A multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network

Abstract

This invention provides a multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network, belonging to the field of construction safety and information technology. This invention achieves intelligent assessment of construction personnel identification and protective equipment wearing status by real-time collection of facial images of construction workers, information on wearing safety helmets and reflective vests, location coordinates, images of the work area, voice interaction data, and network performance parameters, combined with construction plans and specifications. The system determines the access of construction workers based on the identification results, calculates an operation efficiency index to evaluate the matching degree between on-site operation efficiency and remote collaboration capabilities, and optimizes AR guidance display, voice Q&A, and network resource allocation when efficiency is insufficient. Simultaneously, it monitors and regulates network stability to ensure data transmission reliability and remote collaboration continuity, thereby improving construction site safety, operation efficiency, and collaboration reliability.

Description

Technical Field

[0001] This invention relates to the field of construction safety and information technology, specifically to a multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network. Background Technology

[0002] Most smart safety helmets on the market only have positioning, attendance, or simple sensing functions, lacking the ability to intelligently handle complex tasks on construction sites. Existing AR helmets are mostly used in industrial manufacturing or experimental environments and are not optimized for the specific needs of construction sites. Construction sites suffer from signal blind spots, large changes in lighting, severe dust interference, and high communication latency, resulting in the inability to transmit on-site work information in a timely manner. Workers also face difficulties in obtaining specifications, drawings, or task instructions. Furthermore, existing safety helmets cannot achieve real-time personnel identification verification and protective equipment wearing monitoring, which can easily lead to safety hazards such as substitute work, unlicensed work, or improper use of protective equipment. Traditional solutions struggle to guarantee the real-time performance of multi-trade collaboration, remote guidance, and AR annotation, especially in high-density, heavily obstructed construction environments where remote collaboration significantly deteriorates. These shortcomings stem primarily from the lack of dedicated design for construction environments in existing devices, the absence of highly reliable, low-latency data transmission systems, and the fragmented nature of functional modules, hindering data fusion and intelligent management. Consequently, existing technologies cannot effectively support on-site construction management in smart construction sites, exhibiting issues such as insufficient accuracy in worker identification and protective equipment wearing, delayed on-site information acquisition and work guidance, inefficient remote collaboration, and poor equipment stability. The purpose of this invention is to provide a multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network. By integrating AR display, high-definition camera, voice interaction, 5G communication, and intelligent computing modules, it achieves intelligent inspection, voice Q&A, personnel information identification, and remote collaboration functions, improving construction efficiency and safety management levels while ensuring highly reliable, low-latency data transmission to meet the real-time management needs of smart construction sites. Summary of the Invention

[0003] The purpose of this invention is to provide a multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: A multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network includes: The data acquisition module is used to acquire facial images and images of safety helmets and reflective vests being worn, while recording the recognition and processing time; it also acquires location coordinate data, target work area image data, voice input signal data, system voice response data, as well as remote collaboration end-to-end latency, network bandwidth, network packet loss rate, and network end-to-end latency, and processes this data in conjunction with construction plan data and construction specification knowledge base data stored on the edge server. The personnel identification safety analysis module is used to calculate the personnel identification index and compare it with the personnel identification access threshold to determine whether the construction personnel meet the requirements for entering the work area. If they do not meet the requirements, they will be prohibited from entering and a rectification strategy will be prompted. The on-site operation guidance information interaction module is used to calculate the operation efficiency index and compare it with the operation efficiency target threshold to determine whether the on-site operation efficiency matches the remote collaboration capability. If they do not match, it will optimize AR guidance and voice Q&A, adjust the 5G network priority and strengthen the task reminder strategy. The network anomaly handling module is used to calculate the network stability index and compare it with the network stability threshold to determine whether the network stability at the construction site is up to standard. If it is not up to standard, it optimizes the network link, adjusts redundant transmission and cache priority, and implements a synchronous retransmission strategy.

[0005] Furthermore, the data acquisition module includes a personnel identification protection data acquisition unit and a work interaction network data acquisition unit; The personnel identification data acquisition unit is used to monitor the identification status and protective equipment wearing status of construction personnel in real time. It captures facial images through a high-definition camera installed in front of the safety helmet, and captures images of the safety helmet and reflective vest being worn through the same high-definition camera. It also records the processing time of facial recognition and protective equipment recognition through the image recognition process. The work interaction network data acquisition unit is used to monitor the work status, voice interaction effect, and network transmission performance of construction personnel in real time. It collects the current location coordinates of the workers through a Beidou positioning receiver installed on the top of the safety helmet, collects image data of the target work area of ​​the construction site through a high-definition camera installed on the front of the safety helmet, and obtains the progress information of the target work plan through the construction task plan data stored in the edge server. It also collects the voice input signal data of the construction personnel through a microphone array installed under the brim of the safety helmet, records the voice response content data of the system through the voice output channel corresponding to the bone conduction speaker, and obtains standard reference answer data through the construction specification knowledge base stored in the edge server. Finally, it collects the remote collaboration end-to-end latency, network bandwidth, network packet loss rate, and network end-to-end latency through a 5G communication antenna installed on the top of the safety helmet. Furthermore, the personnel identification security analysis module includes an identity protection parameter extraction unit, a first calculation unit, and a first analysis unit; The identity protection parameter extraction unit is used to process the collected face images. It uses face detection and feature extraction algorithms to locate the face region and extract feature vectors. The extracted face features are then matched with the standard face features in the personnel database to calculate similarity. At the same time, the recognition results are compared with the identity labeled in the database for verification. The ratio of the number of correct recognitions to the total number of recognitions is calculated to obtain the face recognition accuracy. Images of helmets and reflective vests being worn are processed. Image preprocessing techniques are used to normalize grayscale, denoise, and enhance the original images, eliminating the effects of lighting variations, dust obstruction, and noise interference. Then, a deep learning algorithm based on target detection is used to detect and locate the helmets and reflective vests in the preprocessed images, obtaining target bounding boxes and category confidence scores. The recognition results are then compared with standard wearing images at the pixel level. Accuracy is evaluated by calculating overlap rate and spatial position deviation. The detection results are verified and statistically analyzed using standard data. The ratio of the number of correct detections to the total number of detections is calculated to obtain the accuracy rate of protective equipment wearing detection. Statistical analysis was conducted on the response performance of face recognition and protective equipment recognition processes. Multiple sets of recognition processing time data were collected under different operating scenarios at the construction site, and the corresponding changes in recognition accuracy were recorded. Performance testing methods were used to fit and analyze the relationship between recognition time and recognition accuracy to determine the maximum processing time threshold under the condition that the recognition accuracy meets the preset accuracy requirements. At the same time, margin correction processing was performed in combination with edge server computing capability parameters. Subsequently, the recorded recognition time was normalized, and the current recognition processing time was proportionally mapped to the maximum allowable recognition time. The recognition time was scaled by constructing a normalization function to obtain a normalized index of recognition speed.

