network status data collected by the network probe module is periodically sent to the trusted execution environment for verification and signature.

[0010] Furthermore, in step three, the dynamic decision-making and scheduling of the transmission path based on the data value level label and the real-time network link status specifically involves the hybrid link scheduler calculating a dynamic decision matrix based on the received data value level label and verified network link status data, combined with the data packet size and the global policy issued by the centralized control platform. For critical alarm data at the third level, the decision matrix prioritizes link reliability and real-time transmission, and prioritizes scheduling power private network or public network links with the best quality for real-time and complete transmission. For routine inspection data at the second level, the decision matrix focuses on balancing link reliability, real-time performance, and bandwidth cost, and the scheduling adopts an intelligent slicing transmission strategy that uses a private network to transmit key frames and a public network to supplement non-key frames. For the first-level background stream data, the decision matrix prioritizes increasing the bandwidth cost weight, and the scheduling adopts a strategy of local caching and asynchronous compression transmission during network idle periods.

[0011] Furthermore, in step four, generating a lightweight evidence header specifically involves: in the trusted execution environment, using a hash algorithm to jointly calculate the data value level label, key feature vectors extracted from the original video stream or image, precise timestamp, unique device identifier, and GPS coordinates to generate a unique aggregate hash value, and the lightweight evidence header at least contains the aggregate hash value.

[0012] Furthermore, in step five, submitting the lightweight evidence header to the distributed ledger node for evidence storage specifically involves the trusted perception-driven edge intelligent agent, acting as a blockchain lightweight node client, sending the lightweight evidence header to a node in the private blockchain network maintained by the operation and management party through a backup communication channel independent of the business link while sending data packets through the business link, thus completing the on-chain evidence storage.

[0013] Furthermore, in step six, the verification and processing by the centralized control platform specifically involves the following steps: after receiving the data packet through the business link, the centralized control platform immediately initiates a query request to the private blockchain network to obtain the on-chain record of the lightweight evidence header corresponding to the data packet. The centralized control platform uses the same hash algorithm to recalculate the hash value of the business data portion in the received data packet, and compares the calculation result with the aggregate hash value in the on-chain record; If the comparison is consistent, the data is determined to be complete and reliable, and a high-priority alarm is displayed and a maintenance work order is dispatched based on the data value level label in the data packet. If the comparison is inconsistent or there is no evidence record on the chain, the data packet is determined to be untrustworthy, triggering the security audit process and recording an alarm.

[0014] Furthermore, the method also includes step seven, executing an anti-interference degradation strategy: when the trusted execution environment detects a security threat or an anomaly in a key sensing module during self-inspection, the hybrid link scheduler automatically triggers a security priority degradation mode, forcibly upgrading the data value level label of all data to be transmitted to the second level or above, and prioritizing the scheduling of pre-configured encrypted dedicated lines for transmission, while simultaneously sending security alarm information to the centralized control platform.

[0015] Furthermore, the method also includes step eight, performing operational closed-loop optimization: the centralized control platform comprehensively analyzes the transmission success rate, evidence verification success rate and link quality historical data of the entire network of drone nests, dynamically generates and distributes updated global scheduling strategies to each of the trusted perception-driven edge intelligent agents, the global scheduling strategy is used to adjust the cost weight and reliability weight threshold of different regions and different levels of data when selecting public network and private network.

[0016] Compared with existing technologies, this method for secure real-time video and data transmission between the UAV nest and the centralized control platform has the following advantages: I. This invention deploys a trusted perception-driven edge intelligent agent on the UAV nest side, relying on a trusted execution environment to achieve secure isolation of data processing and value level label generation, ensuring the authenticity of data value determination; it uses blockchain notarization technology to generate lightweight notarization heads for key data information on the chain, realizing end-to-end traceability and tamper-proof data from processing, transmission to verification; at the same time, it uses a hybrid link scheduler to deeply integrate the trusted verified data value level labels with real-time network link status, and makes dynamic decisions on transmission paths, thereby effectively solving the problem of data trust assurance and transmission path optimization being isolated from each other in the prior art, ensuring reliable real-time transmission of key alarm data, balancing transmission performance and cost for routine inspection data, and efficient and low-cost transmission of background stream data, significantly improving the end-to-end trustworthiness and network adaptability of data transmission.

