Intelligent education recording and broadcasting network transmission quality optimization method and system based on AI prediction

Abstract

The invention provides a smart education recording and broadcasting network transmission quality optimization method and system based on AI prediction, and relates to the technical field of smart education, and the method comprises the steps: monitoring and collecting network performance parameters of a source node, an intermediate routing node and a destination node in a recording and broadcasting transmission path in real time; performing dynamic aggregation analysis on the multi-node quality parameters, establishing a logic association relationship reflecting a global transmission condition, and outputting a comprehensive evaluation result of a network quality change trend; and inputting a quantitative evaluation result into a trained AI time sequence prediction model, predicting a network anomaly occurrence probability and type in a future time window, and obtaining a transmission strategy optimization instruction containing routing adjustment, code rate adaptation and a resource reallocation strategy. According to the method, the problems of jamming, disconnection and audio and video dislocation of remote ballast correction under a time period traffic peak are solved.

Description

Technical Field

[0001] This invention relates to the field of smart education technology, and in particular to a method and system for optimizing the transmission quality of smart education recording and broadcasting networks based on AI prediction. Background Technology

[0002] Spreading high-quality educational resources to underdeveloped areas has become a core approach to narrowing the urban-rural education gap. Among these approaches, smart education recording and broadcasting systems, as key carriers for the cross-regional transmission of high-quality course resources, have been widely used in county-level education systems. However, the actual operational performance of county-level recording and broadcasting transmission systems currently falls short of the aforementioned needs. In particular, during peak traffic periods, remote town schools frequently experience issues such as video lag, audio-visual misalignment, and interrupted transmission of recording files. This causes students in rural schools to miss core teaching segments such as mathematical theorem derivation and Chinese text analysis, requiring them to spend extra time watching the replays after class, which still cannot make up for the gap in their understanding.

[0003] The existing system employs a passive processing mode, where fault occurrence is followed by terminal feedback and server adjustment. It only sends an abnormal signal to the cloud platform after a visible fault is detected at the township school terminal, initiating remedial operations such as bitrate reduction and data retransmission. This mode introduces a time lag; for example, during a morning recorded math lesson, if a remote township school experiences video lag due to bandwidth contention on a branch link of the county's dedicated network, it takes 3-5 seconds from the occurrence of the lag to the terminal's detection and feedback, 5-8 seconds for the cloud platform to receive the signal and calculate the adjustment strategy, and 7-10 seconds for the bitrate reduction command to be issued and take effect, totaling 15-20 seconds. During this time, the teacher has already completed key teaching steps such as drawing auxiliary lines for geometric figures and deriving theorems. Township school students, having missed the core content, struggle to understand the derivation logic even after watching the replay, resulting in a knowledge gap. Summary of the Invention

[0004] The technical problem to be solved by this invention is to provide a method and system for optimizing the transmission quality of smart education recording and broadcasting networks based on AI prediction, so as to solve the problems of lag, disconnection and audio-visual misalignment in remote towns and schools during peak traffic periods.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows: Firstly, a method for optimizing the transmission quality of smart education recording and broadcasting networks based on AI prediction, the method comprising: Step 1: Monitor and collect network performance parameters of source nodes, intermediate routing nodes and destination nodes in the recording and broadcasting transmission path in real time; Step 2: Perform dynamic aggregation analysis on the quality parameters of multiple nodes, establish logical relationships reflecting the overall transmission status, and output a comprehensive evaluation result on the trend of network quality changes. Step 3: Input the quantitative evaluation results into the trained AI time series prediction model to predict the probability and type of network anomalies in the future time window, and obtain transmission strategy optimization instructions including routing adjustment, bit rate adaptation and resource reallocation strategies. Step 4: Based on the transmission strategy optimization instructions, the path selection constraints and bandwidth allocation constraints are parsed out. Through the constraints and quality parameters, the transmission route optimization calculation is performed to determine a low-latency and high-stability transmission path. Step 5: Integrate the adaptive rate control strategy and dynamic bandwidth reservation mechanism in the transmission path and transmission strategy optimization instructions to generate a set of global transmission control strategies that the system can execute. Step 6: Based on the global transmission control strategy set, network parameters are proactively pre-configured before transmission begins, and parameters are dynamically adjusted based on real-time monitored quality data during transmission to achieve closed-loop control.

[0006] Furthermore, real-time monitoring and collection of network performance parameters of source nodes, intermediate routing nodes, and destination nodes in the recording and broadcasting transmission path, including: Step 1.1: Send performance data collection commands to the source node, intermediate routing nodes, and destination node; Step 1.2: Receive network performance parameters periodically reported by each node in response to instructions. These network performance parameters include latency, jitter, packet loss rate, and available bandwidth.

[0007] Furthermore, dynamic aggregation analysis is performed on the quality parameters of multiple nodes to establish logical relationships reflecting the overall transmission status, and a comprehensive evaluation result on the trend of network quality changes is output, including: Step 2.1: Based on the quality parameters of each node, and according to predefined aggregation rules and weight configuration, perform weighted aggregation calculation on performance parameters of different types from different nodes to generate an aggregated index that characterizes the global quality status of the entire transmission path. Step 2.2: Based on the aggregation index and the spatiotemporal correlation between the parameters of each node, construct a global transmission quality map data with nodes as vertices and transmission link quality relationships as edges; Step 2.3: Based on the key performance values ​​extracted from the aggregated indicators and the link relationship weights in the global transmission quality map data, calculate the deviation of the key performance values ​​from the historical baseline data, and combine the link relationship weights to generate the real-time quality attenuation coefficient of each link. Step 2.4: Perform time series smoothing on the quality decay coefficient, and calculate the slope and fluctuation expectation of quality change in the short term based on the processed series. Combine the deviation, quality decay coefficient, slope and fluctuation expectation, and generate a quantitative comprehensive evaluation result including quality score and trend level through a weighted decision algorithm.