[0006] Furthermore, the first calculation unit is used to calculate the personnel recognition index by obtaining the face recognition accuracy rate, protective equipment wearing detection accuracy rate and recognition speed normalization index, after dimensionless processing, and using a weighted linear combination algorithm.

[0007] Furthermore, the first analysis unit is used to obtain a first evaluation result by setting a preset personnel identification access threshold and comparing the personnel identification index with the personnel identification access threshold. When the personnel identification index is greater than or equal to the personnel identification access threshold, it is determined that the construction personnel meet the requirements for entering the work area, and their identity and safety equipment status are automatically recorded and continuously monitored. When the personnel identification index is less than the personnel identification access threshold, it is determined that the construction personnel do not meet the requirements for entering the work area, and there is a risk of identity verification failure or improper wearing of protective equipment. This triggers the first warning instruction and generates the first strategy: prohibit the current construction personnel from entering the work area. The system issues a prompt, requiring the construction personnel to wear safety helmets and reflective vests correctly, and to re-perform facial recognition verification. The personnel identification index value is updated until the personnel identification index is greater than or equal to the personnel identification access threshold.

[0008] Furthermore, the on-site operation guidance information interaction module includes an operation efficiency parameter extraction unit, a second calculation unit, and a second analysis unit; The work efficiency parameter extraction unit is used to spatially match the current position coordinates of the workers with the target work area by using a position matching method, identify the construction status of the target work area image data by using an image recognition method, and perform progress comparison analysis in combination with the target work plan progress information to calculate the work progress completion rate. Speech recognition technology is used to process the voice input signal data of construction personnel into text, resulting in text question information. Semantic matching methods in natural language processing are used to calculate the semantic similarity between the system's voice response content data and the standard reference answer data. Statistical analysis is performed based on multiple question-and-answer results, and the ratio of the number of correct semantic matches to the total number of questions and answers is used as the evaluation index to obtain the voice question-and-answer response accuracy.

[0009] Furthermore, the second calculation unit is used to calculate the work efficiency index by using a weighted linear combination algorithm after dimensionless processing of the acquired work progress completion rate, voice question and answer response accuracy rate, and remote collaboration end-to-end latency.

[0010] Furthermore, the second analysis unit is used to obtain a second evaluation result by comparing the work efficiency index with a preset work efficiency achievement threshold and performing comparative analysis on the work efficiency index with the work efficiency achievement threshold. When the work efficiency index is greater than or equal to the work efficiency threshold, it indicates that the work efficiency at the construction site matches the remote collaboration capability, meets the needs of on-site operations, and is subject to continuous monitoring. When the work efficiency index is less than the work efficiency threshold, it indicates a mismatch between on-site work efficiency and remote collaboration capabilities, posing a risk of work delays or insufficient information transmission. This triggers a second warning instruction and generates a second strategy: Optimize the AR guidance display and voice Q&A response process: Increase the task prompt coverage of AR work guidance by 10-15%, and improve workers' understanding of the work process by increasing the frequency of visual annotations and voice prompts for key steps, thereby improving the accuracy of work guidance; Adjust the 5G network slicing priority: Switch the data stream of remote collaboration services to priority slices, increase video and AR annotation bandwidth resources by 15%, and ensure a reduction in end-to-end latency, thereby improving the real-time performance of remote collaboration; Enhance the task reminder and remote guidance process: Add an automatic reminder trigger mechanism to key work tasks, ensuring a 10% increase in task progress prompt frequency, and combine this with real-time remote guidance feedback to ensure that the timeliness of information transmission meets work efficiency requirements.

[0011] Furthermore, the network anomaly handling module includes a third computing unit and a third analysis unit; The third calculation unit is used to calculate the network stability index by extracting network bandwidth, network packet loss rate and network end-to-end latency, processing them in a dimensionless manner, and then using a weighted linear combination algorithm.

[0012] Furthermore, the third analysis unit is used to obtain a third evaluation result by setting a preset network stability threshold and comparing the network stability index with the network stability threshold. When the network stability index is greater than or equal to the network stability threshold, it indicates that the network stability at the construction site is qualified and should be continuously monitored. When the network stability index is less than the network stability threshold, it indicates that the network stability at the construction site is unqualified, posing a risk of data transmission interruption, information delay, or obstruction of remote collaboration. This triggers a third early warning instruction and generates a third strategy: Adjusting network links: Optimizing the connection sequence of 5G base stations on tower cranes, foundation pits, main structures, and office building rooftops at the construction site, prioritizing terminal access to the base station with the most stable signal to improve the success rate of critical task data transmission; Adjusting redundant data transmission: Enabling multi-channel transmission and retransmission mechanisms for critical inspection records and work data to reduce data loss; Adjusting the priority of cached data: Setting high-priority buffers for remote collaboration videos and real-time work guidance information to ensure timely display and transmission even under network fluctuations; Adjusting synchronization and retransmission: When the network recovers, batch data synchronization is performed according to the order of critical tasks first, then secondary tasks, ensuring the integrity of the work loop information and avoiding information disorder or omission.

[0013] Compared with the prior art, the beneficial effects of the present invention are: This invention monitors the identification status and protective equipment wearing status of construction personnel in real time, and compares the personnel identification index with the preset access threshold to automatically determine whether construction personnel meet the requirements for entering the work area. When they do not meet the requirements, an early warning is triggered in time and a prohibition on entry and rectification strategy are implemented, thereby effectively improving the safety management level of construction personnel and reducing the safety risks caused by identity verification failure or improper wearing of protective equipment.