[0017] Second, this invention, by setting an anti-interference degradation strategy, automatically upgrades data transmission priority and schedules encrypted dedicated lines when a security threat or critical module anomaly is detected in the trusted execution environment, ensuring data transmission security in special scenarios. Through an operational closed-loop optimization mechanism, the centralized control platform dynamically adjusts the global scheduling strategy based on historical data such as the overall network transmission success rate and the evidence verification success rate, continuously optimizing transmission performance. At the same time, by adopting strategies such as local caching, intelligent slicing transmission, and asynchronous compression transmission, bandwidth usage costs are effectively reduced, thereby enhancing the system's adaptability to complex network environments and abnormal equipment states, and improving the overall operational stability, reliability, and resource utilization efficiency.

[0018] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0020] Figure 1 This is a flowchart illustrating the steps of the real-time video and data secure transmission method of the present invention. Figure 2 This is a diagram of the trusted perception-driven edge intelligent agent architecture of the present invention; Figure 3 This is a closed-loop diagram of the end-to-end trusted transmission and verification of the present invention. Detailed Implementation

[0021] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0022] Example like Figures 1 to 3 As shown in the figure, this embodiment discloses a method for secure real-time video and data transmission between a UAV nest and a centralized control platform. This method is mainly applied in the operation and maintenance of power transmission lines. By deploying a trusted perception-driven edge agent on the UAV nest side, secure and efficient transmission of data collected by the UAV to the centralized control platform is achieved. The following provides a detailed description of each step of the method and related technical details.

[0023] The transmission system in this embodiment mainly includes a trusted perception-driven edge intelligent agent on the UAV nest side, a public network link, a private power grid, distributed ledger nodes, and a remote centralized control platform. The trusted perception-driven edge intelligent agent integrates core functional modules such as a trusted execution environment, a network probe module, a hybrid link scheduler, and a blockchain lightweight node client, serving as the core unit for data processing, link awareness, path scheduling, and evidence submission. The private blockchain network is maintained by the power operation and maintenance management party and is used to store lightweight evidence headers during data transmission, ensuring data traceability and immutability. The centralized control platform is deployed in the power operation and maintenance center and is responsible for data reception, verification, processing, and the generation and distribution of global scheduling strategies.

[0024] In this implementation, the detailed implementation process for each step is as follows: Data processing in a trusted execution environment: In the trusted perception-driven edge intelligent agent deployed on the drone nest side, the trusted execution environment is built based on a dedicated security chip. In this embodiment, the national cryptographic-grade security chip SM400 is selected. This chip has independent computing units, storage units and encryption modules, which can provide an isolated and secure operating environment for data processing and effectively resist security threats such as malicious software attacks and data tampering.

[0025] During the inspection of power transmission lines, drones collect high-definition visible light video streams, infrared thermal imaging video streams, and equipment status data in real time, and transmit this raw data back to the drone's nest in real time. The nest then transmits the raw data to the trusted execution environment of the trusted perception-driven edge intelligent agent through a data interface.

[0026] A power equipment defect analysis model is pre-deployed within the trusted execution environment. This model is a lightweight neural network model, specifically optimized using the MobileNetV3 architecture. The model parameter size is controlled within 5MB, and the inference latency is less than 100ms, making it adaptable to the hardware computing power limitations of edge agents. The model is trained using a large number of power equipment defect samples, covering common defect types in key equipment such as transmission line towers, insulators, conductors, and hardware, including insulator damage, conductor strand breakage, and loose hardware.

