A link dynamic selection method for high-voltage power transmission pipe gallery inspection data communication

Abstract

The application belongs to the technical field of ramjet engine, and discloses a link dynamic selection method for high-voltage power transmission pipe gallery inspection data communication. The link dynamic selection method comprises the following steps: determining an overall optimization target; establishing hardware components and software modules running on the hardware components; establishing the connection relationship and the electrical signal flow direction among the components; and performing link dynamic selection. The link dynamic selection method proposes a multi-dimensional link quality evaluation model integrating three key indicators of the physical layer, the link layer and the network layer; introduces a content-aware mechanism to realize content-demand-oriented link matching; designs a data preprocessing strategy to reduce transmission load from the source; realizes adaptive data packeting, and improves the dynamic optimization transmission efficiency and reliability. The link quality evaluation module, the content-aware link selection, the data preprocessing and the adaptive packeting are organically integrated to form a collaborative optimization overall framework, rather than an isolated optimization means, and have engineering practical value.

Description

Technical Field

[0001] This invention belongs to the field of ramjet engine technology, specifically relating to a dynamic link selection method for data communication in high-voltage power transmission corridor inspection. Background Technology

[0002] High-voltage underground power transmission tunnels serve as crucial infrastructure for large-scale experimental facilities, housing numerous high-voltage cast-in-place busbars and power cables, among other auxiliary equipment. To ensure the safe and stable operation of these tunnels, a large number of sensor nodes are deployed to monitor temperature, humidity, smoke, and toxic gases, collecting environmental data, equipment status data, and video surveillance data to construct a comprehensive monitoring and early warning system. Due to the risks of electromagnetic interference and high-voltage discharge, wireless communication is typically used to network and transmit sensor data to a monitoring center for centralized analysis and management. In the unique environment of underground tunnels, the propagation characteristics of wireless signals are extremely complex and dynamically changing due to the enclosed space, the shielding and attenuation effect of the concrete structure on wireless signals, and the potential electromagnetic interference (EMI) generated by high-voltage power equipment. Furthermore, different types of monitoring data (e.g., low-rate sensor data, text alarms, audio distress calls, and high-definition video) have vastly different requirements for the Quality of Service (QoS) of the communication link, with varying demands for latency, reliability, and bandwidth.

[0003] In existing monitoring systems for underground power transmission tunnels, the optimization and management of wireless communication links face numerous challenges. Traditional wireless communication methods often struggle to achieve efficient, reliable communication that meets diverse service requirements in such complex environments.

[0004] Existing Solution 1: A method based on static frequency band allocation or simple signal strength selection. This solution typically pre-assigns fixed operating frequency bands for different communication tasks or devices, or performs simple channel selection or switching based solely on a single indicator such as Received Signal Strength Indicator (RSSI). Specifically, all sensor data might be assigned to be transmitted through low-frequency bands to take advantage of their better penetration, while video data might be assigned to be transmitted through high-frequency bands to take advantage of their larger bandwidth. Alternatively, the device might only monitor the RSSI of the currently connected frequency band, and when the RSSI falls below a certain threshold, attempt to switch to another frequency band with a higher RSSI.

[0005] However, this scheme suffers from several problems due to its use of fixed frequency band allocation or link evaluation and selection based solely on a single metric such as RSSI. First, RSSI, as a physical layer metric, cannot fully reflect the actual data transmission capacity and reliability of the link. For example, high RSSI in high-interference environments may be accompanied by high Bit Error Rate (BER) or Packet Loss Rate (PLR). Second, static allocation or selection based on a single metric cannot adapt to the rapid dynamic changes in the underground utility tunnel channel environment. This may lead to a significant performance degradation of a certain frequency band due to interference or fading at a specific moment, while other available frequency bands remain underutilized. Most critically, this method completely ignores the different priorities of different types of communication content regarding link parameters (such as BER, PLR, and throughput). For example, for text alarm information requiring extremely high reliability, even if a frequency band has high throughput and a good RSSI, if its instantaneous BER or PLR is high, the link is not suitable for transmitting alarms. Conversely, for periodic sensor data with less stringent real-time requirements, lower throughput or even occasional packet loss may be tolerated, with greater emphasis placed on transmission coverage or low power consumption. Static methods cannot dynamically adjust the evaluation weights and selection strategies for link quality parameters based on content type.

[0006] Existing Solution 2: Communication performance monitoring and coarse-grained resource allocation based on link-layer or network-layer metrics. This solution may monitor link-layer or network-layer metrics such as packet loss rate (PLR) and throughput (Thr). Some more complex solutions may implement dynamic spectrum access (DSA) or dynamic channel allocation to attempt to improve spectrum utilization.

[0007] While these methods consider metrics that are closer to actual transmission performance than RSSI, they still have limitations. Firstly, many methods still rely on a combination of single or a few metrics, lacking a truly multi-dimensional link quality assessment model, making it difficult to comprehensively and accurately depict the real channel conditions of different frequency bands under complex electromagnetic environments. In particular, they fail to effectively integrate key metrics from different layers such as the Physical Layer (BER), Link Layer (PLR), and Throughput for comprehensive evaluation. Secondly, although dynamic resource allocation may be implemented, these strategies often tend to optimize aggregate performance metrics (such as maximizing total throughput or user fairness), failing to deeply understand the content type of the data to be transmitted and its differentiated QoS requirements, and thus failing to perform weighted evaluation and refined resource matching of link quality parameters accordingly. For example, a strategy that maximizes total throughput might allocate all bandwidth to video, resulting in the inability to transmit important low-rate alarm information in a timely and reliable manner. Furthermore, these methods typically fail to effectively integrate edge computing capabilities, such as performing content-related preprocessing before data transmission (e.g., text summarization, video keyframe extraction), and fail to adaptively adjust transmission parameters such as packet size based on real-time link quality, thereby further optimizing resource utilization efficiency and transmission reliability.

[0008] Existing technologies suffer from the following prominent problems when applied to complex environments with diverse business requirements, such as underground high-voltage power transmission tunnels: a lack of comprehensive, real-time, and multi-dimensional quality assessment capabilities for wireless links across different frequency bands; failure to effectively perceive the content type of the data to be transmitted and its specific differentiated Quality of Service (QoS) requirements; failure to differentiate the weighting of link quality parameters based on content type and to select links accordingly; and the fact that existing optimization methods are often isolated, failing to coordinate and optimize link assessment, content awareness, data preprocessing, and adaptive transmission parameter adjustment. These problems make it difficult for existing methods to achieve efficient, reliable communication that meets diverse business needs and fully utilizes resources in this environment, especially in scenarios requiring the simultaneous transmission of multiple data types while prioritizing critical information (such as alarms).

[0009] Currently, there is an urgent need to develop a dynamic link selection method for data communication in high-voltage power transmission tunnel inspection. Summary of the Invention

[0010] To address the shortcomings in comprehensive link quality assessment, content-aware adaptability, dynamic environment adaptability, and multi-stage collaborative optimization in the complex environment of underground high-voltage power transmission tunnels when facing diverse monitoring and early warning data transmission needs, this invention proposes a wireless communication link optimization method based on multi-dimensional link quality assessment and content awareness for use in underground high-voltage power transmission tunnels. By deploying edge nodes within the underground tunnel, the method can perceive the wireless channel status and characteristics of the data to be transmitted in real time, dynamically select the optimal communication link, and optimize transmission parameters to ensure efficient and reliable transmission of critical monitoring and early warning information.

[0011] The link dynamic selection method for high-voltage transmission tunnel inspection data communication of the present invention solves the following technical problems: (1) It solves the problem that existing technologies rely on only a single or a few performance indicators (such as RSSI, simple PLR / Thr) to evaluate the quality of wireless communication links, which cannot fully and accurately reflect the real transmission performance in complex channel environments (such as underground pipe corridors); (2) It solves the problem that existing technologies fail to effectively perceive the content type of the data to be transmitted and its differentiated quality of service requirements; (3) To address the problem that existing technologies fail to weight and match the importance of link quality parameters (such as reliability and rate) according to different content types when selecting links; (4) Solve the problem that existing link optimization methods (such as link evaluation and resource allocation) are isolated from each other and fail to coordinate and optimize data preprocessing, link selection and adaptive transmission parameter adjustment, resulting in low resource utilization efficiency and difficulty in ensuring reliable transmission of key data. (5) Solve the problem that existing technologies are unable to achieve efficient, reliable communication that can meet the diverse business needs in complex and dynamically changing wireless environments (especially underground tunnels for high-voltage power transmission), especially when processing multiple types of data at the same time and ensuring the priority of key information.

[0012] The link dynamic selection method for high-voltage transmission tunnel inspection data communication of the present invention includes the following steps: S10. Determine the overall optimization objective; S20. Establish hardware components and software modules running on the hardware components; S30. Establish the connection relationships between various components and the flow of electrical signals; S40. Perform dynamic link selection.