[0008] Furthermore, the quantitative evaluation results are input into the trained AI time-series prediction model to predict the probability and type of network anomalies within future time windows, resulting in transmission strategy optimization instructions that include routing adjustments, bitrate adaptation, and resource reallocation strategies, including: Step 3.1: Based on the quantitative comprehensive evaluation results, perform feature extraction and standardization preprocessing on the results; Step 3.2: Input the preprocessed feature data into the AI ​​time series prediction module to analyze the evolution trend of network quality parameters within a specific future time window; Step 3.3: Based on the evolution trend, calculate the probability of occurrence of various network anomalies and identify the anomaly types; Step 3.4: Based on the occurrence probability and anomaly type, generate corresponding routing adjustment strategies, bitrate adaptation schemes, and resource reallocation strategies; Step 3.5: Integrate the routing adjustment strategy, bitrate adaptation scheme, and resource reallocation strategy into a system-executable transmission strategy optimization instruction.

[0009] Furthermore, based on the transmission strategy optimization instructions, path selection constraints and bandwidth allocation constraints are parsed out. Using these constraints and quality parameters, transmission route optimization calculations are performed to determine low-latency, high-stability transmission paths, including: Step 4.1: Parse the transmission strategy optimization instructions, extract the path selection constraints and bandwidth allocation constraints contained therein, obtain the current network topology data and real-time quality parameters, and construct a multi-objective optimization function based on the constraints; Step 4.2: Based on the multi-objective optimization function, iterative calculations are performed in the set of optional transmission paths to evaluate the delay and stability indices of each path and obtain the evaluation results; Step 4.3: Based on the evaluation results, select the optimized transmission path that simultaneously meets the requirements of low latency and high stability, and obtain the path selection instruction containing detailed configuration information of the optimized transmission path.

[0010] Furthermore, the adaptive rate control strategy and dynamic bandwidth reservation mechanism in the transmission path and transmission strategy optimization instructions are fused to generate a global transmission control strategy set that the system can execute, including: Step 5.1: Based on the path selection instruction and the transmission strategy optimization instruction, parse the path selection instruction to obtain the optimized transmission path configuration information, and extract the adaptive bit rate control strategy parameters and dynamic bandwidth reservation mechanism parameters from the transmission strategy optimization instruction. Step 5.2: Establish the policy mapping relationship between path configuration information, rate control strategy and bandwidth reservation mechanism, and perform multi-policy collaborative optimization calculation; Step 5.3: Based on the collaborative optimization calculation results, generate a three-in-one control strategy that includes transmission path configuration, bit rate adaptive adjustment rules and bandwidth dynamic allocation scheme, and encapsulate the three-in-one control strategy into a global transmission control strategy set that the system can recognize and execute.

[0011] Furthermore, based on the global transmission control strategy set, network parameters are proactively pre-configured before transmission begins, and these parameters are dynamically adjusted during transmission based on real-time monitored quality data to achieve closed-loop control, including: Step 6.1: Generate the corresponding network device configuration instruction set according to the global transmission control policy set. Before the transmission starts, send the configuration instruction set to the source node, intermediate routing node and destination node to complete the active pre-configuration of network parameters. Step 6.2: During the transmission process, continuously receive real-time network performance data reported by each node and compare and analyze it with the preset quality threshold. Step 6.3: When real-time performance data is detected to deviate from the preset threshold, the dynamic adjustment mechanism is activated, the optimization parameters are recalculated, the recalculated optimization parameters are used to generate update instructions, and the instructions are sent to the corresponding network nodes for execution. Step 6.4: Based on the feedback of the execution results, verify the effect of parameter adjustment, and decide whether to start a new round of optimization and adjustment based on the verification results to form closed-loop control.

[0012] Secondly, an AI-based predictive intelligent education recording network transmission quality optimization system includes: The acquisition module is used to monitor and collect network performance parameters of source nodes, intermediate routing nodes and destination nodes in the recording and broadcasting transmission path in real time; The module is used to dynamically aggregate and analyze the quality parameters of multiple nodes, establish logical relationships that reflect the overall transmission status, and output a comprehensive evaluation result of the network quality change trend. The quantitative evaluation result is input into the trained AI time series prediction model to predict the probability and type of network anomalies in the future time window, and obtain transmission strategy optimization instructions including routing adjustment, bit rate adaptation and resource reallocation strategies. The calculation module is used to parse the path selection constraints and bandwidth allocation constraints based on the transmission strategy optimization instructions. Through the constraints and quality parameters, it performs transmission route optimization calculations to determine a low-latency, high-stability transmission path. It then integrates the transmission path with the adaptive rate control strategy and dynamic bandwidth reservation mechanism in the transmission strategy optimization instructions to generate a set of global transmission control strategies that the system can execute. The processing module is used to proactively pre-configure network parameters before transmission begins based on the global transmission control strategy set, and dynamically adjust the parameters based on real-time monitored quality data during transmission to achieve closed-loop control.

[0013] Thirdly, a computing device, comprising: One or more processors; A storage device for storing one or more programs that, when executed by one or more processors, cause the one or more processors to implement the method.

[0014] Fourthly, a computer-readable storage medium storing a program that, when executed by a processor, implements the method.

[0015] The above-described solution of the present invention has at least the following beneficial effects: By real-time, comprehensive monitoring of network performance parameters of source nodes, intermediate routing nodes, and destination nodes in the recording and broadcasting transmission path, the limitations of isolated single-point monitoring are overcome. Dynamic aggregation analysis of multi-node quality parameters accurately constructs the correlation between transmission quality across the entire domain, outputting reliable network quality assessment results. Based on a trained AI time-series prediction model, network anomalies are predicted in advance, and targeted transmission strategy optimization instructions are generated, transforming anomaly response from post-event remediation to pre-event warning, shortening anomaly handling time, and preventing students in rural schools from missing core teaching segments. Through transmission route optimization calculations and multi-strategy fusion, low-latency, high-stability transmission paths are determined, forming a global transmission control strategy set to ensure low latency, high stability, and adaptive bitrate adaptation in recording and broadcasting transmission. Combined with pre-configuration before transmission and dynamic closed-loop control during transmission, network parameters are continuously optimized to ensure that recording and broadcasting transmission quality always meets standards during peak traffic periods. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the method for optimizing the transmission quality of a smart education recording and broadcasting network based on AI prediction, provided in an embodiment of the present invention.