[0027] The power equipment defect analysis model processes the original video stream or image as follows: First, the original video stream is frame-by-frame extracted, with one key image extracted every 10 frames; then, the extracted images are preprocessed, including image scaling and normalization; next, the preprocessed images are input into the model for inference, and the model outputs analysis results, which include three categories: normal inspection, suspected defects, and confirmed faults.

[0028] Based on the analysis results output by the model and preset rules, data value level labels are generated, including Level 1, Level 2, and Level 3. Level 1 corresponds to background stream data, specifically video stream data collected during UAV flight that lacks power equipment or obvious defect characteristics, such as background images of the sky and ground vegetation. Level 2 corresponds to routine inspection data, specifically data related to normal inspections or suspected defects as analyzed by the model, including inspection images and video frames of normal equipment and images and video clips of suspected defect areas. Level 3 corresponds to critical alarm data, which is generated from confirmed fault results output by the model or based on preset equipment importance rules. For example, main line towers and important crossing sections in transmission lines are preset as high-importance equipment; inspection data collected for these devices, regardless of the model analysis results, are all marked as Level 3 critical alarm data.

[0029] Real-time network link status awareness: This step uses the network probe module and security channel in the trusted perception-driven edge agent to collect real-time network link status of public network links and private power grids, respectively. The network link status includes signal strength, network latency, available bandwidth, and link cost.

[0030] The network probe module employs an active probing method based on the TCP / IP protocol to collect the status of currently available public network links. The specific collection process is as follows: the network probe module sends probe data packets to a preset public network probing server every 200ms. By statistically analyzing the sending and receiving response times of the probe data packets, the network latency is calculated. The signal strength of the public network link is obtained by analyzing the signal strength indication value of the received response data packets. The available bandwidth is calculated by the ratio of the number of successfully transmitted data packets to the total bandwidth resources during 10 consecutive probes. In this embodiment, signal strength is expressed in dBm, ranging from -110dBm to -30dBm, with higher values ​​indicating stronger signals. Network latency is expressed in milliseconds, ranging from 0ms to 1000ms. Available bandwidth is expressed in Mbps, ranging from 0Mbps to 100Mbps.

[0031] Simultaneously, the current load status and leased line latency of the dedicated power network are obtained from the power communication network management system via a secure channel. This secure channel is constructed using IPsec encryption to ensure data transmission security and prevent the theft or tampering of link status data. The power communication network management system is the core of the dedicated power network's operation and management, capable of monitoring its operational status in real time. The edge agent obtains the current load status and leased line latency of the corresponding transmission link of the dedicated power network by sending query requests compliant with the IEC 61850 standard to the power communication network management system.

[0032] The raw network status data collected by the network probe module is periodically sent to the Trusted Execution Environment (TEE) for verification and signing, with a verification cycle of 1 second. The verification process is as follows: The TEE uses hash verification to calculate a hash value for the raw network status data and compares it with the verification hash value provided by the network probe module. If the comparison matches, the raw network status data is considered tamper-proof, and the verification passes. If the comparison does not match, the data is considered to be at risk of tampering, the verification fails, and the affected portion of the raw data is discarded, along with relevant alarm information. After successful verification, the TEE signs the network status data using a dedicated private key. The signing employs the RSA asymmetric encryption algorithm with a 2048-bit key length, ensuring data integrity and source reliability.

[0033] Dynamic decision-making and scheduling of transmission paths: In this step, the hybrid link scheduler receives verified and signed network link status data, as well as data value level labels generated by the trusted execution environment. Combining the packet size with the global policy issued by the centralized control platform, it calculates a dynamic decision matrix and uses this matrix to realize dynamic decision-making and scheduling of transmission paths.