[0013] Furthermore, S10 includes the following: A multi-dimensional wireless communication link quality assessment model is constructed, which integrates three key indicators: bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr), to comprehensively, in real time, and accurately characterize the channel status in complex environments such as underground power transmission tunnels. Construct a content-aware adaptive link selection strategy to identify the types of data units to be transmitted, including alarms, video, and sensor data, as well as QoS constraints. Dynamically adjust the evaluation weights of link quality parameters, including bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr), based on the content type to match different data streams to the corresponding wireless frequency bands. This integrates multiple optimization steps, including multi-dimensional link quality assessment, content-aware link selection, edge data preprocessing based on data type and emergency events, and adaptive packet size adjustment based on real-time link quality, into a collaborative comprehensive algorithm framework to achieve dynamic, efficient, and reliable optimization of monitoring and early warning communication links in complex and ever-changing environments. Furthermore, the software modules in S20 include: a. Link Quality Assessment (LQA) module: Responsible for real-time acquisition and processing of wireless link performance data between the device and edge devices; it has the following functions: Obtain physical layer information: Obtain physical layer feedback information by calling the application programming interface of the wireless transceiver chip to assist in estimating the bit error rate (BER); Establish a probing mechanism: periodically or on demand send known probe data packets or sequences to edge devices and receive responses or reports from edge devices; by comparing the sent and received data content and packet header information, calculate the bit error rate (BER) and packet loss rate (PLR). Throughput measurement: During actual data transmission, monitor the total amount of valid data successfully sent and acknowledged within a pre-set time period; divide the total amount of valid data by the length of the time period to obtain the effective transmission rate Thr; Storage data structure: The Link Quality Assessment (LQA) module stores the bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr) measured by each edge device on each available frequency band in an internally maintained link state table; b. Content-Aware Link Selection Module: Receives data units to be transmitted and real-time link status information from the Link Quality Assessment (LQA) module, and makes link selection decisions; it has the following functions: Data Unit Attribute Acquisition: When sensor readings are packaged, video frames are generated, and alarm signals are triggered, edge devices generate or transmit corresponding data units. Each data unit or edge node is accompanied by an identified type (Type) and associated attributes. Types include Sensor data, Text, Alarm, Audio, and Video. Associated attributes include data size (PSize), latency (DelayReq), and maximum tolerable packet loss rate. Maximum tolerable bit error rate The type and associated attributes are configured on the edge device or determined through data content analysis; if a reading exceeding the threshold is received from the smoke sensor, a data unit of type Alarm is generated, with extremely low latency DelayReq and extremely high reliability requirements set; extremely low and extremely high data values ​​are preset. Query link status: Query the bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr) of all available links at the current moment from the link status table maintained by the Link Quality Assessment (LQA) module. Constraint judgment: Traverse all available links, and for each link, determine the maximum tolerable delay (DelayReq) and maximum tolerable packet loss rate for each data unit. Maximum tolerable bit error rate The data size (PSize), along with the link's bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr), are used to determine whether QoS constraints are met; if not, the data is excluded. QoS constraints include... , and ; Obtaining weights and scores: Based on the data unit type (Type), find the preset or dynamically determined weighting factors for the bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr). , , Pre-set weighting factors and , The numerical range; for Type=Alarm, and The value is higher than For Type=Video, The values ​​are relatively high; based on the bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr) provided by the link quality assessment module (LQA), these values ​​are substituted into the preset scoring function for BER, PLR, and Thr. Calculate the score for each item; Utility calculation and selection: For each link that satisfies the constraints, use the weighted sum formula. Calculate the corresponding utility value; select the link with the largest utility value as the target link for this data unit transmission; c. Data Preprocessing Module: Before data units enter the transmission queue, the data preprocessing module preprocesses the data units according to their type and context; it has the following functions: Perform type identification and event judgment: Obtain the type (Type) of the data unit and determine whether there is an emergency event based on the system status and monitoring logic; for the corresponding video type, the edge node continuously runs the video analysis algorithm to detect smoke or flames; for the corresponding audio type, the audio analysis algorithm runs to detect people calling for help or abnormal equipment noises; If the type is Video and a fire, smoke, or unusual human activity event is detected, then keyframe extraction or short video clip extraction is performed. If the type is Audio and abnormal sound events such as people calling for help or explosions are detected, then the audio segment is captured and transmitted. If the type is Text and it is not associated with an alarm event, then text summary processing will be performed according to the configuration; Text includes logs and regular reports. If the type is Alarm, no preprocessing is performed to prevent potential loss of information, and the integrity and timeliness of the original data are guaranteed first. After preprocessing, update the PSize attribute of the data unit; d. Adaptive Data Packet Segmentation Module: After the data unit is selected for the transmission link, the adaptive data packet segmentation module determines the actual data packet size for transmission; the specific process is as follows: Query real-time link quality: Obtain the best link determined by the content-aware link selection module. Obtain the bit error rate (BER) and packet loss rate (PLR) at the current moment from the link quality assessment module (LQA). Configure mapping and adjust PSize: Based on the bit error rate (BER) and packet loss rate (PLR), determine the packet size PSize for this transmission by consulting the link status table or through a mapping function; define three PSize levels (high, medium, and low) and set thresholds for BER and PLR. When both BER and PLR are below the minimum threshold, the link quality is good, and the smallest PSize is selected; when either BER or PLR is above the maximum threshold, the link quality is poor, and the largest PSize is selected; when either is between the two, the link quality is average, and a medium or smaller PSize is selected as needed; for alarms of extremely high priority or extremely stringent reliability requirements, even if the link quality is good, configure the largest PSize to further reduce risk; pre-set the range of PSize values, and simultaneously set the maximum, minimum, medium, and small ranges for PSize.

[0014] Furthermore, in S30, each software module implements function calls and data transfer within the processing unit of the edge computing node; the specific process is as follows: S31. Edge devices periodically or event-drivenly generate raw data units and transmit the raw data units, along with information including their type, priority, and latency requirements, to the data receiving or processing interface on the edge computing node. S32. The Link Quality Assessment (LQA) module on the edge computing node continuously interacts with the edge device by sending probe packets, receiving physical layer feedback, and monitoring the actual transmitted data stream, and periodically updates its internal link status table; the link status data in the link status table is a shared resource for all software modules that need to select links; S33. When a new data unit to be transmitted arrives at the edge computing node, it first enters the data preprocessing module; the data preprocessing module decides whether to perform preprocessing and how to process it based on the data type and possible emergency event detection results, and outputs the processed data unit and the updated Size attribute; S34. The preprocessed data unit and its attributes are passed to the content-aware link selection module; the content-aware link selection module simultaneously obtains the latest link status information from the link quality assessment module (LQA), performs fitness assessment and optimal link selection algorithm based on data attributes and link quality information, and outputs the target wireless link ID or frequency band of the preprocessed data unit. S35. The selected target link ID and data unit are passed together to the adaptive data packet segmentation module; the adaptive data packet segmentation module, based on the current bit error rate (BER) and packet loss rate (PLR) of the target link, queries the link quality assessment module (LQA) again or uses the most recent LQA result to determine the data packet size (PSize) for this transmission. S36. Finally, the preprocessed data unit is divided into data packets of a defined size and sent to the target edge device or next-hop node through the wireless communication interface of the edge node using the selected frequency band; during the data packet transmission and reception process, the Link Quality Assessment module (LQA) continues to monitor and update the link status.