[0017] Figure 2 This is a schematic diagram of an AI-predictive-based smart education recording and broadcasting network transmission quality optimization system provided by an embodiment of the present invention. Detailed Implementation

[0018] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0019] like Figure 1 As shown, embodiments of the present invention propose a method for optimizing the transmission quality of smart education recording and broadcasting networks based on AI prediction. The method includes the following steps: Step 1: Monitor and collect network performance parameters of source nodes, intermediate routing nodes and destination nodes in the recording and broadcasting transmission path in real time; Step 2: Perform dynamic aggregation analysis on the quality parameters of multiple nodes, establish logical relationships reflecting the overall transmission status, and output a comprehensive evaluation result on the trend of network quality changes. Step 3: Input the quantitative evaluation results into the trained AI time series prediction model to predict the probability and type of network anomalies in the future time window, and obtain transmission strategy optimization instructions including routing adjustment, bit rate adaptation and resource reallocation strategies. Step 4: Based on the transmission strategy optimization instructions, the path selection constraints and bandwidth allocation constraints are parsed out. Through the constraints and quality parameters, the transmission route optimization calculation is performed to determine a low-latency and high-stability transmission path. Step 5: Integrate the adaptive rate control strategy and dynamic bandwidth reservation mechanism in the transmission path and transmission strategy optimization instructions to generate a set of global transmission control strategies that the system can execute. Step 6: Based on the global transmission control strategy set, network parameters are proactively pre-configured before transmission begins, and parameters are dynamically adjusted based on real-time monitored quality data during transmission to achieve closed-loop control.

[0020] In this embodiment of the invention, a six-step closed-loop process addresses the pain points of smart education recording and transmission. First, it monitors the performance parameters of the source, midstream, and target nodes across the entire domain, breaking the limitations of single-point monitoring and laying the foundation for optimization. Then, it dynamically aggregates and analyzes parameters from multiple nodes to construct a comprehensive transmission correlation, outputting accurate quality trend assessments. The core relies on an AI time-series prediction model to predict network anomalies in advance and generate optimization strategies, transforming post-event remediation into pre-event warning to prevent the loss of teaching content. Combined with pre-configuration before transmission and dynamic adjustment during the process, closed-loop control is achieved. This ensures low latency and high stability in recording and transmission, resolving issues such as buffering and misalignment in remote schools during peak hours, efficiently supporting the sharing of urban and rural educational resources, helping to narrow the education gap, and improving the teaching quality and student learning outcomes in rural schools.

[0021] In a preferred embodiment of the present invention, step 1 above may include: Step 1.1: Issue performance data collection instructions to the source node, intermediate routing nodes, and destination nodes. Specifically, based on the application scenario of the county-level integrated urban and rural compulsory education recording and broadcasting teaching system, 30 minutes before the start of each day's recorded course, the system's backend acquisition module will issue performance data collection instructions to the source node, intermediate routing node, and destination node in the recording and broadcasting transmission path according to the preset node list. The source node is the edge node of the core school in the county town, the intermediate routing nodes include the switches of the backbone link of the county education network and the access nodes of the branch links connecting various township schools, and the destination nodes include the classroom terminals of 12 township schools and the receiving nodes of the education bureau's cloud platform. When issuing instructions, the collection period will be specified. During the peak recording and broadcasting period of 8:00-9:00 in the morning and the peak playback period of 19:00-21:00 in the evening, the collection period is set to 1 second each time. During off-peak periods, the collection period is adjusted to 5 seconds each time to ensure that data can be acquired frequently during periods of high traffic to capture real-time changes.

[0022] Step 1.2: Receive network performance parameters periodically reported by each node in response to instructions. These parameters include latency, jitter, packet loss rate, and available bandwidth. Specifically, after receiving the performance data collection instruction, each node will continuously collect its own network performance parameters according to the required period and actively report them to the system backend. The source node will report the transmission latency of the encoded recording content, data transmission jitter, packet loss rate, and current available bandwidth. Intermediate routing nodes will report the forwarding latency, link jitter, packet loss rate, and remaining bandwidth when data passes through them. Among the destination nodes, the rural school classroom terminals will report the latency of receiving the recording signal, audio and video synchronization jitter, receiving end packet loss rate, and terminal available bandwidth. The education bureau cloud platform will report the latency of receiving the recording file, storage transmission jitter, file transmission packet loss rate, and cloud platform receiving bandwidth. The system backend acquisition module will receive these periodically reported data in real time and store them according to node type and parameter category.

[0023] In this embodiment of the invention, by actively sending collection instructions to the source node, intermediate routing nodes and destination node, the limitation of traditional systems that only rely on some nodes to passively report data is broken, and full-coverage monitoring of all nodes in the transmission path is realized. At the same time, core performance parameters such as receiving delay, jitter, packet loss rate and available bandwidth are clearly defined, and the timeliness of data is ensured by periodic reporting, avoiding evaluation bias caused by isolated single-point data.

[0024] In a preferred embodiment of the present invention, step 2 above may include: Step 2.1: Based on the quality parameters of each node, and using predefined aggregation rules and weight configurations, perform weighted aggregation calculations on performance parameters of different types originating from different nodes to generate an aggregation index characterizing the global quality status of the entire transmission path. Specifically, this includes: first, retrieving all network performance parameters collected from the edge source nodes of the county core school, intermediate routing nodes of the backbone and branch links of the county education network, township school terminals, and destination nodes of the education bureau cloud platform. These parameters specifically cover the latency, jitter, packet loss rate, and available bandwidth of each node. Referring to the impact of each node parameter on transmission quality during the historical operation of the county recording and broadcasting system, predefined aggregation rules and weight configurations are established. The aggregation rules specify that different types of parameters must first be uniformly converted to standardized values ​​of 0-100. The weight configuration is set according to the actual teaching transmission needs. For example, the latency parameter of intermediate routing nodes directly affects the real-time performance of recording and broadcasting, so the weight is set to 30%; the packet loss rate parameter of the destination node is related to the reception effect of the township school terminal screen, so the weight is set to 25%; the available bandwidth parameter of the source node is set to 20%; and the jitter parameter is set to 25%.