[0034] The construction process of the dynamic decision matrix is ​​as follows: First, the decision indicators are determined, including link reliability, transmission real-time performance, and bandwidth cost. The weights of each indicator are dynamically adjusted according to the data value level label and the global strategy. Then, parameters such as signal strength, network latency, and available bandwidth in the network link status data are converted into quantitative values ​​of the decision indicators. For example, the quantitative value of link reliability = (quantitative score of signal strength + quantitative score of available bandwidth) / 2, where the quantitative score of signal strength is obtained by mapping the signal strength value (-110dBm to -90dBm is 1 point, -89dBm to -70dBm is 2 points, -69dBm to -50dBm is 3 points, -49dBm to -30dBm is 4 points, and so on). 0dBm is 4 points), the available bandwidth quantification score is obtained by mapping the available bandwidth value (0Mbps to 10Mbps is 1 point, 11Mbps to 30Mbps is 2 points, 31Mbps to 60Mbps is 3 points, and 61Mbps to 100Mbps is 4 points); the transmission real-time quantification value = 4 - (network latency / 250) (network latency 0ms to 250ms is 4 points, 251ms to 500ms is 3 points, 501ms to 750ms is 2 points, and 751ms to 1000ms is 1 point); the bandwidth cost quantification value is calculated based on the public network link tariff standard and available bandwidth, and the bandwidth cost of the power private network is preset to a fixed value.

[0035] Based on the aforementioned quantified values ​​and weight allocations, a comprehensive score is calculated for each link, and the dynamic decision matrix uses this comprehensive score as the basis for path selection. The global policies issued by the centralized control platform include link usage priority rules for different regions and time periods. For example, during peak electricity consumption periods, when the load on the dedicated power grid is high, its weight in routine inspection data transmission can be appropriately reduced; in remote mountainous areas and other areas with weak public network signals, the weight of the dedicated power grid is increased.

[0036] The transmission path scheduling strategy is as follows, depending on the data level: For critical alarm data at level three, the decision matrix prioritizes link reliability and real-time transmission (link reliability weight 0.5, real-time transmission weight 0.4, bandwidth cost weight 0.1). The hybrid link scheduler prioritizes links with a quantitative score of ≥3 for both link reliability and real-time transmission, selecting the link with the highest overall score as the transmission path, and prioritizing the scheduling of power grid private network or the best-quality public network link for real-time complete transmission. For example, when the quantitative score for the reliability of the power private network link is 3.5, the quantitative score for real-time transmission is 3.8, and the quantitative score for bandwidth cost is 2 yuan / GB, and the corresponding quantitative scores for a certain 5G public network link are 3.6, 3.9, and 3 yuan / GB respectively, the comprehensive score is calculated as follows: Comprehensive score for the power private network = 3.5×0.5 + 3.8×0.4 + 2×0.1 = 1.75 + 1.52 + 0.2 = 3.47; Comprehensive score for the 5G public network link = 3.6×0.5 + 3.9×0.4 + 3×0.1 = 1.8 + 1.56 + 0.3 = 3.66. In this case, the 5G public network link is selected for real-time and complete transmission of critical alarm data to ensure that the data can be transmitted to the centralized control platform quickly and reliably.

[0037] For the second-level routine inspection data, the decision matrix primarily balances link reliability, real-time performance, and bandwidth cost (link reliability weight 0.3, real-time performance weight 0.3, bandwidth cost weight 0.4). Scheduling employs an intelligent slicing transmission strategy, using a dedicated network to transmit key frames and the public network to supplement non-key frames. First, the video stream corresponding to the routine inspection data is sliced. Key frames include video frames corresponding to suspected defect areas identified by the model and video frames of key equipment components, accounting for 20% of the total frames; non-key frames are the remaining routine video frames, accounting for 80%. Then, key frames are transmitted via the dedicated power network, while non-key frames are transmitted via the public network link. If the available bandwidth of the public network link is insufficient, non-key frames can be H.265 compressed before transmission, ensuring reliable transmission of core inspection data while controlling bandwidth costs.