[0015] Furthermore, the specific process of S40 is as follows; In the wireless environment of underground utility tunnels, a closed-loop process of continuous sensing, evaluation, decision-making, and execution is established through dynamic link selection. Real-time link evaluation and content-demand-weighted intelligent decision-making are performed, and transmission performance and resource utilization are optimized by dynamically adjusting communication strategies. The Link Quality Assessment (LQA) module monitors or probes wireless links in various available frequency bands, calculates key performance indicators including Bit Error Rate (BER), Packet Loss Rate (PLR), and Effective Transmission Rate (Thr), and stores them in the link state database. The simplified steps are as follows: The arriving data units first enter the data preprocessing module; the data preprocessing module performs preprocessing according to the preprocessing rules and event detection model in the system configuration. The event detection model is set to detect fire, smoke or people calling for help, and performs type and content processing on the data units, extracts video keyframes or text summaries, and then outputs the processed data and the updated data packet size PSize. Next, the processed data unit is sent to the content-aware link selection module. The content-aware link selection module obtains real-time link quality information from the link status database and combines it with the selection weights and QoS constraints set for different data types in the system configuration. It first checks whether each available link meets the constraints of the data unit, then calculates the utility value based on content type for the links that meet the conditions, and finally selects the link with the best utility. If no suitable link can be found, the processing failure process is entered, and the data is cached or discarded. If the optimal link is selected, the link quality information along with the data unit is transmitted to the adaptive data packet segmentation module. The adaptive data packet segmentation module obtains the current bit error rate (BER) and packet loss rate (PLR) of the optimal link from the link status database, and determines the data packet size (PSize) to be used in this transmission based on the packet size threshold or rules in the system configuration. Finally, based on the determined data packet size, the data unit is segmented into data packets and wirelessly transmitted through the selected optimal link; feedback information during the transmission process is fed back to the monitoring part of the Link Quality Assessment (LQA) module to update the link status and form a closed loop; the processing flow of a single data unit is re-evaluated after successful transmission or processing failure. The steps for dynamic link selection are as follows: S41. System startup and initialization; The edge computing node starts up, initializes the Link Quality Assessment (LQA) module, and begins periodically probing or monitoring the link quality with associated edge devices on all available frequency bands, and establishes and updates the link status table; the content-aware link selection module, data preprocessing module, and adaptive data packet segmentation module enter standby state, loading configurations including data type attributes, weights, scoring functions, preprocessing strategies, and packet segmentation thresholds; S42. Data generation and reception; Edge devices continuously sense the environment or their own status, generating different types of data units, including sensor data, text, audio, and video. These data units are either initially packaged and transmitted to the edge node directly at the edge device, or collected by the edge node on demand. When emergency events, including fires and calls for help, occur, the corresponding alarm information (Type=Alarm) is generated with high priority. S43. Data preprocessing; Data units arriving at edge nodes are first processed by the data preprocessing module; depending on the data type and whether it is associated with an emergency event, corresponding compression, filtering, or key frame / key segment extraction are performed; then the processed data units are output; when smoke is detected, video preprocessing extracts key video frames. S44. Perform content-aware link selection; The preprocessed data units are sent to the content-aware link selection module. The content-aware link selection module queries the latest link status table of the link quality assessment module (LQA) to obtain the real-time multi-dimensional quality of all available links; it finds the corresponding weights and QoS constraints according to the data unit type and urgency; for links that meet the constraints, it calculates the weighted utility value and selects the link with the highest utility as the transmission path. S45. Perform adaptive data packetization; After determining the transmission link, the adaptive data packet segmentation module queries the link's real-time bit error rate (BER) and packet loss rate (PLR); based on the BER and PLR, it dynamically adjusts the data packet size (PSize) to balance efficiency and reliability. S46. Perform data transmission and link quality assessment module LQA update; After adjusting the data packet size (PSize), the data is segmented into data packets and transmitted through the selected wireless link. During transmission, the Link Quality Assessment (LQA) module continues to monitor the transmission status of the corresponding link and update the link status table. S47. Perform periodic optimization; The entire assessment, decision-making, and execution process is carried out periodically. By reassessing the priorities of links and data and adjusting strategies at fixed intervals, the process adapts to dynamic changes in the environment and business. For high-priority data, the execution decision-making process is triggered immediately.

[0016] Furthermore, the software module also includes a log recording and analysis module, which records information including each link quality assessment result, content-aware decision-making process, selected link, preprocessing behavior, and adaptive packet sub-parameters, for system performance analysis, fault diagnosis, and subsequent algorithm optimization; it also includes a visualization interface, which displays real-time link quality, transmission status of various types of data, and link selection results.

[0017] Furthermore, the Link Quality Assessment (LQA) module is responsible for real-time and comprehensive evaluation of the communication link quality of edge nodes and associated edge devices across all available wireless frequency bands. The LQA module comprehensively considers the Bit Error Rate (BER), Packet Loss Rate (PLR), and Effective Transmission Rate (Thr). Edge nodes periodically obtain estimates of BER, PLR, and Thr through probing or monitoring mechanisms. The link quality at time t is represented by a vector. a. The function expression for the bit error rate (BER) is as follows: Bit error rate (BER) is used to measure the probability of errors in physical layer bit transmission; it is estimated by monitoring the error correction performance of forward error correction (FEC) or by sending known sequences. The estimation formula is: ; in, This indicates that within a time window ending at time t, on the link Total number of bits sent. Indicates the number of bits that were erroneous at the receiving end; b. The packet loss rate (PLR) is expressed as a function of: Packet loss rate (PLR) measures the probability of packet loss, comprehensively reflecting channel errors, interference, and collisions. It is calculated by monitoring packet sequence numbers and using ACK or NACK mechanisms. The formula for estimating PLR is: ; in, This indicates that within a time window ending at time t, on the link Total number of data packets sent. Indicates the number of data packets that failed to be successfully received or acknowledged; c. The effective transmission rate Thr is expressed as a function of: The effective transmission rate (Thr) measures the number of valid data bits successfully transmitted per unit of time; it is calculated by monitoring the total amount of data successfully transmitted within a specific time interval; the formula for estimating the effective transmission rate (Thr) is: ; in, This indicates a sequence of lengths ending at time t. During the time interval, on the link The total number of valid data bits successfully transmitted. The vector output by the Link Quality Assessment (LQA) module This provides refined input for subsequent content-aware link selection.

[0018] Furthermore, the content-aware link selection module intelligently selects the optimal transmission link or frequency band based on the content attributes of the data unit to be transmitted and the real-time multi-dimensional link quality assessment results; including the following tasks: a. Data Content Awareness: Identify the type (Type) of the data unit to be transmitted; where Sensor data includes temperature, humidity, and gas concentration; Text information includes device status and system logs; Alarm information includes fire, gas leak, and intrusion alarms; Audio includes personnel calls for help and abnormal equipment sounds; Video includes on-site monitoring footage; Simultaneously, acquire key attributes of the data unit, including but not limited to: pre-processed data size (Size), maximum tolerable latency (DelayReq), and minimum reliability requirements, including the maximum tolerable packet loss rate. and maximum tolerable bit error rate The above attributes are preset or dynamically determined based on the data type and its importance and application requirements in the monitoring and early warning system. Among them, the fire alarm information Type=Alarm has low latency and high reliability requirements. b. Fitness evaluation: At time t, for each data unit to be transmitted... and every available wireless link Calculate a quantization link Transmission Data Unit A function of fitness or utility First, check the link. real-time quality Does it meet the data unit requirement? The Quality of Service (QoS) constraints include: Reliability Constraint 1: ; Reliability Constraint 2: ; Latency or rate constraints: Link The effective transmission rate must be sufficient in data units Transmit the corresponding amount of data within the maximum tolerable delay: ; If any constraint is not satisfied, the corresponding link is considered unsuitable, and the function is... The utility value is set to a pre-defined, extremely small negative value; c. Weighted utility calculation: If the link satisfies all constraints, then based on the data unit type and its attributes, it is a function. Utility value: ; in, It is for data units The non-negative weighting factor determined by the data type reflects the degree of emphasis different data types place on bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr), and satisfies... Among them, for alarm information Type=Alarm, and High weight; for video with Type=Video, High weighting; for ordinary sensor data Type=Sensor, reliability weighting is high. It is to include link quality indicators The scoring function maps to the [0,1] interval, with higher scores for indicators that favor transmission; that is, lower BER, lower PLR, and higher Thr scores. Therefore: ; in, It is the maximum acceptable bit error rate. It is the upper limit of packet loss rate or a normalized reference value. This is a reference throughput value, which is the maximum theoretical rate or average rate achieved by the system. d. Optimal Link Selection: Select the link with the highest utility value as the optimal link for this transmission. ; If none of the links meet the constraints, the corresponding data units cannot be transmitted, preprocessed, or cached temporarily.

[0019] Furthermore, the data preprocessing module performs intelligent preprocessing on some data units at the edge node or edge device to reduce the actual amount of data transmitted and optimize bandwidth utilization; the preprocessing strategy depends on the data type and whether it is associated with an emergency event. a. Text data: including text and alarm data. If it does not involve an emergency, lightweight text summarization technology is used to extract the core information and only the summary is transmitted. For alarm information including fire and gas leak, Type=Alarm. Even if it is text, completeness and timeliness are guaranteed. To ensure that it has the highest priority, preprocessing is skipped and the original alarm information is transmitted directly. b. Video surveillance data: This is the largest data type; edge nodes run lightweight video analysis algorithms; if fire smoke, abnormal movement of people, or falling events are detected, key frames or short video clips containing the events are immediately extracted and transmitted, and an alarm is triggered, generating a Type=Alarm data unit; when there are no abnormalities, the frame rate and resolution of video transmission are reduced, or low-quality images are transmitted periodically. c. Audio data: If used to monitor abnormal equipment noises or people calling for help, the edge node performs audio event detection; when abnormal sound patterns, including explosions, equipment failure screams, and people calling for help, are detected, audio transmission is triggered and alarm information is generated; The data preprocessing module outputs the actual data content to be transmitted and its updated data size (PSize) attribute, based on the type and importance of the data unit and the real-time analysis results.