[0025] According to the aggregation rules and weight configuration, the various performance parameters of each node are standardized and weighted respectively. Then, the weighted parameter values ​​of all nodes are summed to generate an aggregated index that can characterize the global quality status of the entire recording and broadcasting transmission path from the core school source node to the township school destination node.

[0026] Step 2.2: Based on the aggregation indicators and the spatiotemporal correlation between the parameters of each node, construct a global transmission quality map with nodes as vertices and transmission link quality relationships as edges. Specifically, this includes: based on the generated aggregation indicators, and combined with the parameter data of each node at different times and between different nodes collected continuously, analyze the spatiotemporal correlation between the parameters of each node. The temporal correlation is reflected in the relationship between the change of the delay parameter of the same intermediate routing node during the peak recording and broadcasting period of 8:00-9:00 in the morning and the off-peak period. The spatial correlation is reflected in the relationship that after the delay of the intermediate node of a branch link connecting a remote town school increases, the delay of the corresponding destination node in the remote town school also increases.

[0027] After clarifying the spatiotemporal correlation, we began to construct a comprehensive transmission quality map. The edge source node of the core school in the county, the intermediate routing nodes of all backbone and branch links in the county education network, the terminals of 12 township schools, and the destination node of the education bureau cloud platform were used as vertices of the map. Each vertex was labeled with the corresponding node name and current key performance parameters. The transmission links between the nodes were used as edges of the map. Each edge was labeled with the quality level and impact range of the link based on the spatiotemporal correlation analysis results of the aggregation indicators and node parameters. For example, the backbone link connecting the core school and the education bureau cloud platform was labeled as high quality, which affected all township schools. The branch link connecting remote township schools was labeled as medium quality, which only affected the corresponding remote township schools. This clearly presents the quality correlation between each node and link in the entire recording and broadcasting transmission path.

[0028] Step 2.3: Based on the key performance values ​​extracted from the aggregated indicators and the link relationship weights in the global transmission quality map data, calculate the deviation of the key performance values ​​from the historical baseline data, and generate the real-time quality attenuation coefficient for each link by combining the link relationship weights. Specifically, this includes: first, extracting key performance values ​​from the generated aggregated indicators, including core data such as the average latency, average packet loss rate, and minimum available bandwidth of the entire transmission path. For example, extracting the current average latency as 180ms, the average packet loss rate as 2%, and the minimum available bandwidth as 18Mbps; then, extracting the link relationship weights corresponding to each link from the constructed global transmission quality map data. These weights are set according to the importance of the link in the transmission path. For example, the relationship weight of the backbone link is 40%, the weight of the branch link connecting the central town / school is 30%, and the weight of the branch link connecting the remote town / school is 30%.

[0029] Historical baseline data of the county's recording and broadcasting system under normal operating conditions was retrieved. The average latency of the historical baseline data was 150ms, the average packet loss rate was 1%, and the minimum available bandwidth was 20Mbps. By comparing the extracted current key performance values ​​with the historical baseline data, the deviation of each key performance value was calculated. For example, the deviation of the average latency was (180-150)÷150×100%=20%, the deviation of the average packet loss rate was (2-1)÷1×100%=100%, and the deviation of the minimum available bandwidth was (20-18)÷20×100%=10%. The deviation of each key performance value was multiplied by the link relationship weight of the corresponding link to generate the real-time quality attenuation coefficient of each link. For example, the average latency deviation of the branch link connecting to remote town schools was multiplied by the link relationship weight of 30% to obtain the real-time quality attenuation coefficient of 6%, which accurately reflects the attenuation of the current quality of each link compared with the normal state.

[0030] Step 2.4 involves performing time-series smoothing on the quality decay coefficient and calculating the slope and expected fluctuation of quality changes in the near future based on the processed sequence. Combining the deviation, quality decay coefficient, slope, and expected fluctuation, a weighted decision algorithm is used to generate a quantitative comprehensive evaluation result including a quality score and trend level. Specifically, this includes: performing time-series smoothing on the generated real-time quality decay coefficients of each link; selecting quality decay coefficient data collected every second over the past 5 minutes as the processing object; and calculating the average of 10 adjacent data points to eliminate the interference of instantaneous fluctuations on the data. For example, a remote town school branch link that originally fluctuated between 5% and 8%... The quality attenuation coefficient, after smoothing, stabilizes at around 6.5%, ensuring that the processed coefficient accurately reflects the link quality change trend. Based on the smoothed quality attenuation coefficient sequence, the slope of quality change in the next 5 minutes is calculated. By comparing the average attenuation coefficient of adjacent 2 minutes, if the average attenuation coefficient of the first 2 minutes is 6% and the average attenuation coefficient of the next 2 minutes is 6.5%, then the slope of quality change is 0.5% / 2 minutes, indicating that the link quality is slowly declining. At the same time, combined with the fluctuation range of the quality attenuation coefficient of the link in the same period of the past month (such as 8:00-8:05 in the morning), the expected fluctuation in the next 5 minutes is ±0.3%.

[0031] The deviation, real-time quality decay coefficient, and quality change slope and fluctuation expectation calculated in this step are combined and processed according to a predefined weighted decision algorithm, where the deviation accounts for 30%, the real-time quality decay coefficient accounts for 35%, the quality change slope accounts for 20%, and the fluctuation expectation accounts for 15%. Finally, a quantitative comprehensive evaluation result containing quality score and trend level is generated. For example, the quality score of a branch link of a remote town school is 75 points, and the trend level is slightly declining.

[0032] In this embodiment of the invention, weighted aggregation calculation is used to integrate performance parameters of different nodes and types to generate an aggregated index reflecting the global quality of the entire path, avoiding the one-sidedness of evaluating a single node or parameter; a global transmission quality map is constructed to clearly present the quality correlation between nodes and links, solving the problem of isolated analysis of node parameters; deviation and real-time quality attenuation coefficient are calculated, and the current quality status is measured in combination with historical baseline data, enhancing the accuracy of the assessment; after smoothing, the slope of change and fluctuation expectation are calculated, and quantitative results are generated through weighted decision-making, realizing accurate prediction of network quality trends and improving the comprehensiveness of network quality assessment.