[0038] For the first-level background stream data, the decision matrix prioritizes increasing bandwidth cost (link reliability weight 0.1, transmission real-time performance weight 0.1, bandwidth cost weight 0.8). Scheduling employs a strategy of local caching and asynchronous compressed transmission during network idle periods. The trusted perception-driven edge agent is equipped with a 1TB local solid-state drive for storing background stream data. Network idle periods are determined by real-time network link status; specifically, when the average load rate of both the public network and the dedicated power grid is below 30% for more than 5 minutes, it is considered a network idle period. During this period, the cached background stream data is compressed using H.265 and then transmitted asynchronously through the link with the lowest current bandwidth cost. If the network status changes during transmission and the load rate exceeds 30%, transmission is paused and resumed when the idle period returns.

[0039] Generation of lightweight evidence header: In a trusted execution environment, data value level labels, feature information of the original video stream or image, timestamps, device identifiers, and geographic location information are bound together, and an aggregate hash value is calculated to generate a lightweight evidence header.

[0040] The specific implementation process is as follows: Feature information extraction: For the original video stream, feature information of key frames is extracted. The SIFT algorithm is used to detect and describe feature points of key frames. 500-1000 feature points are extracted for each key frame to generate a 128-dimensional key feature vector. For a single image, the SIFT algorithm is directly used to extract feature points and generate a 128-dimensional key feature vector.

[0041] Timestamp Acquisition: The high-precision clock module built into the Trusted Execution Environment (TEE) is used to obtain accurate timestamps. The timestamp format is YYYY-MM-DDHH:MM:SS.ssssss, accurate to the microsecond level, ensuring that the timestamp of each data packet is unique and accurate.

[0042] Device identification is determined: There is a one-to-one correspondence between the drone nest and the edge agent. Each edge agent is assigned a unique device identifier, which is represented by a 16-bit hexadecimal string. The device identifier is uniformly assigned and entered into the system by the operation and management party when the device is deployed to ensure the uniqueness of the device identifier.

[0043] Geographic location information acquisition: Geographic location information is acquired through the GPS module integrated into the edge agent, specifically GPS coordinates, with longitude and latitude accurate to 6 decimal places and altitude accurate to the meter level, ensuring the accuracy of data collection location.

[0044] After collecting the aforementioned information, the SHA-256 hash algorithm is used to jointly calculate a unique 32-byte aggregate hash value based on the data value level label, key feature vector, precise timestamp, unique device identifier, and GPS coordinates. This aggregate hash value is contained in the lightweight evidence header, which may also include some basic information as needed. The total length of the lightweight evidence header is controlled within 48 bytes to ensure it does not significantly increase the amount of data transmitted.

[0045] Data packet transmission and evidence header preservation: The data packet containing the lightweight evidence header is sent to the centralized control platform according to the transmission path scheduled in step three. The data packet has the structure of a lightweight evidence header + business data, where the business data is processed accordingly according to the transmission strategy.

[0046] Simultaneously, a lightweight evidence header is submitted to the distributed ledger node for evidence storage. The trusted perception-driven edge intelligent agent, acting as a blockchain lightweight node client, supports communication protocols with the private blockchain network. This private blockchain network is maintained by the operator and adopts a consortium blockchain architecture. Its nodes include an operations and maintenance center node, regional management nodes, and audit nodes, with a total of 10 nodes deployed to ensure the decentralization and reliability of the blockchain network.

[0047] During the notarization process, the edge agent transmits a lightweight notarization header through a backup communication channel independent of the business link. This backup channel utilizes a 4G network to ensure successful on-chain notarization even if the business link fails. Following the blockchain network's transaction format, the edge agent encapsulates the lightweight notarization header into transaction data and sends it via a P2P network to any node in the private blockchain network. Upon receiving the transaction data, the node verifies the transaction. If verification is successful, the node broadcasts the transaction to the entire blockchain network. Other nodes synchronize the transaction and record it in their local ledgers, completing the on-chain notarization. After notarization is complete, the blockchain network returns a successful notarization response to the edge agent, containing the transaction hash value and block height. The edge agent records this response information for future reference.