[0020] Furthermore, after selecting the optimal transmission link, the adaptive data packet segmentation module dynamically adjusts the data packet size PSize at the physical layer or link layer based on the real-time bit error rate (BER) and packet loss rate (PLR) evaluation results of the link. a. Adjustment logic: When the link quality is good, the selected link has a low bit error rate (BER) and packet loss rate (PLR), so the minimum PSize is used to reduce protocol overhead and increase the effective transmission rate; when the link quality deteriorates, the BER or PLR increases, so the maximum, medium, or small PSize is used as needed to increase the probability of successful transmission of a single data packet, reduce retransmission costs, and maintain transmission robustness. b. Mapping function: The mapping function maps the current link's bit error rate (BER) and packet loss rate (PLR) to a value in the set of packet sizes (PSize); ; The mapping rules are adjusted hierarchically based on preset thresholds: like and Use large bags ; like or Using medium package ; like or Use small bags ; threshold and The specific data packet size PSize is determined based on actual channel characteristics and system requirements. For Alarm data types with extremely high latency or reliability requirements, even with good link quality, the maximum PSize is used to reduce the risk of retransmission due to momentary errors. This is achieved by adding the type Type to the mapping function. To reflect this.

[0021] The link dynamic selection method for high-voltage transmission tunnel inspection data communication of the present invention achieves the following technical effects: (1) Improved the accuracy and comprehensiveness of communication link quality assessment: The multi-dimensional link quality assessment adopts three key indicators including bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr), which is different from the existing technology that only relies on a single or a few indicators (such as RSSI). This comprehensive assessment can more accurately reflect the real transmission performance of wireless channels in complex underground utility tunnel environments, especially in the presence of interference and fading, providing more reliable basic data for subsequent optimization decisions; (2) Significantly improves adaptability to complex and ever-changing wireless environments: By periodically or in real-time conducting multi-dimensional link quality assessments and dynamically selecting links and adjusting transmission parameters (such as data packet size) based on the assessment results, this invention can quickly respond to changes in the channel environment. This overcomes the limitations of traditional static configuration methods, enabling the communication system to better adapt to the dynamic fluctuations of the wireless environment in underground utility tunnels, ensuring that the current optimal link state can be utilized for transmission at different times; (3) Differentiated service guarantees for different types of monitoring and early warning information needs are achieved: This invention introduces a content-aware mechanism that can identify the type of data unit to be transmitted (such as alarms, video, audio, sensor data, etc.) and its specific QoS requirements. By setting different link quality parameter evaluation weights for different types of data (for example, increasing the weight of alarm information for low BER and low PLR), and selecting links accordingly, this invention can prioritize the low-latency and high-reliability transmission of critical information (such as fire alarms and personnel calls for help), while optimizing the transmission efficiency of other non-emergency data. This makes the communication service more in line with the actual business needs of the monitoring and early warning system; (4) Optimized wireless spectrum resource utilization efficiency: Through content-aware link selection, this invention can match different types of data streams to the frequency bands most suitable for their needs, avoiding congestion of all data on a single "optimal" frequency band, and improving the overall utilization efficiency of multi-band resources. Combined with the data preprocessing module, it reduces unnecessary data transmission at the edge (e.g., only transmitting video keyframes or text summaries), reducing the load on the wireless link from the source and further saving valuable bandwidth resources; (5) Enhanced reliability and robustness of data transmission: Adaptive adjustment of data packet size based on real-time link quality (especially BER and PLR) can balance transmission efficiency and reliability. Reducing the data packet size when the channel quality is poor can reduce the probability of single packet error or loss and reduce the cost of retransmission, thereby maintaining a certain transmission success rate and robustness in harsh environments, which is especially important for environments such as underground utility tunnels that are susceptible to interference; (6) A collaborative optimization communication strategy was constructed: This invention organically integrates several independent optimization steps, such as multi-dimensional LQA, content-aware link selection, data preprocessing, and adaptive packet splitting, into a unified framework, forming a closed-loop control system. This collaborative optimization enables complementary advantages among modules. For example, the reduced data volume from preprocessing alleviates the pressure on link selection, real-time LQA results guide the packet splitting strategy, and content requirements drive link selection and preprocessing behavior, thereby achieving a more significant overall performance improvement than existing isolated methods.

[0022] In summary, the dynamic link selection method for high-voltage transmission tunnel inspection data communication proposed in this invention proposes a multi-dimensional link quality assessment model that integrates three key indicators: physical layer, link layer, and network layer, rather than relying on a single or a few basic indicators. A content-aware mechanism is introduced to identify the type of data units to be transmitted and their differentiated quality of service requirements, and dynamically adjusts the evaluation weights of link quality parameters based on these requirements to achieve content-demand-oriented link matching. An edge data preprocessing strategy based on data type and emergency events such as fires and emergency calls is designed to reduce transmission load at the source. An adaptive data packetization strategy based on real-time link quality is implemented to dynamically optimize transmission efficiency and reliability. Multi-dimensional LQA, content-aware link selection, data preprocessing, and adaptive packetization are organically integrated to form a synergistic optimization framework, rather than isolated optimization methods, which has practical engineering value. Attached Figure Description

[0023] The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this disclosure. It is obvious that the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort. Furthermore, the same reference numerals denote the same parts throughout the drawings.

[0024] Figure 1 This is a flowchart of the link dynamic selection method for high-voltage power transmission tunnel inspection data communication according to the present invention. Detailed Implementation

[0025] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0026] Example: The link dynamic selection method for high-voltage transmission tunnel inspection data communication in this example includes the following steps: S10. Determine the overall optimization objective; S20. Establish hardware components and software modules running on the hardware components; S30. Establish the connection relationships between various components and the flow of electrical signals; S40. Perform dynamic link selection.

[0027] Furthermore, S10 includes the following: A multi-dimensional wireless communication link quality assessment model is constructed, which integrates three key indicators: bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr), to comprehensively, in real time, and accurately characterize the channel status in complex environments such as underground power transmission tunnels. Construct a content-aware adaptive link selection strategy to identify the types of data units to be transmitted, including alarms, video, and sensor data, as well as QoS constraints. Dynamically adjust the evaluation weights of link quality parameters, including bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr), based on the content type to match different data streams to the corresponding wireless frequency bands. This system integrates multiple optimization steps, including multi-dimensional link quality assessment, content-aware link selection, edge data preprocessing based on data type and emergency events, and adaptive packet size adjustment based on real-time link quality, into a collaborative comprehensive algorithm framework. This enables dynamic, efficient, and reliable optimization of monitoring and early warning communication links in complex and ever-changing environments.

[0028] Furthermore, the software modules in S20 include: a. Link Quality Assessment (LQA) module: Responsible for real-time acquisition and processing of wireless link performance data between the device and edge devices; it has the following functions: Obtain physical layer information: Obtain physical layer feedback information by calling the application programming interface of the wireless transceiver chip to assist in estimating the bit error rate (BER); Establish a probing mechanism: periodically or on demand send known probe data packets or sequences to edge devices and receive responses or reports from edge devices; by comparing the sent and received data content and packet header information, calculate the bit error rate (BER) and packet loss rate (PLR). Throughput measurement: During actual data transmission, monitor the total amount of valid data successfully sent and acknowledged within a pre-set time period; divide the total amount of valid data by the length of the time period to obtain the effective transmission rate Thr; Storage data structure: The Link Quality Assessment (LQA) module stores the bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr) measured by each edge device on each available frequency band in an internally maintained link state table; b. Content-Aware Link Selection Module: Receives data units to be transmitted and real-time link status information from the Link Quality Assessment (LQA) module, and makes link selection decisions; it has the following functions: Data Unit Attribute Acquisition: When sensor readings are packaged, video frames are generated, and alarm signals are triggered, edge devices generate or transmit corresponding data units. Each data unit or edge node is accompanied by an identified type (Type) and associated attributes. Types include Sensor data, Text, Alarm, Audio, and Video. Associated attributes include data size (PSize), latency (DelayReq), and maximum tolerable packet loss rate. Maximum tolerable bit error rate The type and associated attributes are configured on the edge device or determined through data content analysis; if a reading exceeding the threshold is received from the smoke sensor, a data unit of type Alarm is generated, with extremely low latency DelayReq and extremely high reliability requirements set; extremely low and extremely high data values ​​are preset. Query link status: Query the bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr) of all available links at the current moment from the link status table maintained by the Link Quality Assessment (LQA) module. Constraint judgment: Traverse all available links, and for each link, determine the maximum tolerable delay (DelayReq) and maximum tolerable packet loss rate for each data unit. Maximum tolerable bit error rate The data size (PSize), along with the link's bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr), are used to determine whether QoS constraints are met; if not, the data is excluded. QoS constraints include... , and ; Obtaining weights and scores: Based on the data unit type (Type), find the preset or dynamically determined weighting factors for the bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr). , , Pre-set weighting factors , , The numerical range; for Type=Alarm, and The value is higher than In this embodiment, and The value is much higher than The difference is at least an order of magnitude; for Type=Video, The values ​​are relatively high; based on the bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr) provided by the link quality assessment module (LQA), these values ​​are substituted into the preset scoring function for BER, PLR, and Thr. Calculate the score for each item; Utility calculation and selection: For each link that satisfies the constraints, use the weighted sum formula. Calculate the corresponding utility value; select the link with the largest utility value as the target link for this data unit transmission; c. Data Preprocessing Module: Before data units enter the transmission queue, the data preprocessing module preprocesses the data units according to their type and context; it has the following functions: Perform type identification and event judgment: Obtain the type (Type) of the data unit and determine whether there is an emergency event based on the system status and monitoring logic; for the corresponding video type, the edge node continuously runs the video analysis algorithm to detect smoke or flames; for the corresponding audio type, the audio analysis algorithm runs to detect people calling for help or abnormal equipment noises; If the type is Video and a fire, smoke, or unusual human activity event is detected, then keyframe extraction or short video clip extraction is performed. If the type is Audio and abnormal sound events such as people calling for help or explosions are detected, then the audio segment is captured and transmitted. If the type is Text and it is not associated with an alarm event, then text summary processing will be performed according to the configuration; Text includes logs and regular reports. If the type is Alarm, no preprocessing is performed to prevent potential loss of information, and the integrity and timeliness of the original data are guaranteed first. After preprocessing, update the PSize attribute of the data unit; d. Adaptive Data Packet Segmentation Module: After the data unit is selected for the transmission link, the adaptive data packet segmentation module determines the actual data packet size for transmission; the specific process is as follows: Query real-time link quality: Obtain the best link determined by the content-aware link selection module. Obtain the bit error rate (BER) and packet loss rate (PLR) at the current moment from the link quality assessment module (LQA). Configure mapping and adjust PSize: Based on the bit error rate (BER) and packet loss rate (PLR), determine the packet size PSize for this transmission by consulting the link status table or through a mapping function; define three PSize levels (high, medium, and low) and set thresholds for BER and PLR. When both BER and PLR are below the minimum threshold, the link quality is good, and the smallest PSize is selected; when either BER or PLR is above the maximum threshold, the link quality is poor, and the largest PSize is selected; when either is between the two, the link quality is average, and a medium or smaller PSize is selected as needed; for alarms of extremely high priority or extremely stringent reliability requirements, even if the link quality is good, configure the largest PSize to further reduce risk; pre-set the range of PSize values, and simultaneously set the maximum, minimum, medium, and small ranges for PSize.