[0033] In a preferred embodiment of the present invention, step 3 above may include: Step 3.1: Based on the quantitative comprehensive evaluation results, feature extraction and standardization preprocessing are performed on the results. Specifically, this includes retrieving the generated quantitative comprehensive evaluation results, which contain the quality scores and trend levels of each link in the county-level recording and broadcasting transmission path. For example, the quality score of the branch link connecting remote towns and schools is 72 points, and the trend level is slightly declining; the quality score of the backbone link is 90 points, and the trend level is stable. Feature extraction is performed on these evaluation results, focusing on extracting features directly related to the recording and broadcasting transmission quality of towns and schools. These features include the magnitude of quality score changes for each link, the duration of the trend level, and the deviation features of the current value of key performance parameters from the historical baseline. For example, features such as the quality score of the remote town and school branch link decreasing by 8 points and the latency increasing by 30ms compared to the baseline within the past 5 minutes are extracted. The extracted features are then standardized preprocessed, mapping the feature values ​​of different dimensions to a uniform range of 0-1. For example, the latency deviation is converted from an increase of 30ms to a standardized value of the corresponding proportion to avoid the impact of differences in feature units or value ranges on the accuracy of subsequent predictive analysis. This ensures that the preprocessed feature data can accurately reflect the actual quality status of the county-level recording and broadcasting network.

[0034] Step 3.2 involves inputting the preprocessed feature data into the AI ​​time series prediction module to analyze the evolution trend of network quality parameters within a specific future time window. This includes collecting historical data from county-level smart education recording scenarios. Data sources include network performance parameters of source nodes, intermediate routing nodes, and destination nodes in the county-level recording transmission path over the past 6 months, as well as recording course schedules, student playback data, and network equipment operation logs for the corresponding time periods. Data from key time periods such as the morning peak recording period of 8:00-9:00 AM and the evening peak playback period of 7:00-9:00 PM are highlighted to ensure the data comprehensively reflects the actual operating characteristics of the county-level recording network. The collected historical data is cleaned to remove abnormal and invalid data caused by equipment failure or temporary power outages, and a small amount of data missing due to temporary node offline events is supplemented. Data continuity is ensured by using the mean of adjacent time periods for padding. The cleaned data is then divided into training and testing datasets in a 7:3 ratio.

[0035] A suitable basic model architecture for time-series data prediction was selected and customized for county-level recording and broadcasting scenarios. Addressing the significant time-variety fluctuations in county-level recording and broadcasting traffic, a time-period feature embedding layer was added to the model, enabling it to identify traffic patterns during different time periods (e.g., peak recording times, peak playback times, and off-peak times). Simultaneously, considering the sensitivity of township and school terminals to latency and bandwidth, the weighting of latency and available bandwidth parameters in the model's loss function was increased, making the model training process more closely aligned with actual teaching transmission needs. The adjusted model was trained using a pre-defined training dataset, with real-time monitoring of the model's prediction accuracy on the test dataset. The model's prediction errors for link latency and available bandwidth within the next two minutes were controlled to be within 10%. When the speed is within ms and 2Mbps, training is stopped and the current model parameters are saved to form a preliminary AI time series prediction model. The preliminary model is deployed to the test environment of the county-level recording and broadcasting system to simulate different scenarios such as morning recording and broadcasting peaks and evening playback peaks. Real-time collected preprocessed feature data is input, and the model's analysis results on the evolution trend of network quality parameters are observed. The model is iteratively optimized by comparing it with the actual network changes. For example, when it is found that the model has a large deviation in predicting the bandwidth changes of remote town and school branch links, historical data of such links are added to retrain the model until the model's prediction accuracy is stable at over 90% for 30 consecutive days in the test environment. Finally, the construction of the AI ​​time series prediction analysis module adapted to the county-level smart education recording and broadcasting scenario is completed.

[0036] The preprocessed feature data is imported into the AI ​​time-series prediction and analysis module. This module is preloaded with analysis logic optimized for county-level smart education recording scenarios. It can perform trend analysis based on the time-specific characteristics of county-level recording traffic and set specific future time windows. Based on the actual needs of county-level recording teaching, the time window is set to the next 2 minutes. This duration can provide early warning of anomalies that may affect teaching and avoid the results becoming invalid due to excessively long prediction periods. Based on the preprocessed feature data and combined with historical recording network quality change data for the same period, the AI ​​time-series prediction and analysis module analyzes the evolution trend of network quality parameters of each link within the next 2 minutes. For example, the analysis shows that the latency of the branch link connecting to remote town schools will continue to increase by 20ms and the available bandwidth will decrease by 5Mbps within the next 2 minutes, while the backbone link quality will remain stable. This clarifies the direction and magnitude of quality changes of different links in a short period of time.

[0037] Step 3.3: Based on the evolution trend, calculate the probability of occurrence of various network anomalies and identify the anomaly types. Specifically, this includes: establishing a correspondence between trends and anomaly types based on the evolution trend of network quality parameters obtained from the analysis, combined with common network anomaly types in county-level recording and broadcasting systems. For example, when the available bandwidth of a link continues to decrease and the latency increases rapidly, it corresponds to the link bandwidth preemption anomaly type. Then, based on this correspondence, calculate the probability of occurrence of various network anomalies, referring to the frequency of anomalies under similar trends in historical data. For example, if the link bandwidth preemption anomaly occurred 20 times under similar trends in the past month, and the current trend matches this type of situation 90%, then calculate the probability of occurrence of the link bandwidth preemption anomaly in the current scenario. At the same time, further identify the anomaly type based on trend characteristics. For example, for the trend of increased latency and reduced bandwidth of remote town and school branch links, combined with the actual situation that the county-level private network branch links are prone to bandwidth preemption due to the transmission of resources from surrounding schools, it is clearly identified that the anomaly type that will occur on this link is branch link bandwidth preemption, and its probability of occurrence is as high as 85%.