[0014] This invention also achieves dynamic optimization of the accuracy of work process guidance, the real-time nature of remote collaboration, and the reliability of data transmission by calculating and evaluating the work efficiency index and network stability index at the construction site in real time, and combining strategies such as AR guidance display, voice Q&A, 5G network slicing optimization, and network link adjustment, thereby significantly improving the work efficiency at the construction site and the continuity and stability of information interaction. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the overall system flow of the present invention. Detailed Implementation

[0016] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0017] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0018] Example 1 Please see Figure 1 This invention provides a technical solution: a multi-functional AR smart safety helmet and construction site operation management system based on a 5G private network, comprising: The data acquisition module is used to acquire facial images and images of safety helmets and reflective vests being worn, while recording the recognition and processing time; it also acquires location coordinate data, target work area image data, voice input signal data, system voice response data, as well as remote collaboration end-to-end latency, network bandwidth, network packet loss rate, and network end-to-end latency, and processes this data in conjunction with construction plan data and construction specification knowledge base data stored on the edge server. The personnel identification safety analysis module is used to calculate the personnel identification index and compare it with the personnel identification access threshold to determine whether the construction personnel meet the requirements for entering the work area. If they do not meet the requirements, they will be prohibited from entering and a rectification strategy will be prompted. The on-site operation guidance information interaction module is used to calculate the operation efficiency index and compare it with the operation efficiency target threshold to determine whether the on-site operation efficiency matches the remote collaboration capability. If they do not match, it will optimize AR guidance and voice Q&A, adjust the 5G network priority and strengthen the task reminder strategy. The network anomaly handling module is used to calculate the network stability index and compare it with the network stability threshold to determine whether the network stability at the construction site is up to standard. If it is not up to standard, it optimizes the network link, adjusts redundant transmission and cache priority, and implements a synchronous retransmission strategy.

[0019] In this embodiment, by real-time monitoring and dynamic evaluation of construction personnel identification, protective equipment wearing status, work efficiency, and network stability, and by automatically triggering safety, work, and network optimization strategies based on preset thresholds, a closed-loop control of construction personnel access management, work process guidance, and remote collaboration optimization is achieved, thereby effectively improving the safety level, work efficiency, and information transmission reliability of the construction site.

[0020] Example 2 Please see Figure 1 In the explanation of Embodiment 1, the data acquisition module specifically includes a personnel identification protection data acquisition unit and a work interaction network data acquisition unit. The personnel identification and protection data acquisition unit is used to monitor the identification status and protective equipment wearing status of construction personnel in real time. It captures facial images (referred to as F) through a high-definition camera installed in front of the safety helmet, and captures images of the safety helmet and reflective vest being worn (referred to as S) through the same high-definition camera. It also records the processing time of facial recognition and protective equipment recognition through the image recognition process (referred to as T). The operation interaction network data acquisition unit is used to monitor the operation status, voice interaction effect, and network transmission performance of construction personnel in real time. It collects the current position coordinates of the workers (denoted as Pd) through a Beidou positioning receiver installed on the top of the safety helmet; it collects image data of the target operation area at the construction site (denoted as Zs) through a high-definition camera installed in front of the safety helmet; it obtains the target operation plan progress information (denoted as Gm) through construction task plan data stored in the edge server; it collects the voice input signal data of the construction personnel (denoted as As) through a microphone array installed under the brim of the safety helmet; it records the system's voice response content data (denoted as Rs) through the voice output channel corresponding to the bone conduction speaker; and it obtains standard reference answer data (denoted as Cb) through the construction specification knowledge base stored in the edge server. Finally, it collects the remote collaboration end-to-end delay (denoted as Ly), network bandwidth (denoted as Bw), network packet loss rate (denoted as Pw), and network end-to-end delay (denoted as Dd) through a 5G communication antenna installed on the top of the safety helmet. In this embodiment, by integrating face and protective equipment image acquisition, work location and on-site image acquisition, voice interaction and 5G network performance monitoring into the safety helmet, real-time all-round monitoring of construction personnel's identity, protection status, work progress, voice communication effect and network transmission status is achieved, thereby significantly improving the accuracy of construction site safety management and remote collaboration efficiency.

[0021] Example 3 Please see Figure 1 In the explanation of Embodiment 2, the personnel identification security analysis module specifically includes an identity protection parameter extraction unit, a first calculation unit, and a first analysis unit. The identity protection parameter extraction unit is used to process the collected face image F. It uses face detection and feature extraction algorithms to locate the face region and extract feature vectors. The extracted face features are then matched with the standard face features in the personnel database for similarity calculation. At the same time, the recognition results are compared with the identity labeled in the database for verification. The ratio of the number of correct recognitions to the total number of recognitions is used to obtain the face recognition accuracy, denoted as FL. The acquired images S of helmet and reflective vest wearing are processed. Image preprocessing techniques are used to normalize grayscale, denoise, and enhance the original images to eliminate the effects of lighting changes, dust obstruction, and noise interference. Then, a deep learning algorithm based on target detection is used to detect and locate the helmet and reflective vest targets in the preprocessed images, obtaining target bounding boxes and category confidence scores. The recognition results are compared with standard wearing images at the pixel level. The accuracy is evaluated by calculating the overlap rate and spatial position deviation. The detection results are verified and statistically analyzed using standard data. The ratio of the number of correct detections to the total number of detections is calculated to obtain the accuracy rate of protective equipment wearing detection, denoted as SG. Statistical analysis was conducted on the response performance of face recognition and protective equipment recognition processes. Multiple sets of recognition processing time data T were collected under different operating scenarios at the construction site, and the corresponding changes in recognition accuracy were recorded. Performance testing methods were used to fit and analyze the relationship between recognition time and recognition accuracy to determine the maximum processing time threshold, denoted as Tmax, under the condition that the recognition accuracy meets the preset accuracy requirements. Margin correction was also performed by combining the edge server computing power parameters. Subsequently, the recorded recognition time T was normalized, and the current recognition processing time was proportionally mapped to the maximum allowable recognition time Tmax. The recognition time was scaled by constructing a normalization function to obtain the recognition speed normalization index, denoted as TH.