[0029] Furthermore, in S30, each software module implements function calls and data transfer within the processing unit of the edge computing node; the specific process is as follows: S31. Edge devices periodically or event-drivenly generate raw data units and transmit the raw data units, along with information including their type, priority, and latency requirements, to the data receiving or processing interface on the edge computing node. S32. The Link Quality Assessment (LQA) module on the edge computing node continuously interacts with the edge device by sending probe packets, receiving physical layer feedback, and monitoring the actual transmitted data stream, and periodically updates its internal link status table; the link status data in the link status table is a shared resource for all software modules that need to select links; S33. When a new data unit to be transmitted arrives at the edge computing node, it first enters the data preprocessing module; the data preprocessing module decides whether to perform preprocessing and how to process it based on the data type and possible emergency event detection results, and outputs the processed data unit and the updated Size attribute; S34. The preprocessed data unit and its attributes are passed to the content-aware link selection module; the content-aware link selection module simultaneously obtains the latest link status information from the link quality assessment module (LQA), performs fitness assessment and optimal link selection algorithm based on data attributes and link quality information, and outputs the target wireless link ID or frequency band of the preprocessed data unit. S35. The selected target link ID and data unit are passed together to the adaptive data packet segmentation module; the adaptive data packet segmentation module, based on the current bit error rate (BER) and packet loss rate (PLR) of the target link, queries the link quality assessment module (LQA) again or uses the most recent LQA result to determine the data packet size (PSize) for this transmission. S36. Finally, the preprocessed data units are divided into data packets of a defined size and transmitted to the target edge device or next-hop node via the wireless communication interface of the edge node using the selected frequency band; during the transmission and reception of data packets, the Link Quality Assessment (LQA) module continues to monitor and update the link status. Furthermore, the specific process of S40 is as follows; In the wireless environment of underground utility tunnels, a closed-loop process of continuous sensing, evaluation, decision-making, and execution is established through dynamic link selection. Real-time link evaluation and content-demand-weighted intelligent decision-making are performed, and transmission performance and resource utilization are optimized by dynamically adjusting communication strategies. The Link Quality Assessment (LQA) module monitors or probes wireless links in various available frequency bands, calculates key performance indicators including Bit Error Rate (BER), Packet Loss Rate (PLR), and Effective Transmission Rate (Thr), and stores them in the link state database. The simplified steps are as follows: The arriving data units first enter the data preprocessing module; the data preprocessing module performs preprocessing according to the preprocessing rules and event detection model in the system configuration. The event detection model is set to detect fire, smoke or people calling for help, and performs type and content processing on the data units, extracts video keyframes or text summaries, and then outputs the processed data and the updated data packet size PSize. Next, the processed data unit is sent to the content-aware link selection module. The content-aware link selection module obtains real-time link quality information from the link status database and combines it with the selection weights and QoS constraints set for different data types in the system configuration. It first checks whether each available link meets the constraints of the data unit, then calculates the utility value based on content type for the links that meet the conditions, and finally selects the link with the best utility. If no suitable link can be found, the processing failure process is entered, and the data is cached or discarded. If the optimal link is selected, the link quality information along with the data unit is transmitted to the adaptive data packet segmentation module. The adaptive data packet segmentation module obtains the current bit error rate (BER) and packet loss rate (PLR) of the optimal link from the link status database, and determines the data packet size (PSize) to be used in this transmission based on the packet size threshold or rules in the system configuration. Finally, based on the determined data packet size, the data unit is segmented into data packets and wirelessly transmitted through the selected optimal link; feedback information during the transmission process is fed back to the monitoring part of the Link Quality Assessment (LQA) module to update the link status and form a closed loop; the processing flow of a single data unit is re-evaluated after successful transmission or processing failure. Among them, such as Figure 1 As shown, the steps for dynamic link selection are as follows: S41. System startup and initialization; The edge computing node starts up, initializes the Link Quality Assessment (LQA) module, and begins periodically probing or monitoring the link quality with associated edge devices on all available frequency bands, and establishes and updates the link status table; the content-aware link selection module, data preprocessing module, and adaptive data packet segmentation module enter standby state, loading configurations including data type attributes, weights, scoring functions, preprocessing strategies, and packet segmentation thresholds; S42. Data generation and reception; Edge devices continuously sense the environment or their own status, generating different types of data units, including sensor data, text, audio, and video. These data units are either initially packaged and transmitted to the edge node directly at the edge device, or collected by the edge node on demand. When emergency events, including fires and calls for help, occur, the corresponding alarm information (Type=Alarm) is generated with high priority. S43. Data preprocessing; Data units arriving at edge nodes are first processed by the data preprocessing module; depending on the data type and whether it is associated with an emergency event, corresponding compression, filtering, or key frame / key segment extraction are performed; then the processed data units are output; when smoke is detected, video preprocessing extracts key video frames. S44. Perform content-aware link selection; The preprocessed data units are sent to the content-aware link selection module. The content-aware link selection module queries the latest link status table of the link quality assessment module (LQA) to obtain the real-time multi-dimensional quality of all available links; it finds the corresponding weights and QoS constraints according to the data unit type and urgency; for links that meet the constraints, it calculates the weighted utility value and selects the link with the highest utility as the transmission path. S45. Perform adaptive data packetization; After determining the transmission link, the adaptive data packet segmentation module queries the link's real-time bit error rate (BER) and packet loss rate (PLR); based on the BER and PLR, it dynamically adjusts the data packet size (PSize) to balance efficiency and reliability. S46. Perform data transmission and link quality assessment module LQA update; After adjusting the data packet size (PSize), the data is segmented into data packets and transmitted through the selected wireless link. During transmission, the Link Quality Assessment (LQA) module continues to monitor the transmission status of the corresponding link and update the link status table. S47. Perform periodic optimization; The entire assessment, decision-making, and execution process is carried out periodically. By reassessing the priorities of links and data and adjusting strategies at fixed intervals, the process adapts to dynamic changes in the environment and business. For high-priority data, the execution decision-making process is triggered immediately.

[0030] Furthermore, the software module also includes a log recording and analysis module, which records information including each link quality assessment result, content-aware decision-making process, selected link, preprocessing behavior, and adaptive packet sub-parameters, for system performance analysis, fault diagnosis, and subsequent algorithm optimization; it also includes a visualization interface, which displays real-time link quality, transmission status of various types of data, and link selection results.