[0048] Verification and processing of the centralized control platform: After receiving the data packet through the business link, the centralized control platform immediately initiates the verification process, which is implemented as follows: Evidence storage record query: The centralized control platform initiates a query request to the private blockchain network through the blockchain client. The query request carries key information such as lightweight evidence storage header or device identifier and timestamp in the data packet. The blockchain network retrieves the corresponding on-chain evidence storage record according to the query conditions and returns the evidence storage record to the centralized control platform.

[0049] Hash value recalculation: The centralized control platform extracts the business data portion from the received data packet, extracts feature information, obtains data value level labels, timestamps, device identifiers, and geographical location information using the same method as in step four, and recalculates the aggregate hash value using the SHA-256 hash algorithm.

[0050] Verification by comparison: The recalculated aggregate hash value is compared with the aggregate hash value in the on-chain evidence record. If the two are completely consistent, the data is determined to be complete, reliable, and untampered; if the comparison is inconsistent or there is no corresponding evidence record on the chain, the data packet is determined to be untrustworthy.

[0051] Based on the verification results and data value level labels, the centralized control platform performs corresponding processing: When the data is reliable: For Level 3 critical alarm data, the centralized control platform immediately triggers a high-priority alarm, displaying alarm information in a red pop-up window on the monitoring interface, including device identification, geographical location, fault type, timestamp, etc., and automatically generates a maintenance work order, which is then dispatched to the corresponding maintenance personnel. For Level 2 routine inspection data, the centralized control platform stores the data in the corresponding database, and simultaneously marks suspected defective data and pushes it to the manual review interface for further confirmation by professionals. For Level 1 background stream data, the centralized control platform performs routine storage for subsequent traceability or data analysis.

[0052] When data is unreliable: The centralized control platform triggers a security audit process, records alarm information, suspends subsequent data processing of the device, and notifies maintenance personnel to inspect the drone's nest and edge intelligence agent to investigate security risks in the data transmission process.

[0053] Implementation of anti-interference degradation strategy: The Trusted Execution Environment has a self-testing function, which performs a self-test every 5 seconds on its own operating status, the working status of the security chip, and the working status of the key perception modules to detect whether there are security threats or module abnormalities.

[0054] When the self-test detects a security threat or an anomaly in a critical sensing module, the hybrid link scheduler automatically triggers a security-first degradation mode: Data level adjustment: The data value level label of all data to be transmitted will be forcibly upgraded to the second level or above. Data originally in the first level will be upgraded to the second level, data originally in the second level will remain unchanged, and data originally in the third level will remain in the third level, to ensure the transmission priority of core data.

[0055] Transmission path adjustment: Priority will be given to using pre-configured encrypted dedicated lines for transmission. These lines are dedicated encrypted channels within the power grid, employing the AES-256 encryption algorithm to encrypt transmitted data, further ensuring data transmission security. If the encrypted dedicated line is unavailable, the public network link with the best current network status (link reliability score ≥ 3 points, transmission real-time performance score ≥ 3 points) will be selected for transmission.

[0056] Security alarm sending: Simultaneously send security alarm information to the centralized control platform. The alarm information includes the device identifier, anomaly type, occurrence time, and current data transmission strategy adjustment status, so that the centralized control platform can promptly grasp the device status and take corresponding measures.

[0057] Execution of operational closed-loop optimization: The centralized control platform collects and stores real-time data on transmission-related information of all drones in the network, including transmission success rate, certificate verification success rate, and historical link quality data, with a statistical period of 1 hour.

[0058] The centralized control platform uses big data analytics algorithms to comprehensively analyze the aforementioned historical data. This includes analyzing the stability and cost-effectiveness of various links in different regions and time periods, as well as the transmission requirements and link compatibility for different data levels. Based on the analysis results, an updated global scheduling strategy is dynamically generated. This strategy includes cost and reliability weight thresholds for selecting public or private networks for different regions and data levels.