[0022] In this embodiment, by performing high-precision feature extraction, deep learning target detection, and pixel-level consistency comparison on face images and safety protection equipment images, and combining the recognition processing time normalization analysis, accurate and rapid assessment of the identity verification of construction personnel and the wearing status of protective equipment can be achieved, thereby effectively improving the safety and real-time performance of personnel management at the construction site.

[0023] Example 4 Please see Figure 1 In the explanation of Embodiment 3, specifically, the first calculation unit is used to calculate the personnel recognition index, denoted as PRI, by using the obtained face recognition accuracy FL, protective equipment wearing detection accuracy SG, and recognition speed normalization index TH after dimensionless processing, and employing a weighted linear combination algorithm. The formula is as follows:

[0024] In the formula, w1, w2 and w3 represent weighting coefficients.

[0025] The face recognition accuracy (FL) represents the impact of the face recognition index on the personnel identification index and has the highest weight. Face recognition directly determines the core reliability of identity verification. When the recognition accuracy is low, it will lead to identity verification failure or misjudgment. Therefore, it is given the highest weight to reflect its dominant role in the overall personnel identification. The second highest weight is given to the accuracy rate of protective equipment wearing detection (SG) on the personnel identification index. The wearing status of protective equipment reflects the degree of compliance with work safety, and its detection accuracy directly affects safety management judgment. When the wearing detection is incorrect, there is a potential safety risk. Therefore, it is given the second highest weight to reflect its safety control value. The normalized index TH, which characterizes the recognition speed, has a medium weight on the personnel recognition index. Recognition speed affects on-site traffic efficiency. Although its direct impact on safety judgment is not as critical as accuracy, it still reflects the system's response performance. Therefore, it is given a medium weight to reflect the auxiliary role of speed in recognition efficiency. The weighted combination described above comprehensively reflects the overall level of construction personnel's identity verification accuracy, compliance with protective measures, and identification speed. A higher PRI value indicates high reliability and rapid response in personnel identification; a lower PRI value suggests inaccurate identification or improper protective equipment use.

[0026] In this embodiment, the personnel identification index PRI is calculated by weighting and combining the accuracy of face recognition, the accuracy of protective equipment wearing detection, and the normalized index of recognition speed. This enables a comprehensive assessment of the identity of construction workers and the status of their safety equipment, thereby improving the scientific nature and reliability of the identification decision.

[0027] Example 5 Please see Figure 1 In the explanation of Embodiment 4, specifically, the first analysis unit is used to compare and analyze the personnel identification index PRI with the personnel identification access threshold Pth using a preset personnel identification access threshold to obtain a first evaluation result, including: When the personnel identification index PRI is greater than or equal to the personnel identification access threshold Pth, it is determined that the construction personnel meet the requirements for entering the work area, and their identity and safety equipment status are automatically recorded and continuously monitored. When the Personnel Identification Index (PRI) is less than the Personnel Identification Access Threshold (Pth), it is determined that the construction personnel do not meet the requirements for entering the work area and there is a risk of identity verification failure or improper wearing of protective equipment. This triggers the first warning instruction and generates the first strategy: prohibiting the current construction personnel from entering the work area. The system issues a prompt, requiring the construction personnel to wear safety helmets and reflective vests correctly and to re-perform facial recognition verification. The Personnel Identification Index (PRI) value is updated until the Personnel Identification Index (PRI) is greater than or equal to the Personnel Identification Access Threshold (Pth).

[0028] Method for obtaining the personnel identification access threshold Pth: The goal of determining the personnel identification access threshold Pth is to identify a critical personnel identification index (PRI) value that can effectively distinguish between "construction personnel identity verification qualified status" and "construction personnel identity verification unqualified status". First, a database of standard construction personnel identification data is constructed. This database is formed by collecting continuous identity verification data in actual construction site operation scenarios, with a sample size of no less than 500 sets, covering various states such as accurate identity recognition, partial recognition anomalies, and recognition failure. Each sample in the database includes the face recognition accuracy rate (FL), the protective equipment wearing detection accuracy rate (SG), and the recognition speed normalization index (TH). Three or more engineers with construction safety management experience, based on on-site observation records and work assessment results, assign a "gold standard" label to the identity verification status of each sample, clearly classifying it as "identity verification qualified status" or "unqualified status". Subsequently, all sample data in the database are uniformly calculated according to the personnel identification index (PRI) calculation formula to obtain the PRI value corresponding to each sample. Next, a statistical distribution analysis of the PRI values ​​for the two states was performed, and probability density function curves were plotted. Theoretically, the PRI value for "qualified state" is generally close to or greater than 1, while the PRI value for "unqualified state" is generally lower. Based on this, the classification performance was evaluated by constructing an ROC curve, and the point with the largest Youden index was selected as the optimal split point. The PRI value corresponding to this point achieves the best balance between sensitivity and specificity, thereby effectively identifying identity verification risks. The preferred personnel identification access threshold Pth is 0.85. When PRI ≥ 0.85, the construction personnel are deemed to meet the requirements for entering the work area; when PRI < 0.85, the construction personnel are deemed to not meet the requirements for entering the work area, triggering a prohibition on entry and prompting a rectification strategy.

[0029] In this embodiment, by comparing the personnel identification index PRI with the preset access threshold Pth, real-time access control of the construction personnel's identity and protective equipment status is achieved. When the requirements are not met, an early warning is automatically triggered and entry is restricted, thereby effectively improving the safety management level of the work area.

[0030] Example 6 Please see Figure 1 In the explanation of Embodiment 2, the on-site operation guidance information interaction module specifically includes an operation efficiency parameter extraction unit, a second calculation unit, and a second analysis unit. The operation efficiency parameter extraction unit is used to spatially match the current position coordinate data Pd of the operator with the target operation area by using a position matching method, to identify the construction status of the target operation area image data Zs at the construction site by using an image recognition method, and to perform progress comparison analysis in combination with the target operation plan progress information Gm to calculate the operation progress completion rate, denoted as Wz. Speech recognition technology is used to process the voice input signal data As of the construction personnel into text, resulting in text question information. Semantic matching method in natural language processing is used to calculate the semantic similarity between the system's voice response content data Rs and the standard reference answer data Cb. Statistical analysis is performed based on multiple question-and-answer results. The ratio of the number of correct semantic matches to the total number of questions and answers is used as the evaluation index to obtain the voice question-and-answer response accuracy, denoted as Vy.