[0038] Step 3.4: Based on the occurrence probability and anomaly type, generate corresponding routing adjustment strategies, bitrate adaptation schemes, and resource redistribution strategies. Specifically, this includes: Developing specific strategies for county-level recording scenarios based on the calculated anomaly occurrence probability and identified anomaly type. If the anomaly type is bandwidth preemption on a remote town / school branch link and the occurrence probability reaches 85%, a corresponding routing adjustment strategy is generated to plan a new path from the core school source node through a backup branch link to the remote town / school destination node, avoiding the impact of the original congested link on transmission; generating a bitrate adaptation scheme, considering that the remote town / school terminal needs to maintain 1080P image clarity, appropriately reducing the recording bitrate from the current 6Mbps to 5Mbps, ensuring image quality while reducing bandwidth consumption; generating a resource redistribution strategy, allocating an additional 5Mbps of bandwidth from the county-level education network backbone link to the backup branch link of the remote town / school, ensuring the stability of the new path transmission; if the anomaly occurrence probability is low, such as below 30%, only a mild adjustment strategy is generated, such as slightly optimizing the bitrate parameters, to avoid excessive adjustment affecting normal transmission, ensuring that the generated strategy matches the severity of the anomaly.

[0039] Step 3.5 integrates the routing adjustment strategy, bitrate adaptation scheme, and resource reallocation strategy into a system-executable transmission strategy optimization instruction. Specifically, this includes: collecting the generated routing adjustment strategy, bitrate adaptation scheme, and resource reallocation strategy; clarifying the execution targets, execution parameters, and execution timing of each strategy; for example, the routing adjustment strategy targets the core campus edge nodes and intermediate routing nodes, the execution parameter is the node connection order of the new path, and the execution timing is within the next minute; the bitrate adaptation scheme targets the core campus recording and broadcasting encoding equipment, the execution parameter is the adjusted bitrate value, and the execution timing is synchronized with the routing switch; the resource reallocation strategy targets the county-level private network bandwidth management nodes, the execution parameter is the additional allocated bandwidth value, and the execution timing is 5 seconds before the routing switch. Then, according to the instruction format requirements of the county-level recording and broadcasting system, these strategies are integrated into a system-executable transmission strategy optimization instruction. The instruction clearly marks the priority of each strategy to ensure that each node can execute the instructions accurately and in the correct order after receiving them, avoiding transmission interruptions due to chaotic strategy execution, and ultimately achieving early intervention for anomalies in the county-level recording and broadcasting network.

[0040] In this embodiment of the invention, by performing feature extraction and standardized preprocessing on the quantitative comprehensive evaluation results, key quality features can be screened out and data difference interference can be eliminated. Inputting the preprocessed data into the AI ​​time-series prediction module allows for early assessment of network quality evolution trends within a specific future time window, breaking the traditional post-fault response model and achieving early warning of anomalies. Calculating the probability of anomalies based on trends and identifying their types enables precise location of problem attributes, avoiding blind strategy formulation. Generating three specific strategies—routing, bitrate, and resource allocation—for different anomaly types allows for the formation of multi-dimensional response solutions. Finally, these are integrated into executable system instructions, ensuring rapid strategy implementation and resolving issues of buffering and disconnections in remote towns and schools during peak recording and broadcasting periods in county-level areas, thus guaranteeing the stability of teaching transmission. In a preferred embodiment of the present invention, step 4 above may include: Step 4.1: Parse the transmission strategy optimization instructions, extract the path selection constraints and bandwidth allocation constraints contained therein, obtain the current network topology data and real-time quality parameters, and construct a multi-objective optimization function based on the constraints. Specifically, this includes: parsing the generated system-executable transmission strategy optimization instructions. The strategy optimization instructions are formulated for county-level smart education recording and broadcasting scenarios, and it is necessary to extract the constraints related to path selection and bandwidth allocation. The path selection constraints are set according to the recording and broadcasting transmission needs of township schools. For example, it is specified that the signal transmission delay of recording and broadcasting in remote township schools should be ≤200ms and the packet loss rate during transmission should be ≤1%, while the delay in central township schools should be ≤180ms. The bandwidth allocation constraints are determined based on the bandwidth scale of each township school. For example, the transmission bandwidth of central township schools should be ≥50Mbps and the transmission bandwidth of remote township schools should be ≥20Mbps.

[0041] The network topology data is obtained through the network management of the county-level education network. This data includes the connection relationships of the edge source nodes of the core school in the county seat, intermediate routing nodes of all backbone and branch links within the county-level network, terminals of 12 township schools, and destination nodes of the education bureau's cloud platform, as well as the current activation status of the links between each node. Simultaneously, real-time quality parameters such as link latency, jitter, packet loss rate, and available bandwidth are retrieved. The extracted path selection constraints and bandwidth allocation constraints are combined with the current network topology data and real-time quality parameters to construct a multi-objective optimization function. ,in, For path The delayed optimization subfunction is expressed as follows: , This represents the real-time transmission delay throughout path X. This indicates the maximum permissible delay set according to the constraints. For path The stability optimization subfunction is expressed as follows: ,in, Representing a path The real-time packet loss rate throughout the entire process, This represents the maximum allowable packet loss rate set by the constraint conditions. Representing a path The entire process of real-time shaking, This indicates the maximum permissible jitter set based on historical stable transmission data; and These are the weight coefficients for the delayed optimization subfunction and the stability optimization subfunction, respectively. This represents any one of the county-level recording and broadcasting transmission paths to be evaluated.

[0042] Step 4.2: Based on the multi-objective optimization function, iterative calculations are performed on the set of optional transmission paths to evaluate the latency and stability indicators of each path and obtain the evaluation results. Specifically, this includes: based on the network topology data of the county-level education network, identifying all optional transmission paths from the edge source node of the core school in the county seat to the destination nodes of each township school; using the constructed multi-objective optimization function as the evaluation standard, iterative calculations are performed on each optional transmission path in the set; for each path, its current real-time quality parameters such as latency, packet loss rate, and available bandwidth are substituted, and the path's performance in terms of latency and stability is calculated through the function. The system uses numerical values, with the latency index directly using the transmission latency data of the entire path, and the stability index calculated by comprehensively considering the packet loss rate fluctuations and bandwidth stability of each link in the path. For example, when calculating the path corresponding to branch link A, its latency is found to be 190ms and its stability index score is 92; when calculating the path corresponding to branch link B, its latency is found to be 210ms and its stability index score is 88; when calculating the path corresponding to branch link C, its latency is found to be 185ms and its stability index score is 94. Through such iterative calculations, the latency and stability index evaluation results of all optional paths are obtained.