[0059] For example, if analysis reveals that in a certain area between 2:00 PM and 4:00 PM, the average available bandwidth of the public network link is low (≤10Mbps), with a transmission success rate of only 85%, while the load rate of the power grid is low (≤40%), with a transmission success rate of 99%, then the global strategy for that area during that period should be adjusted: the reliability weight of the power grid for second-level routine inspection data should be increased from 0.3 to 0.4, the cost weight of the public network link should be decreased from 0.4 to 0.3, and the reliability weight threshold of the power grid should be adjusted from 3 points to 2.8 points to ensure that more routine inspection data is transmitted through the power grid and improve the transmission success rate.

[0060] The centralized control platform distributes the updated global scheduling policy to each trusted perception-driven edge agent through a secure channel. After receiving the policy, the edge agents update their local policy configuration for dynamic decision-making on subsequent transmission paths, forming an operational closed-loop optimization that continuously improves the performance and reliability of the entire transmission system.

[0061] This invention integrates data processing, link awareness, path scheduling, and evidence generation into one by deploying a trusted perception-driven edge agent on the UAV nest side. This achieves a deep fusion of data value and network status, and can dynamically adjust transmission strategies according to data importance and actual network conditions. While ensuring the real-time and reliable transmission of critical data, it effectively reduces bandwidth costs.

[0062] By combining a trusted execution environment with blockchain-based evidence storage technology, the trustworthiness of data is ensured throughout the entire chain from processing and value determination to transmission and verification. This effectively prevents security risks such as data tampering and forgery, and provides strong support for the reliability of power inspection data.

[0063] The anti-interference degradation strategy and operation closed-loop optimization settings enhance the stability and adaptability of the system, enabling it to cope with complex network environments and equipment anomalies, continuously optimize transmission performance, and meet the high security, high reliability, and high efficiency requirements for data transmission in power transmission line operation and maintenance scenarios.

[0064] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for secure real-time video and data transmission between a UAV nest and a centralized control platform, characterized in that, Deploying a trusted perception-driven edge agent at the drone nest side, the method includes the following steps: Step 1: The data is processed in the trusted execution environment on the UAV nest side. The processing includes analyzing the original video stream or images transmitted by the UAV through a power equipment defect analysis model, and generating data value level labels based on the analysis results and preset rules. Step 2: Sensing the real-time network link status between the public network link and the private power grid. The network link status includes signal strength, network latency, available bandwidth, and link cost. Step 3: Based on the data value level label and the real-time network link status, the transmission path is dynamically decided and scheduled through the hybrid link scheduler. Step 4: Bind the data value level label, the feature information of the original video stream or image, the timestamp, the device identifier, and the geographical location information in the trusted execution environment and calculate the aggregate hash value to generate a lightweight evidence header; Step 5: Send the data packet containing the lightweight evidence header to the centralized control platform through the scheduled transmission path, and simultaneously submit the lightweight evidence header to the distributed ledger node for evidence storage. Step six: After receiving the data packet, the centralized control platform queries the corresponding evidence storage record from the distributed ledger node for verification, and processes it accordingly based on the verification result and the data value level label.

2. The method for secure real-time video and data transmission between the UAV nest and the centralized control platform according to claim 1, characterized in that, In step one, the trusted execution environment is built on a dedicated security chip, and the power equipment defect analysis model is a lightweight neural network model deployed within the trusted execution environment. The analysis results of the power equipment defect analysis model include normal inspection, suspected defects, and confirmed faults.

3. The method for secure real-time video and data transmission between the UAV nest and the centralized control platform according to claim 1, characterized in that, In step one, the data value level label includes at least a first level, a second level, and a third level, wherein the first level corresponds to background stream data, the second level corresponds to routine inspection data, and the third level corresponds to critical alarm data. The critical alarm data is generated from the confirmed fault results output by the power equipment defect analysis model, or generated according to preset equipment importance rules.

4. The method for secure real-time video and data transmission between the UAV nest and the centralized control platform according to claim 1, characterized in that, In step two, sensing the real-time network link status of the public network link and the power private network specifically involves: collecting the signal strength, network latency, and available bandwidth of the public network link through the network probe module in the trusted sensing-driven edge agent, and simultaneously obtaining the current load status and dedicated line latency of the power private network from the power communication network management system through a secure channel. The raw network status data collected by the network probe module is periodically sent to the trusted execution environment for verification and signature.