[0031] In this embodiment, by real-time monitoring and analysis of the location, work status, and voice interaction of construction personnel, the completion rate of work progress Wz and the accuracy rate of voice question and answer response Vy are calculated, thereby achieving accurate evaluation of on-site work efficiency and improving the accuracy of work guidance and operational coordination.

[0032] Example 7 Please see Figure 1 In the explanation of Embodiment Six, specifically, the second calculation unit is used to calculate the work efficiency index, denoted as AEI, by using the acquired work progress completion rate Wz, voice question-and-answer response accuracy Vy, and remote collaboration end-to-end latency Ly, after dimensionless processing, and employing a weighted linear combination algorithm. The formula is as follows:

[0033] In the formula, a1, a2, and a3 represent weighting coefficients, and Lmax represents the maximum allowable latency. The method of obtaining Lmax is as follows: the real-time interactive performance of the remote collaboration process is tested by collecting the end-to-end latency Ly of remote collaboration under different network load conditions, and evaluating it in combination with AR annotation stability and video smoothness indicators. The impact of latency on operational accuracy is analyzed by using user experience evaluation methods and latency threshold testing methods. The maximum allowable latency is determined under the condition that there is no obvious drift in remote annotation and normal audio and video synchronization. The maximum allowable latency Lmax is obtained by combining the 5G private network communication performance.

[0034] : The percentage of completion of work progress (Wz) represents the impact of work efficiency index and has the highest weight; work progress directly reflects the completion of construction tasks and is a core indicator for efficiency evaluation. When progress is lagging, the efficiency index drops significantly, so it is given the highest weight. The accuracy rate of voice question and answer response (Vy) represents the impact of the work efficiency index and has the second highest weight. Voice question and answer response reflects the efficiency of interaction between personnel and the system. A high accuracy rate can improve the quality of operation guidance, but its impact is slightly lower than that of progress completion. Therefore, it is given the second highest weight. The end-to-end latency Ly in remote collaboration represents the impact on the work efficiency index and has a medium weight. Latency affects the real-time performance of AR guidance and remote operation. Although it is not a direct indicator of work results, high latency will reduce the work efficiency experience. Therefore, it is given a medium weight to reflect its auxiliary impact on overall efficiency. A weighted linear combination can comprehensively reflect the progress of on-site operations, the quality of voice Q&A interaction, and the response performance of remote collaboration. When the AEI value increases, it indicates high operational efficiency and smooth remote collaboration; when the AEI value is low, it indicates that the operational progress is lagging or that there is a delay in remote interaction.

[0035] In this embodiment, the second calculation unit of the present invention can perform dimensionless processing on the completion rate of construction site work progress, the accuracy of voice question and answer response, and the end-to-end latency of remote collaboration, and use a weighted linear combination algorithm to calculate the work efficiency index AEI, thereby realizing a quantitative evaluation of construction site work efficiency and remote collaboration capability. This enables real-time monitoring of the progress of each work link, accurate assessment of the construction personnel's understanding of task instructions, ensuring the real-time nature of remote collaboration and reducing the risk of information lag. It unifies and quantifies work efficiency into a single indicator, making it easier for managers to quickly determine whether the on-site work efficiency meets the preset standard. At the same time, when the efficiency is lower than the threshold, it triggers optimization strategies, improves the closed-loop management level of the work process, and significantly enhances the intelligence and refinement of construction site work management.

[0036] Example 8 Please see Figure 1 In the explanation of Embodiment Seven, specifically, the second analysis unit is used to compare and analyze the work efficiency index AEI with the work efficiency threshold Ath by setting a preset work efficiency achievement threshold, denoted as Ath, to obtain a second evaluation result, including: When the work efficiency index AEI is greater than or equal to the work efficiency threshold Ath, it indicates that the work efficiency at the construction site matches the remote collaboration capability, meets the needs of on-site operations, and is continuously monitored. When the Operational Efficiency Index (AEI) is less than the Operational Efficiency Target Threshold (Ath), it indicates a mismatch between on-site operational efficiency and remote collaboration capabilities, posing a risk of operational delays or insufficient information transmission. This triggers a second warning instruction and generates a second strategy: Optimize the AR guidance display and voice Q&A response process: Increase the task prompt coverage of AR operational guidance by 10-15%, and improve workers' understanding of the operational process by increasing the frequency of visual annotations and voice prompts for key steps, thereby improving the accuracy of operational guidance; Adjust the 5G network slicing priority: Switch the data stream of remote collaboration services to priority slices, increase video and AR annotation bandwidth resources by 15%, ensuring a reduction in end-to-end latency, thereby improving the real-time performance of remote collaboration; Enhance the task reminder and remote guidance process: Add an automatic reminder trigger mechanism to key operational tasks, ensuring a 10% increase in task progress prompt frequency, and combine this with real-time remote guidance feedback to ensure that the timeliness of information transmission meets operational efficiency requirements.

[0037] Method for obtaining the work efficiency compliance threshold Ath: The goal of calibrating the work efficiency compliance threshold Ath is to determine a critical work efficiency index (AEI) value that can effectively distinguish between "work efficiency compliance status" and "work efficiency non-compliance status" on the construction site. First, a standard database of on-site work efficiency is constructed. This database is formed by continuously collecting data on construction progress completion (Wz), voice question-and-answer response accuracy (Vy), and remote collaboration end-to-end latency (Ly) during operations at multiple construction sites and across multiple work teams. The database sample size is no less than 500 sets, covering efficient operations, general operations, and states of delayed operations or insufficient information transmission. Three or more engineers with construction management experience, combining on-site work records and remote collaboration effectiveness, assign a "gold standard" label to the work status of each sample, classifying it as either "work efficiency compliance status" or "non-compliance status." Subsequently, the AEI (Advanced Effect Index) was uniformly calculated for all sample data in the database according to the formula, yielding the AEI value for each group of samples. Statistical distribution analysis was performed on the AEI values ​​for the two states, and probability density function curves were plotted. The AEI value for the "compliant state" was generally higher, while the AEI value for the "non-compliant state" was generally lower. ROC curves were used to evaluate classification performance, and the point with the largest Youden index was selected as the optimal split point. This point corresponds to an AEI value that achieves the best balance between sensitivity and specificity, thus effectively identifying operational efficiency risks. The optimal operational efficiency compliance threshold, Ath, was set at 0.80. When AEI ≥ 0.80, on-site operational efficiency was considered compliant; when AEI < 0.80, operational efficiency was considered insufficient, triggering optimization of AR guidance and voice Q&A, adjustment of 5G network priority, and enhanced task reminder strategies.