[0031] Furthermore, the Link Quality Assessment (LQA) module is responsible for real-time and comprehensive evaluation of the communication link quality of edge nodes and associated edge devices across all available wireless frequency bands. The LQA module comprehensively considers the Bit Error Rate (BER), Packet Loss Rate (PLR), and Effective Transmission Rate (Thr). Edge nodes periodically obtain estimates of BER, PLR, and Thr through probing or monitoring mechanisms. The link quality at time t is represented by a vector. a. The function expression for the bit error rate (BER) is as follows: Bit error rate (BER) is used to measure the probability of errors in physical layer bit transmission; it is estimated by monitoring the error correction performance of forward error correction (FEC) or by sending known sequences. The estimation formula is: ; in, This indicates that within a time window ending at time t, on the link Total number of bits sent. Indicates the number of bits that were erroneous at the receiving end; b. The packet loss rate (PLR) is expressed as a function of: Packet loss rate (PLR) measures the probability of packet loss, comprehensively reflecting channel errors, interference, and collisions. It is calculated by monitoring packet sequence numbers and using ACK or NACK mechanisms. The formula for estimating PLR is: ; in, This indicates that within a time window ending at time t, on the link Total number of data packets sent. Indicates the number of data packets that failed to be successfully received or acknowledged; c. The effective transmission rate Thr is expressed as a function of: The effective transmission rate (Thr) measures the number of valid data bits successfully transmitted per unit of time; it is calculated by monitoring the total amount of data successfully transmitted within a specific time interval; the formula for estimating the effective transmission rate (Thr) is: ; in, This indicates a sequence of lengths ending at time t. During the time interval, on the link The total number of valid data bits successfully transmitted. The vector output by the Link Quality Assessment (LQA) module This provides refined input for subsequent content-aware link selection.

[0032] Furthermore, the content-aware link selection module intelligently selects the optimal transmission link or frequency band based on the content attributes of the data unit to be transmitted and the real-time multi-dimensional link quality assessment results; including the following tasks: a. Data Content Awareness: Identify the type (Type) of the data unit to be transmitted; where Sensor data includes temperature, humidity, and gas concentration; Text information includes device status and system logs; Alarm information includes fire, gas leak, and intrusion alarms; Audio includes personnel calls for help and abnormal equipment sounds; Video includes on-site monitoring footage; Simultaneously, acquire key attributes of the data unit, including but not limited to: pre-processed data size (Size), maximum tolerable latency (DelayReq), and minimum reliability requirements, including the maximum tolerable packet loss rate. and maximum tolerable bit error rate The above attributes are preset or dynamically determined based on the data type and its importance and application requirements in the monitoring and early warning system. Among them, the fire alarm information Type=Alarm has low latency and high reliability requirements. b. Fitness evaluation: At time t, for each data unit to be transmitted... and every available wireless link Calculate a quantization link Transmission Data Unit A function of fitness or utility First, check the link. real-time quality Does it meet the data unit requirement? The Quality of Service (QoS) constraints include: Reliability Constraint 1: ; Reliability Constraint 2: ; Latency or rate constraints: Link The effective transmission rate must be sufficient in data units Transmit the corresponding amount of data within the maximum tolerable delay: ; If any constraint is not satisfied, the corresponding link is considered unsuitable, and the function is... The utility value is set to a pre-defined, extremely small negative value; c. Weighted utility calculation: If the link satisfies all constraints, then based on the data unit type and its attributes, it is a function. Utility value: ; in, It is for data units The non-negative weighting factor determined by the data type reflects the degree of emphasis different data types place on bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr), and satisfies... Among them, for alarm information Type=Alarm, and High weight; for video with Type=Video, High weighting; for ordinary sensor data (Type=Sensor), reliability weighting is high. This refers to link quality metrics The scoring function maps to the [0,1] interval, with higher scores for indicators that favor transmission; that is, lower BER, lower PLR, and higher Thr scores. Therefore: ; in, It is the maximum acceptable bit error rate. It is the upper limit or normalized reference value for packet loss rate. This is a reference throughput value, which is the maximum theoretical rate or average rate achieved by the system. d. Optimal Link Selection: Select the link with the highest utility value as the optimal link for this transmission. ; If none of the links meet the constraints, the corresponding data units cannot be transmitted temporarily, or data preprocessing or buffering will be required.

[0033] Furthermore, the data preprocessing module performs intelligent preprocessing on some data units at the edge node or edge device to reduce the actual amount of data transmitted and optimize bandwidth utilization; the preprocessing strategy depends on the data type and whether it is associated with an emergency event. a. Text data: including text and alarm data. If it does not involve an emergency, lightweight text summarization technology is used to extract the core information and only the summary is transmitted. For alarm information including fire and gas leak, Type=Alarm. Even if it is text, completeness and timeliness are guaranteed. To ensure that it has the highest priority, preprocessing is skipped and the original alarm information is transmitted directly. b. Video surveillance data: This is the largest data type; edge nodes run lightweight video analysis algorithms; if fire smoke, abnormal movement of people, or falling events are detected, key frames or short video clips containing the events are immediately extracted and transmitted, and an alarm is triggered, generating a Type=Alarm data unit; when there are no abnormalities, the frame rate and resolution of video transmission are reduced, or low-quality images are transmitted periodically. c. Audio data: If used to monitor abnormal equipment noises or people calling for help, the edge node performs audio event detection; when abnormal sound patterns, including explosions, equipment failure screams, and people calling for help, are detected, audio transmission is triggered and alarm information is generated; The data preprocessing module outputs the actual data content to be transmitted and its updated data size (PSize) attribute, based on the type and importance of the data unit and the real-time analysis results.

[0034] Furthermore, after selecting the optimal transmission link, the adaptive data packet segmentation module dynamically adjusts the data packet size PSize at the physical layer or link layer based on the real-time bit error rate (BER) and packet loss rate (PLR) evaluation results of the link. a. Adjustment logic: When the link quality is good, the selected link has a low bit error rate (BER) and packet loss rate (PLR), so the minimum PSize is used to reduce protocol overhead and increase the effective transmission rate; when the link quality deteriorates, the BER or PLR increases, so the maximum, medium, or small PSize is used as needed to increase the probability of successful transmission of a single data packet, reduce retransmission costs, and maintain transmission robustness. b. Mapping function: The mapping function maps the current link's bit error rate (BER) and packet loss rate (PLR) to a value in the set of packet sizes (PSize); ; The mapping rules are adjusted hierarchically based on preset thresholds: like and Use large bags ; like or Using medium package ; like or Use small bags ; threshold and The specific data packet size PSize is determined based on actual channel characteristics and system requirements. For Alarm data types with extremely high latency or reliability requirements, even with good link quality, the maximum PSize is used to reduce the risk of retransmission due to momentary errors. This is achieved by adding the type Type to the mapping function. To reflect this.

[0035] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. For those skilled in the art, all features disclosed in the present invention, or all steps in all methods or processes disclosed, except for mutually exclusive features and / or steps, can be combined in any way without departing from the principles of the present invention. The present invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. A method for dynamic link selection in data communication for high-voltage transmission pipeline inspection, characterized in that, Includes the following steps: S10. Determine the overall optimization objective; S20. Establish hardware components and software modules running on the hardware components; S30. Establish the connection relationships between various components and the flow of electrical signals; S40. Perform dynamic link selection.

2. The link dynamic selection method for high-voltage transmission tunnel inspection data communication according to claim 1, characterized in that, S10 includes the following: Construct a multi-dimensional wireless communication link quality assessment model that integrates three key indicators: bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr). Construct a content-aware adaptive link selection strategy to identify the types of data units to be transmitted, including alarms, video, and sensor data, as well as QoS constraints. Dynamically adjust the evaluation weights of link quality parameters, including bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr), based on the content type to match different data streams to the corresponding wireless frequency bands. This integrates multiple optimization steps, including multi-dimensional link quality assessment, content-aware link selection, edge data preprocessing based on data type and emergency events, and adaptive packet size adjustment based on real-time link quality, into a comprehensive algorithm framework that works collaboratively.