5. The method for secure real-time video and data transmission between the UAV nest and the centralized control platform according to claim 1, characterized in that, In step three, the dynamic decision-making and scheduling of transmission paths based on the data value level label and the real-time network link status specifically involves the hybrid link scheduler calculating a dynamic decision matrix based on the received data value level label and verified network link status data, combined with the data packet size and the global policy issued by the centralized control platform. For critical alarm data at the third level, the decision matrix prioritizes link reliability and real-time transmission, and prioritizes scheduling power private network or public network links with the best quality for real-time and complete transmission. For routine inspection data at the second level, the decision matrix focuses on balancing link reliability, real-time performance, and bandwidth cost, and the scheduling adopts an intelligent slicing transmission strategy that uses a private network to transmit key frames and a public network to supplement non-key frames. For the first-level background stream data, the decision matrix prioritizes increasing the bandwidth cost weight, and the scheduling adopts a strategy of local caching and asynchronous compression transmission during network idle periods.

6. The method for secure real-time video and data transmission between the UAV nest and the centralized control platform according to claim 1, characterized in that, In step four, generating a lightweight evidence header specifically involves: in the trusted execution environment, using a hash algorithm to jointly calculate the data value level label, key feature vectors extracted from the original video stream or image, precise timestamp, unique device identifier, and GPS coordinates to generate a unique aggregate hash value, and the lightweight evidence header contains at least the aggregate hash value.

7. The method for secure real-time video and data transmission between the UAV nest and the centralized control platform according to claim 1, characterized in that, In step five, submitting the lightweight evidence header to the distributed ledger node for evidence storage specifically involves the trusted perception-driven edge intelligent agent, acting as a blockchain lightweight node client, sending the lightweight evidence header to a node in a private blockchain network maintained by the operation and management party through a backup communication channel independent of the business link while sending data packets through the business link, thus completing the on-chain evidence storage.

8. The method for secure real-time video and data transmission between the UAV nest and the centralized control platform according to claim 1, characterized in that, In step six, the verification and processing by the centralized control platform specifically involves the following steps: After receiving the data packet through the business link, the centralized control platform immediately initiates a query request to the private blockchain network to obtain the on-chain record of the lightweight evidence header corresponding to the data packet. The centralized control platform uses the same hash algorithm to recalculate the hash value of the business data portion in the received data packet, and compares the calculation result with the aggregate hash value in the on-chain record; If the comparison is consistent, the data is determined to be complete and reliable, and a high-priority alarm is displayed and a maintenance work order is dispatched based on the data value level label in the data packet. If the comparison is inconsistent or there is no evidence record on the chain, the data packet is determined to be untrustworthy, triggering the security audit process and recording an alarm.

9. The method for secure real-time video and data transmission between the UAV nest and the centralized control platform according to claim 1, characterized in that, The method further includes step seven, executing an anti-interference degradation strategy: when the trusted execution environment detects a security threat or an abnormality in a key sensing module during self-inspection, the hybrid link scheduler automatically triggers a security priority degradation mode, forcibly upgrades the data value level label of all data to be transmitted to the second level or above, prioritizes scheduling the pre-configured encrypted leased line for transmission, and sends security alarm information to the centralized control platform.

10. The method for secure real-time video and data transmission between the UAV nest and the centralized control platform according to claim 1, characterized in that, The method also includes step eight, performing operational closed-loop optimization: the centralized control platform comprehensively analyzes the transmission success rate, evidence verification success rate and link quality historical data of the entire network of drone nests, dynamically generates and distributes updated global scheduling strategies to each of the trusted perception-driven edge intelligent agents, the global scheduling strategy is used to adjust the cost weight and reliability weight threshold of different regions and different levels of data when selecting public network and private network.

Images