[0038] In this embodiment, the second analysis unit compares the work efficiency index AEI with the target threshold Ath in real time, which can effectively identify the mismatch between the work efficiency and remote collaboration capabilities at the construction site, detect the risk of work delays or insufficient information transmission in advance, ensure that key tasks can be completed as planned, significantly improve the execution efficiency of on-site operations and the timeliness of information transmission, improve the level of construction safety management, enhance the reliability and stability of remote collaboration, and realize the intelligent and efficient management of construction site operations.

[0039] Example 9 Please see Figure 1 In the explanation of Embodiment 2, specifically, the network anomaly handling module includes a third computing unit and a third analysis unit; The third calculation unit is used to extract network bandwidth Bw, network packet loss rate Pw, and network end-to-end delay Dd, process them without dimensions, and then use a weighted linear combination algorithm to calculate the network stability index, denoted as NSI, as shown in the following formula:

[0040] In the formula, s1, s2, and s3 represent weighting coefficients, and Bmax represents the maximum design bandwidth. Bmax is obtained by analyzing the data transmission requirements of construction site services, statistically analyzing the data volume of high-definition video streams, voice data, AR annotation data, and control commands, constructing a multi-service overlay data flow model, calculating the total bandwidth requirement under peak operating conditions using bandwidth demand modeling methods, and performing capacity planning calculations in conjunction with 5G base station configuration parameters (frequency band, bandwidth, MIMO capability) to obtain the maximum design bandwidth Bmax. Dmax represents the maximum allowable latency. Dmax is obtained by conducting performance tests on the 5G private network data transmission process, collecting the network end-to-end latency Dd under different construction site scenarios, analyzing data packet integrity and system response stability, determining the maximum latency threshold under the condition of ensuring data transmission reliability and system response continuity using network performance evaluation methods, and optimizing and correcting it in conjunction with network slice priority configuration to obtain the maximum allowable latency Dmax.

[0041] : Represents the impact of network bandwidth Bw on the network stability index, and has the highest weight; bandwidth determines data transmission capacity. When bandwidth is insufficient, it may lead to obstruction of video, AR annotation and voice transmission. Therefore, it is given the highest weight to reflect its dominant role in network carrying capacity. The packet loss rate (Pw) represents the impact of the network packet loss rate on the network stability index and has the second highest weight. A high packet loss rate will affect data integrity and the continuity of remote operations, so it is given the second highest weight to reflect the importance of network reliability. The value of Dd represents the impact of end-to-end latency on the network stability index and has a medium weight. Latency affects the real-time interactive experience, but its role is relatively auxiliary under the guarantee of bandwidth and packet loss. Therefore, it is given a medium weight to reflect its auxiliary role in the overall network performance. By using a weighted average, the bandwidth capacity, transmission reliability, and response latency of the construction site network can be comprehensively reflected. When the NSI value increases, the network performance is stable and the transmission is smooth; when the NSI is low, it indicates that the network has congestion, packet loss, or latency problems, and optimization measures need to be taken.

[0042] In this embodiment, the network stability index (NSI) is calculated in real time by the third computing unit, which can accurately reflect the network performance of the 5G private network at the construction site under high load and multiple service overlay conditions. It can detect potential problems such as insufficient bandwidth, data packet loss, or excessive end-to-end latency in advance, thereby effectively reducing the risk of interruption and information delay in the transmission of critical operation data, ensuring the stable transmission of video streams, voice interaction, and AR annotation data, realizing the continuity of remote collaboration and operation guidance, improving the reliability of information management at the construction site and the stability of network operation, and ensuring the safety and efficiency of construction operations.

[0043] Example 10 Please see Figure 1 In the explanation of Embodiment Nine, specifically, the third analysis unit is used to compare and analyze the network stability index NSI with the network stability threshold Nth using a preset network stability threshold, denoted as Nth, to obtain a third evaluation result, including: When the network stability index NSI is greater than or equal to the network stability threshold Nth, it indicates that the network stability at the construction site is qualified and should be continuously monitored. When the Network Stability Index (NSI) is less than the Network Stability Threshold (Nth), it indicates that the network stability at the construction site is unqualified, posing a risk of data transmission interruption, information delay, or obstruction of remote collaboration. This triggers a third early warning instruction and generates a third strategy: Adjusting network links: Optimizing the connection order of 5G base stations on tower cranes, foundation pits, main structures, and office building rooftops at the construction site, prioritizing terminal access to the base station with the most stable signal to improve the success rate of data transmission for critical tasks; Adjusting redundant data transmission: Enabling multi-channel transmission and retransmission mechanisms for critical inspection records and work data to reduce data loss; Adjusting the priority of cached data: Setting high-priority buffers for remote collaboration videos and real-time work guidance information to ensure timely display and transmission even under network fluctuations; Adjusting synchronization and retransmission: When the network recovers, batch data synchronization is performed according to the order of critical tasks first, then secondary tasks, ensuring the integrity of the closed-loop work information and avoiding information disorder or omission.