3. The link dynamic selection method for high-voltage transmission pipeline inspection data communication according to claim 2, characterized in that, The software modules in S20 include: a. Link Quality Assessment (LQA) module: Responsible for real-time acquisition and processing of wireless link performance data between the device and edge devices; it has the following functions: Obtain physical layer information: Obtain physical layer feedback information by calling the application programming interface of the wireless transceiver chip to assist in estimating the bit error rate (BER); Establish a probing mechanism: periodically or on demand send known probe data packets or sequences to edge devices and receive responses or reports from edge devices; by comparing the sent and received data content and packet header information, calculate the bit error rate (BER) and packet loss rate (PLR). Throughput measurement: During actual data transmission, monitor the total amount of valid data successfully sent and acknowledged within a pre-set time period; divide the total amount of valid data by the length of the time period to obtain the effective transmission rate Thr; Storage data structure: The Link Quality Assessment (LQA) module stores the bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr) measured by each edge device on each available frequency band in an internally maintained link state table; b. Content-Aware Link Selection Module: Receives data units to be transmitted and real-time link status information from the Link Quality Assessment (LQA) module, and makes link selection decisions; it has the following functions: Data Unit Attribute Acquisition: When sensor readings are packaged, video frames are generated, and alarm signals are triggered, edge devices generate or transmit corresponding data units. Each data unit or edge node is accompanied by an identified type (Type) and associated attributes. Type includes Sensor data, Text, Alarm, Audio, and Video. Associated attributes include data size (PSize), latency (DelayReq), and maximum tolerable packet loss rate. Maximum tolerable bit error rate The type and associated attributes are configured on the edge device or determined through data content analysis; if a reading exceeding the threshold is received from the smoke sensor, a data unit of type Alarm is generated, with extremely low latency DelayReq and extremely high reliability requirements set; extremely low and extremely high data values ​​are preset. Query link status: Query the bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr) of all available links in the link status table maintained by the Link Quality Assessment (LQA) module at the current moment; Constraint judgment: Traverse all available links, and for each link, determine the maximum tolerable delay (DelayReq) and maximum tolerable packet loss rate for each data unit. Maximum tolerable bit error rate The data size (PSize), along with the link's bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr), are used to determine whether QoS constraints are met; if not, the data is excluded. QoS constraints include... , and ; Obtaining weights and scores: Based on the data unit type (Type), find the preset or dynamically determined weighting factors for the bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr). , , Pre-set weighting factors , , The numerical range; for Type=Alarm, and The value is higher than For Type=Video, The values ​​are relatively high; based on the bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr) provided by the link quality assessment module (LQA), these values ​​are substituted into the preset scoring function for BER, PLR, and Thr. Calculate the score for each item; Utility calculation and selection: For each link that satisfies the constraints, use the weighted sum formula. Calculate the corresponding utility value; select the link with the largest utility value as the target link for this data unit transmission; c. Data Preprocessing Module: Before data units enter the transmission queue, the data preprocessing module preprocesses the data units according to their type and context; it has the following functions: Perform type identification and event judgment: Obtain the type (Type) of the data unit and determine whether there is an emergency event based on the system status and monitoring logic; for the corresponding video type, the edge node continuously runs the video analysis algorithm to detect smoke or flames; for the corresponding audio type, the audio analysis algorithm runs to detect people calling for help or abnormal equipment noises; If the type is Video and a fire, smoke, or unusual human activity event is detected, then keyframe extraction or short video clip extraction is performed. If the type is Audio and abnormal sound events such as people calling for help or explosions are detected, then the audio segment is captured and transmitted. If the type is Text and it is not associated with an alarm event, then text summary processing will be performed according to the configuration; Text includes logs and regular reports. If the type is Alarm, no preprocessing is performed to prevent potential loss of information, and the integrity and timeliness of the original data are guaranteed first. After preprocessing, update the PSize attribute of the data unit; d. Adaptive Data Packet Segmentation Module: After the data unit is selected for the transmission link, the adaptive data packet segmentation module determines the actual data packet size for transmission; the specific process is as follows: Query real-time link quality: Obtain the best link determined by the content-aware link selection module. Obtain the bit error rate (BER) and packet loss rate (PLR) at the current moment from the link quality assessment module (LQA). Configure mapping and adjust PSize: Based on the bit error rate (BER) and packet loss rate (PLR), determine the packet size PSize for this transmission by consulting the link status table or through a mapping function; define three PSize levels (high, medium, and low) and set thresholds for BER and PLR. When both BER and PLR are below the minimum threshold, the link quality is good, and the smallest PSize is selected; when either BER or PLR is above the maximum threshold, the link quality is poor, and the largest PSize is selected; when either is between the two, the link quality is average, and a medium or smaller PSize is selected as needed; for alarms of extremely high priority or extremely stringent reliability requirements, even if the link quality is good, configure the largest PSize to further reduce risk; pre-set the range of PSize values, and simultaneously set the maximum, minimum, medium, and small ranges for PSize.

4. The link dynamic selection method for high-voltage transmission tunnel inspection data communication according to claim 3, characterized in that, In S30, each software module implements function calls and data transfer within the processing unit of the edge computing node; the specific process is as follows: S31. Edge devices periodically or event-drivenly generate raw data units and transmit the raw data units, along with information including their type, priority, and latency requirements, to the data receiving or processing interface on the edge computing node. S32. The Link Quality Assessment (LQA) module on the edge computing node continuously interacts with the edge device by sending probe packets, receiving physical layer feedback, and monitoring the actual transmitted data stream, and periodically updates its internal link status table; the link status data in the link status table is a shared resource for all software modules that need to select links; S33. When a new data unit to be transmitted arrives at the edge computing node, it first enters the data preprocessing module; The data preprocessing module determines whether and how to perform preprocessing based on the data type and possible emergency event detection results, and outputs the processed data units and the updated Size attribute; S34. The preprocessed data unit and its attributes are passed to the content-aware link selection module; the content-aware link selection module simultaneously obtains the latest link status information from the link quality assessment module (LQA), performs fitness assessment and optimal link selection algorithm based on data attributes and link quality information, and outputs the target wireless link ID or frequency band of the preprocessed data unit. S35. The selected target link ID and data unit are passed together to the adaptive data packet module; The adaptive data packet segmentation module determines the packet size PSize for this transmission by querying the link quality assessment module (LQA) again or using the most recent LQA result based on the current bit error rate (BER) and packet loss rate (PLR) of the target link. S36. Finally, the preprocessed data unit is divided into data packets of a defined size and sent to the target edge device or next-hop node through the wireless communication interface of the edge node using the selected frequency band; during the data packet transmission and reception process, the Link Quality Assessment module (LQA) continues to monitor and update the link status.

5. The link dynamic selection method for high-voltage transmission pipeline inspection data communication according to claim 4, characterized in that, The specific process of S40 is as follows; In the wireless environment of underground utility tunnels, a closed-loop process of continuous sensing, evaluation, decision-making, and execution is established through dynamic link selection. Real-time link evaluation and content-demand-weighted intelligent decision-making are performed, and transmission performance and resource utilization are optimized by dynamically adjusting communication strategies. The Link Quality Assessment (LQA) module monitors or probes wireless links in various available frequency bands, calculates key performance indicators including Bit Error Rate (BER), Packet Loss Rate (PLR), and Effective Transmission Rate (Thr), and stores them in the link state database. The simplified steps are as follows: The arriving data units first enter the data preprocessing module; The data preprocessing module performs preprocessing based on the preprocessing rules and event detection model in the system configuration. The event detection model is set to detect fire, smoke or people calling for help. It processes the data units by type and content, extracts video keyframes or text summaries, and then outputs the processed data and the updated data packet size PSize. Next, the processed data unit is sent to the content-aware link selection module. The content-aware link selection module obtains real-time link quality information from the link status database and combines it with the selection weights and QoS constraints set for different data types in the system configuration. It first checks whether each available link meets the constraints of the data unit, then calculates the utility value based on content type for the links that meet the conditions, and finally selects the link with the best utility. If no suitable link can be found, the processing failure process is entered, and the data is cached or discarded. If the best link is selected, the link quality information, along with the data unit, is transmitted to the adaptive data packet module. The adaptive data packet segmentation module obtains the current bit error rate (BER) and packet loss rate (PLR) of the optimal link from the link state database, and determines the data packet size (PSize) to be used in this transmission based on the packet size threshold or rules in the system configuration. Finally, based on the determined data packet size, the data unit is segmented into data packets and wirelessly transmitted through the selected optimal link; feedback information during the transmission process is fed back to the monitoring part of the Link Quality Assessment (LQA) module to update the link status and form a closed loop; the processing flow of a single data unit is re-evaluated after successful transmission or processing failure. The steps for dynamic link selection are as follows: S41. System startup and initialization; The edge computing node starts up, initializes the Link Quality Assessment (LQA) module, and begins periodically probing or monitoring the link quality with associated edge devices on all available frequency bands, and establishes and updates the link status table; the content-aware link selection module, data preprocessing module, and adaptive data packet segmentation module enter standby state, loading configurations including data type attributes, weights, scoring functions, preprocessing strategies, and packet segmentation thresholds; S42. Data generation and reception; Edge devices continuously sense the environment or their own status, generating different types of data units, including sensor data, text, audio, and video. These data units are either initially packaged and transmitted to the edge node directly at the edge device, or collected by the edge node on demand. When emergency events, including fires and calls for help, occur, the corresponding alarm information (Type=Alarm) is generated with high priority. S43. Data preprocessing; Data units arriving at edge nodes are first processed by the data preprocessing module; depending on the data type and whether it is associated with an emergency event, corresponding compression, filtering, or key frame / key segment extraction are performed; then the processed data units are output; when smoke is detected, video preprocessing extracts key video frames. S44. Perform content-aware link selection; The preprocessed data units are sent to the content-aware link selection module. The content-aware link selection module queries the latest link status table of the link quality assessment module (LQA) to obtain the real-time multi-dimensional quality of all available links; it finds the corresponding weights and QoS constraints according to the data unit type and urgency; for links that meet the constraints, it calculates the weighted utility value and selects the link with the highest utility as the transmission path. S45. Perform adaptive data packetization; After determining the transmission link, the adaptive data packet segmentation module queries the link's real-time bit error rate (BER) and packet loss rate (PLR); based on the BER and PLR, it dynamically adjusts the data packet size (PSize) to balance efficiency and reliability. S46. Perform data transmission and link quality assessment module LQA update; After adjusting the data packet size (PSize), the data is segmented into data packets and transmitted through the selected wireless link. During transmission, the Link Quality Assessment (LQA) module continues to monitor the transmission status of the corresponding link and update the link status table. S47. Perform periodic optimization; The entire assessment, decision-making, and execution process is carried out periodically. By reassessing the priorities of links and data and adjusting strategies at fixed intervals, the process adapts to dynamic changes in the environment and business. For high-priority data, the execution decision-making process is triggered immediately.