[0043] Step 4.3: Based on the evaluation results, select the optimized transmission path that simultaneously meets the requirements of low latency and high stability. Obtain the path selection instruction containing detailed configuration information of the optimized transmission path. Specifically, this includes: clarifying the core requirement standard for county-level smart education recording and broadcasting transmission, namely, the transmission path of all township school destination nodes must simultaneously meet the requirements of low latency and high stability. The low latency requirement is latency ≤200ms, and the high stability requirement is a stability index score ≥90 points. Compare the evaluation results of each optional path obtained in the previous step with this requirement standard, and select the path that meets both requirements. For example, for three optional paths of a remote township school, the path corresponding to branch link A has a latency of 190ms and a stability score of 92 points, and the path corresponding to branch link C has a latency of 185ms and a stability score of 94 points, both of which meet the requirement standard. However, the path corresponding to branch link B has a latency of 210ms and a stability score of 88 points, which does not meet the requirement standard. Therefore, the first two paths are selected as candidate optimized transmission paths.

[0044] Among the candidate optimized transmission paths, their latency and stability metrics are further compared, and the path with better overall performance is selected as the final optimized transmission path. For example, the path corresponding to branch link C has lower latency and higher stability score, so this path is determined as the optimized transmission path for this remote town school. A path selection instruction containing detailed configuration information of the optimized transmission path is generated. The instruction clearly marks the names of all nodes involved in the path, the connection order of links between nodes, the bandwidth configuration parameters of each link, the data transmission protocol type, and other information to ensure that the core school edge nodes, intermediate routing nodes, and township school destination nodes can accurately complete the path configuration according to the instruction, ensuring stable transmission of the recording signal.

[0045] In this embodiment of the invention, by parsing the transmission strategy optimization instructions to extract path selection and bandwidth allocation constraints, and combining the current network topology and real-time quality parameters to construct a multi-objective optimization function, it is possible to ensure that the function design closely revolves around the core needs of recording and broadcasting transmission in rural schools, avoiding ineffective calculations detached from the actual scenario. Based on this function, latency and stability indicators are iteratively calculated and evaluated among the available paths, allowing for a comprehensive comparison of the performance differences of different paths, breaking the limitations of traditional static routing, and accurately identifying potential paths that meet the requirements. Finally, based on the evaluation results, a path that simultaneously meets low latency and high stability is selected and configuration instructions are generated. This effectively solves the problems of lag and disconnection caused by path congestion during peak traffic periods in remote rural schools, ensuring that the transmission of recording and broadcasting signals meets the requirements of real-time viewing of 1080P images and latency ≤200ms, providing reliable path support for the efficient transmission of urban and rural educational resources.

[0046] In a preferred embodiment of the present invention, step 5 above may include: Step 5.1: Based on the path selection instruction and transmission strategy optimization instruction, parse the path selection instruction to obtain optimized transmission path configuration information. Simultaneously, extract adaptive bitrate control strategy parameters and dynamic bandwidth reservation mechanism parameters from the transmission strategy optimization instruction. Specifically, this includes: obtaining the generated path selection instruction and the transmission strategy optimization instruction generated in Step 3.5. Both instructions are designed for the transmission needs of county-level smart education recording scenarios. Parse the optimized transmission path configuration information from the path selection instruction, specifically including the identifiers of the edge source nodes of the county's core schools, the connection order of intermediate routing nodes, the identifiers of the destination nodes of township schools, and the transmission protocol types and current activation status of the links between each node. Simultaneously, extract the adaptive bitrate control strategy parameters from the transmission strategy optimization instruction. These parameters are set according to the type of recording content and the terminal requirements of township schools, including the base bitrate value for Chinese language teaching sessions, the upper limit bitrate value for math and geometry demonstration sessions, and the trigger threshold for bitrate adjustment. Extract the dynamic bandwidth reservation mechanism parameters, including the bandwidth reserved for each branch link during the morning recording peak period, the bandwidth adjustment cycle during the evening playback peak period, and the priority allocation rules when bandwidth reservation is insufficient.

[0047] Step 5.2 establishes a policy mapping relationship between path configuration information, bitrate control strategy, and bandwidth reservation mechanism, and performs multi-strategy collaborative optimization calculations. Specifically, this includes: based on the actual needs of county-level recording and broadcasting transmission, establishing a policy mapping relationship between the acquired path configuration information, bitrate control strategy, and bandwidth reservation mechanism. For example, for an optimized transmission path in a remote town school, if the current available bandwidth of the backup branch link in this path is 25Mbps, it is associated with the bitrate upper limit parameter in the adaptive bitrate control strategy to determine that the maximum recording and broadcasting bitrate under this path can be set to 5.5Mbps; at the same time, the bandwidth requirement of this path is associated with the reserved quota in the dynamic bandwidth reservation mechanism. The system specifies that when the available bandwidth for a path is less than 20Mbps, an additional bandwidth reservation process is triggered, allocating 3Mbps of bandwidth from the backbone link to supplement the branch link. Subsequently, multi-strategy collaborative optimization calculations are carried out. During the calculation process, if it is found that the bitrate cap for a certain path is set too high, causing the reserved bandwidth to be insufficient to meet the demand, the bitrate cap is appropriately lowered to 5.5Mbps to match the bitrate demand with the bandwidth reservation. If the bandwidth reservation amount for a certain path exceeds the actual demand, the excess 5Mbps of reserved bandwidth is allocated to other bandwidth-constrained remote town and school paths to ensure that the bandwidth resources of the county education network are used efficiently, while ensuring that the recording and broadcasting transmission quality of each path meets the requirements.