[0044] The method for obtaining the network stability threshold Nth: The goal of calibrating the network stability threshold Nth is to determine a critical network stability index (NSI) value that can effectively distinguish between "stable network state" and "unstable network state" at the construction site. First, a network stability standard database is constructed. This database is formed by continuously collecting network bandwidth (Bw), packet loss rate (Pw), and end-to-end latency (Dd) data under different construction site environments and network load conditions. The database has a sample size of no less than 500 groups, covering high stability, general stability, and significant network fluctuation states. Three or more engineers with experience in construction site information management, combining network monitoring records and task completion status, assign a "gold standard" label to the network state of each sample group, classifying it as either "stable network state" or "unstable state." Subsequently, all sample data in the database are uniformly calculated according to the NSI calculation formula to obtain the corresponding NSI value for each sample group. Statistical distribution analysis is performed on the NSI values ​​of the two states, and probability density function curves are plotted. The NSI value for "stable state" is generally higher, while the NSI value for "unstable state" is generally lower. The classification performance was evaluated using ROC curves, and the point with the maximum Youden index was selected as the optimal split point. This point corresponds to an NSI value that strikes the best balance between sensitivity and specificity, thus enabling effective identification of network anomalies and risks. The preferred network stability threshold Nth is 0.90. When NSI ≥ 0.90, the network stability is considered acceptable; when NSI < 0.90, the network is considered unstable, triggering optimization of network links, adjustment of redundant transmission and buffer priorities, and implementation of a synchronization retransmission strategy.

[0045] In this embodiment, the third analysis unit performs real-time comparison and analysis of the network stability index (NSI) and the preset threshold (Nth). This allows for timely determination of whether the network stability at the construction site is up to standard, early detection of potential data transmission interruptions, information delays, and risks of hindered remote collaboration, and optimization of the access sequence of key base stations through network link switching to improve the success rate of data transmission for critical tasks. Multi-channel transmission and retransmission mechanisms reduce data loss rates, ensuring the continuous transmission of inspection records, work data, and real-time guidance information even under network fluctuations. Furthermore, by adjusting cache priority and implementing batch synchronization strategies after network recovery, key work information is ensured to be transmitted sequentially and completely, achieving continuous operation and information integrity, thereby improving the efficiency of remote collaboration at the construction site, the reliability of work management, and the level of construction safety assurance.

[0046] It should be noted that all calculation formulas in this application employ regression analysis, including but not limited to machine learning algorithms, to deeply analyze the collected parameters and identify their natural trends and interrelationships. Specialized software, such as Python's Scikit-learn library or the R language, is used to automatically generate mathematical models that match the data. Then, cross-validation and other methods are used to objectively evaluate the model performance, and continuous feedback and optimization are combined to ensure that the created formulas truly reflect the inherent laws of the data, thereby guaranteeing their effectiveness and accuracy. In all calculation formulas in this application, the parameters in each formula undergo dimensionless processing within a consistent range to ensure that different physical quantities are compared on the same scale; dimensionless processing techniques include, but are not limited to, min-max-normalization and Z-score standardization. The algorithm of this invention is implemented as a Python script. Before executing the core logic, the program first executes a data loading module (e.g., using the widely used pandas library in Python) configured to read the aforementioned spreadsheet file and load its contents into the program's working memory (e.g., a DataFrame data structure). Subsequent algorithm steps will directly query and retrieve the required configuration parameters from this in-memory data structure.

[0047] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network, characterized in that: include: The data acquisition module is used to acquire facial images and images of people wearing safety helmets and reflective vests, while recording the recognition and processing time. The system collects location coordinate data, target work area image data, voice input signal data, system voice response data, as well as remote collaboration end-to-end latency, network bandwidth, network packet loss rate, and network end-to-end latency, and processes them in conjunction with construction plan data and construction specification knowledge base data stored on the edge server. The personnel identification safety analysis module is used to calculate the personnel identification index and compare it with the personnel identification access threshold to determine whether the construction personnel meet the requirements for entering the work area. If they do not meet the requirements, they will be prohibited from entering and a rectification strategy will be prompted. The on-site operation guidance information interaction module is used to calculate the operation efficiency index and compare it with the operation efficiency target threshold to determine whether the on-site operation efficiency matches the remote collaboration capability. If they do not match, it will optimize AR guidance and voice Q&A, adjust the 5G network priority and strengthen the task reminder strategy. The network anomaly handling module is used to calculate the network stability index and compare it with the network stability threshold to determine whether the network stability at the construction site is up to standard. If it is not up to standard, it optimizes the network link, adjusts redundant transmission and cache priority, and implements a synchronous retransmission strategy.

2. The multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network according to claim 1, characterized in that: The data acquisition module includes a personnel identity protection data acquisition unit and a work interaction network data acquisition unit; The personnel identification data acquisition unit is used to monitor the identification status and protective equipment wearing status of construction personnel in real time. It captures facial images through a high-definition camera installed in front of the safety helmet, and captures images of the safety helmet and reflective vest being worn through the same high-definition camera. It also records the processing time of facial recognition and protective equipment recognition through the image recognition process. The work interaction network data acquisition unit is used to monitor the work status, voice interaction effect, and network transmission performance of construction personnel in real time. It collects the current location coordinates of the workers through a Beidou positioning receiver installed on the top of the safety helmet, collects image data of the target work area of ​​the construction site through a high-definition camera installed on the front of the safety helmet, and obtains the progress information of the target work plan through the construction task plan data stored in the edge server. It also collects the voice input signal data of the construction personnel through a microphone array installed under the brim of the safety helmet, records the voice response content data of the system through the voice output channel corresponding to the bone conduction speaker, and obtains standard reference answer data through the construction specification knowledge base stored in the edge server. Finally, it collects the remote collaboration end-to-end latency, network bandwidth, network packet loss rate, and network end-to-end latency through a 5G communication antenna installed on the top of the safety helmet.

3. The multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network according to claim 2, characterized in that: The personnel identification security analysis module includes an identity protection parameter extraction unit, a first calculation unit, and a first analysis unit; The identity protection parameter extraction unit is used to process the collected face images. It uses face detection and feature extraction algorithms to locate the face region and extract feature vectors. The extracted face features are then matched with the standard face features in the personnel database to calculate similarity. At the same time, the recognition results are compared with the identity labeled in the database for verification. The ratio of the number of correct recognitions to the total number of recognitions is calculated to obtain the face recognition accuracy. The collected images of safety helmets and reflective vests are processed. Image preprocessing techniques are used to normalize grayscale, denoise, and enhance the original images to eliminate the effects of lighting changes, dust obstruction, and noise interference. Then, a deep learning algorithm based on object detection is used to detect and locate the safety helmet and reflective vest targets in the preprocessed image, and obtain the target bounding box and category confidence. The recognition results are compared with standard wearing images at the pixel level. The accuracy is evaluated by calculating the overlap rate and spatial position deviation. The detection results are verified and statistically analyzed in combination with standard data. The ratio of the number of correct detections to the total number of detections is calculated to obtain the accuracy rate of protective equipment wearing detection. Statistical analysis was conducted on the response performance of face recognition and protective equipment recognition processes. Multiple sets of recognition processing time data were collected under different work scenarios at the construction site, and the corresponding changes in recognition accuracy were recorded. Performance testing methods were used to fit and analyze the relationship between recognition time and recognition accuracy to determine the maximum processing time threshold under the condition that the recognition accuracy meets the preset accuracy requirements. At the same time, margin correction processing was carried out in combination with edge server computing capability parameters. The recorded recognition time is then normalized. The current recognition time is mapped proportionally to the maximum allowed recognition time. A normalization function is constructed to scale the recognition time and obtain a normalized index of recognition speed.