6. The link dynamic selection method for high-voltage transmission tunnel inspection data communication according to claim 5, characterized in that, The software module also includes a log recording and analysis module, which records information including each link quality assessment result, content-aware decision-making process, selected link, preprocessing behavior, and adaptive packet sub-parameters, for system performance analysis, fault diagnosis, and subsequent algorithm optimization; it also includes a visualization interface, which displays real-time link quality, transmission status of various types of data, and link selection results.

7. The link dynamic selection method for high-voltage transmission tunnel inspection data communication according to claim 5, characterized in that, The Link Quality Assessment (LQA) module is responsible for real-time and comprehensive evaluation of the communication link quality of edge nodes and associated edge devices across all available wireless frequency bands. The LQA module comprehensively considers the Bit Error Rate (BER), Packet Loss Rate (PLR), and Effective Transmission Rate (Thr). Edge nodes periodically obtain estimates of BER, PLR, and Thr through probing or monitoring mechanisms. The link quality at time t is represented by a vector. a. The function expression for the bit error rate (BER) is as follows: Bit error rate (BER) is used to measure the probability of errors in physical layer bit transmission; it is estimated by monitoring the error correction performance of forward error correction (FEC) or by sending known sequences. The estimation formula is: ； in, This indicates that within a time window ending at time t, on the link Total number of bits sent. Indicates the number of bits that were erroneous at the receiving end; b. The packet loss rate (PLR) is expressed as a function of: Packet loss rate (PLR) measures the probability of packet loss, comprehensively reflecting channel errors, interference, and collisions. It is calculated by monitoring packet sequence numbers and using ACK or NACK mechanisms. The formula for estimating PLR is: ； in, This indicates that within a time window ending at time t, on the link Total number of data packets sent. Indicates the number of data packets that failed to be successfully received or acknowledged; c. The effective transmission rate Thr is expressed as a function of: The effective transmission rate (Thr) measures the number of valid data bits successfully transmitted per unit of time; it is calculated by monitoring the total amount of data successfully transmitted within a specific time interval; the formula for estimating the effective transmission rate (Thr) is: ； in, This indicates a sequence of lengths ending at time t. During the time interval, on the link The total number of valid data bits successfully transmitted. The vector output by the Link Quality Assessment (LQA) module This provides refined input for subsequent content-aware link selection.

8. The link dynamic selection method for high-voltage transmission tunnel inspection data communication according to claim 5, characterized in that, The content-aware link selection module intelligently selects the optimal transmission link or frequency band based on the content attributes of the data unit to be transmitted and the real-time multi-dimensional link quality assessment results. The work includes the following: a. Data Content Awareness: Identify the type (Type) of the data unit to be transmitted; where Sensor data includes temperature, humidity, and gas concentration; Text information includes device status and system logs; Alarm information includes fire, gas leak, and intrusion alarms; Audio includes personnel calls for help and abnormal equipment sounds; Video includes on-site monitoring footage; Simultaneously, acquire key attributes of the data unit, including but not limited to: pre-processed data size (Size), maximum tolerable latency (DelayReq), and minimum reliability requirements, including the maximum tolerable packet loss rate. and maximum tolerable bit error rate The above attributes are preset or dynamically determined based on the data type and its importance and application requirements in the monitoring and early warning system. Among them, the fire alarm information Type=Alarm has low latency and high reliability requirements. b. Fitness evaluation: At time t, for each data unit to be transmitted... and every available wireless link Calculate a quantization link Transmission Data Unit A function of fitness or utility First, check the link. real-time quality Does it meet the data unit requirement? The Quality of Service (QoS) constraints include: Reliability Constraint 1: ; Reliability Constraint 2: ; Latency or rate constraints: Link The effective transmission rate must be sufficient in data units Transmit the corresponding amount of data within the maximum tolerable delay: ； If any constraint is not satisfied, the corresponding link is considered unsuitable, and the function is... The utility value is set to a pre-defined, extremely small negative value; c. Weighted utility calculation: If the link satisfies all constraints, then based on the data unit type and its attributes, it is a function. Utility value: ； in, It is for data units The non-negative weighting factor determined by the data type reflects the degree of emphasis different data types place on bit error rate (BER), packet loss rate (PLR), and effective transmission rate (Thr), and satisfies... Among them, for alarm information Type=Alarm, and High weight; for video with Type=Video, High weighting; for ordinary sensor data Type=Sensor, reliability weighting is high. This refers to link quality metrics The scoring function maps to the [0,1] interval, with higher scores for indicators that favor transmission; that is, lower BER, lower PLR, and higher Thr all result in higher scores. Therefore: ； in, It is the maximum acceptable bit error rate. It is the upper limit of packet loss rate or a normalized reference value. This is a reference throughput value, which is the maximum theoretical rate or average rate achieved by the system. d. Optimal Link Selection: Select the link with the highest utility value as the optimal link for this transmission. ； If none of the links meet the constraints, the corresponding data units cannot be transmitted temporarily, or data preprocessing or buffering will be required.

9. The link dynamic selection method for high-voltage transmission pipeline inspection data communication according to claim 5, characterized in that, The data preprocessing module performs intelligent preprocessing on some data units at edge nodes or edge devices to reduce the actual amount of data transmitted and optimize bandwidth utilization; the preprocessing strategy depends on the data type and whether it is associated with an emergency event. a. Text data: including text and alarm data. If it does not involve an emergency, lightweight text summarization technology is used to extract the core information and only the summary is transmitted. For alarm information including fire and gas leak, Type=Alarm. Even if it is text, completeness and timeliness are guaranteed. To ensure that it has the highest priority, preprocessing is skipped and the original alarm information is transmitted directly. b. Video surveillance data: This is the largest data type. The edge nodes run a lightweight video analytics algorithm; if fire smoke, abnormal movement of people, or falling events are detected, the key frames or short video clips containing the events are immediately extracted and transmitted, and an alarm is triggered, generating a Type=Alarm data unit; when there are no abnormalities, the frame rate and resolution of the video transmission are reduced, or low-quality images are transmitted periodically. c. Audio data: If used to monitor abnormal equipment noises or people calling for help, the edge node performs audio event detection; when abnormal sound patterns, including explosions, equipment failure screams, and people calling for help, are detected, audio transmission is triggered and alarm information is generated; The data preprocessing module outputs the actual data content to be transmitted and its updated data size (PSize) attribute, based on the type and importance of the data unit and the real-time analysis results.

10. The link dynamic selection method for high-voltage transmission tunnel inspection data communication according to claim 5, characterized in that, After selecting the optimal transmission link, the adaptive data packet segmentation module dynamically adjusts the data packet size (PSize) at the physical layer or link layer based on the real-time bit error rate (BER) and packet loss rate (PLR) evaluation results of the link. a. Adjustment logic: When the link quality is good, the selected link has a low bit error rate (BER) and packet loss rate (PLR), so the minimum PSize is used to reduce protocol overhead and increase the effective transmission rate; when the link quality deteriorates, the BER or PLR increases, so the maximum, medium, or small PSize is used as needed to increase the probability of successful transmission of a single data packet, reduce retransmission costs, and maintain transmission robustness. b. Mapping function: The mapping function maps the current link's bit error rate (BER) and packet loss rate (PLR) to a value in the set of packet sizes (PSize); ； The mapping rules are adjusted hierarchically based on preset thresholds: like and Use large bags ; like or Using medium package ; like or Use small bags ; threshold and The specific data packet size PSize is determined based on actual channel characteristics and system requirements. For Alarm data types with extremely high latency or reliability requirements, even with good link quality, the maximum PSize is used to reduce the risk of retransmission due to momentary errors. This is achieved by adding the type Type to the mapping function. To reflect this.

Images