4. The multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network according to claim 3, characterized in that: The first calculation unit is used to calculate the personnel recognition index by obtaining the face recognition accuracy rate, protective equipment wearing detection accuracy rate and recognition speed normalization index, after dimensionless processing, and using a weighted linear combination algorithm.

5. The multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network according to claim 4, characterized in that: The first analysis unit is used to obtain a first evaluation result by comparing the personnel identification index with a preset personnel identification access threshold and performing comparative analysis on the personnel identification index. When the personnel identification index is greater than or equal to the personnel identification access threshold, it is determined that the construction personnel meet the requirements for entering the work area, and their identity and safety equipment status are automatically recorded and continuously monitored. When the personnel identification index is less than the personnel identification access threshold, it is determined that the construction personnel do not meet the requirements for entering the work area, and there is a risk of identity verification failure or improper wearing of protective equipment. This triggers the first warning instruction and generates the first strategy: prohibit the current construction personnel from entering the work area. The system issues a prompt, requiring the construction personnel to wear safety helmets and reflective vests correctly, and to re-perform facial recognition verification. The personnel identification index value is updated until the personnel identification index is greater than or equal to the personnel identification access threshold.

6. The multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network according to claim 2, characterized in that: The on-site operation guidance information interaction module includes an operation efficiency parameter extraction unit, a second calculation unit, and a second analysis unit. The work efficiency parameter extraction unit is used to spatially match the current position coordinates of the workers with the target work area by using a position matching method, identify the construction status of the target work area image data by using an image recognition method, and perform progress comparison analysis in combination with the target work plan progress information to calculate the work progress completion rate. Speech recognition technology is used to process the voice input signal data of construction personnel into text, resulting in text question information. Semantic matching methods in natural language processing are used to calculate the semantic similarity between the system's voice response content data and the standard reference answer data. Statistical analysis is performed based on multiple question-and-answer results, and the ratio of the number of correct semantic matches to the total number of questions and answers is used as the evaluation index to obtain the voice question-and-answer response accuracy.

7. The multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network according to claim 6, characterized in that: The second calculation unit is used to calculate the work efficiency index by using the acquired work progress completion rate, voice question and answer response accuracy rate and remote collaboration end-to-end latency, after dimensionless processing, and employing a weighted linear combination algorithm.

8. The multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network according to claim 7, characterized in that: The second analysis unit is used to obtain a second evaluation result by comparing the work efficiency index with a preset work efficiency achievement threshold and performing comparative analysis on the work efficiency index with the work efficiency achievement threshold. When the work efficiency index is greater than or equal to the work efficiency threshold, it indicates that the work efficiency at the construction site matches the remote collaboration capability, meets the needs of on-site operations, and is subject to continuous monitoring. When the work efficiency index is less than the work efficiency threshold, it indicates a mismatch between on-site work efficiency and remote collaboration capabilities, posing a risk of work delays or insufficient information transmission. This triggers a second warning instruction and generates a second strategy: Optimize the AR guidance display and voice Q&A response process: Increase the task prompt coverage of AR work guidance by 10-15%, and improve workers' understanding of the work process by increasing the frequency of visual annotations and voice prompts for key steps, thereby improving the accuracy of work guidance; Adjust the 5G network slicing priority: Switch the data stream of remote collaboration services to priority slices, increase video and AR annotation bandwidth resources by 15%, and ensure a reduction in end-to-end latency, thereby improving the real-time performance of remote collaboration; Enhance the task reminder and remote guidance process: Add an automatic reminder trigger mechanism to key work tasks, ensuring a 10% increase in task progress prompt frequency, and combine this with real-time remote guidance feedback to ensure that the timeliness of information transmission meets work efficiency requirements.

9. A multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network according to claim 2, characterized in that: The network anomaly handling module includes a third computing unit and a third analysis unit; The third calculation unit is used to calculate the network stability index by extracting network bandwidth, network packet loss rate and network end-to-end latency, processing them in a dimensionless manner, and then using a weighted linear combination algorithm.

10. A multifunctional AR smart safety helmet and construction site operation management system based on a 5G private network as described in claim 9, characterized in that: The third analysis unit is used to obtain a third evaluation result by comparing the network stability index with a preset network stability threshold, including: When the network stability index is greater than or equal to the network stability threshold, it indicates that the network stability at the construction site is qualified and should be continuously monitored. When the network stability index is less than the network stability threshold, it indicates that the network stability at the construction site is unqualified, posing a risk of data transmission interruption, information delay, or obstruction of remote collaboration. This triggers a third early warning instruction and generates a third strategy: Adjusting network links: Optimizing the connection sequence of 5G base stations on tower cranes, foundation pits, main structures, and office building rooftops at the construction site, prioritizing terminal access to the base station with the most stable signal to improve the success rate of critical task data transmission; Adjusting redundant data transmission: Enabling multi-channel transmission and retransmission mechanisms for critical inspection records and work data to reduce data loss; Adjusting the priority of cached data: Setting high-priority buffers for remote collaboration videos and real-time work guidance information to ensure timely display and transmission even under network fluctuations; Adjusting synchronization and retransmission: When the network recovers, batch data synchronization is performed according to the order of critical tasks first, then secondary tasks, ensuring the integrity of the work loop information and avoiding information disorder or omission.

Images