[0048] Step 5.3: Based on the collaborative optimization calculation results, a three-in-one control strategy is generated, including transmission path configuration, bitrate adaptive adjustment rules, and bandwidth dynamic allocation scheme. This three-in-one control strategy is encapsulated into a global transmission control strategy set that the system can recognize and execute. Specifically, this includes: generating a three-in-one control strategy for recording and broadcasting transmission in all township schools within the county, based on the multi-strategy collaborative optimization calculation results; specifying the optimized path node sequence, link transmission protocol, and activation time for each township school, such as completing path switching before 7:50 AM to ensure timely transmission of the 8:00 AM recording session; and specifying the bitrate adaptive adjustment rules, detailing the bitrate baseline value and adjustment logic for different township schools and different teaching periods. For example, the bitrate baseline value for a remote township school's Chinese language recording session is 3Mbps, and when the latency rises to 190ms... The bandwidth is initially set at 2.8Mbps, and will recover to 3Mbps when the latency drops to 170ms. The base bitrate for math recording lessons is 5.5Mbps, which can be increased to 6Mbps when bandwidth is sufficient. The dynamic bandwidth allocation scheme clearly defines the bandwidth reservation quotas and allocation rules for each time period and path. For example, during the peak recording period from 8:00 to 9:00 AM, 5Mbps bandwidth is reserved for each path in remote towns and schools. When the bandwidth utilization of a path exceeds 90%, 2-3Mbps bandwidth is allocated from idle paths. This three-in-one control strategy is encapsulated according to the instruction format requirements of the county-level smart education recording system, generating a global transmission control strategy set that the system can recognize and execute. The strategy set marks the execution node, execution priority, and exception handling plan for each strategy, ensuring that each node in the system can accurately receive and execute the strategy, guaranteeing stable and reliable recording transmission throughout the county.

[0049] In this embodiment of the invention, by parsing the path selection instruction and the transmission strategy optimization instruction, the optimized path configuration, bitrate control parameters, and bandwidth reservation parameters are accurately obtained, ensuring that the basic data of each strategy closely matches the needs of the county-level recording and broadcasting scenario. By establishing the strategy mapping relationship among the three and carrying out collaborative optimization calculations, resource waste or transmission stuttering caused by mismatch between path, bitrate, and bandwidth can be avoided. For example, the bitrate upper limit can be dynamically adjusted and the reserved bandwidth quota can be matched according to the real-time bandwidth of the optimized path. Finally, a three-in-one control strategy is generated and encapsulated into a global transmission control strategy set, which enables all links such as core schools, intermediate nodes, and township school terminals to uniformly execute collaborative strategies, effectively ensuring that remote township schools can stably receive 1080P recording and broadcasting signals even during peak traffic periods, solving the problems of screen stuttering and audio-visual misalignment, and providing integrated strategy support for the efficient sharing of urban and rural education resources.

[0050] In a preferred embodiment of the present invention, step 6 above may include: Step 6.1: Based on the global transmission control policy set, generate the corresponding network device configuration instruction set. Before transmission begins, distribute the configuration instruction set to the source node, intermediate routing nodes, and destination node to complete the proactive pre-configuration of network parameters. Specifically, this includes: based on the generated global transmission control policy set and combined with the model and configuration requirements of each network device in the county-level smart education recording and broadcasting system, generating the corresponding network device configuration instruction set. For the source node of the core school in the county, the configuration instruction includes recording and broadcasting encoding parameters, data output port number, and connection protocol with intermediate routing nodes; for intermediate routing nodes, the configuration instruction includes node forwarding rules for optimizing the transmission path and bandwidth reservation. The configuration includes data receiving port number, decoding parameters, and cache threshold for the destination node of township schools and the education bureau's cloud platform. Then, according to the schedule of the county-wide recorded courses, the configuration instruction set is distributed to the source node, intermediate routing node, and destination node through the management channel of the county-wide education network 10 minutes before the start of the morning recorded course and 15 minutes before the start of the evening playback service. After receiving the instruction, each node automatically executes the parameter configuration operation to complete the proactive pre-configuration of network parameters, ensuring that each node is in a running state that meets the quality requirements when the transmission starts, and avoiding stuttering problems caused by parameter mismatch during the initial transmission stage.

[0051] Step 6.2: During the transmission process, continuously receive real-time network performance data reported by each node and compare and analyze it with preset quality thresholds. Specifically, during the county-level recording and broadcasting transmission process, including real-time recording and broadcasting transmission from 8:00 to 9:00 in the morning and playback data transmission from 19:00 to 21:00 in the evening, the system backend continuously receives real-time network performance data reported by the source node, intermediate routing nodes, and destination node at a cycle of 1 second. The real-time network performance data specifically includes the encoding output delay of the source node, the forwarding delay and available bandwidth of the link of the intermediate routing node, the receiving delay of the destination node, the packet loss rate, and the image decoding clarity. Simultaneously, the system retrieves the pre-set quality thresholds from the global transmission control strategy. These thresholds are set according to the recording and broadcasting needs of rural schools. For example, the latency threshold for real-time recording and broadcasting is ≤200ms, the packet loss rate threshold is ≤1%, and the image clarity threshold is ≥1080P. The latency threshold for playback transmission is ≤300ms, and the packet loss rate threshold is ≤0.5%. The system continuously compares and analyzes the received real-time performance data with the corresponding thresholds one by one. If the real-time latency reported by a destination node in a remote town school is 180ms and the packet loss rate is 0.8%, it is determined to meet the threshold requirements. If the real-time latency reported by another destination node in a remote town school rises to 220ms and the packet loss rate rises to 1.2%, it is determined to deviate from the threshold and needs to be marked as an anomaly to be processed.

[0052] Step 6.3: When real-time performance data deviates from the preset threshold, a dynamic adjustment mechanism is activated to recalculate the optimization parameters. The recalculated optimization parameters are then used to generate update instructions, which are sent to the corresponding network nodes for execution. Specifically, this includes: through comparative analysis, if the real-time performance data of a destination node in a remote town or school deviates from the preset threshold (e.g., a latency of 220ms exceeds the 200ms threshold, or a packet loss rate of 1.2% exceeds the 1% threshold), the dynamic adjustment mechanism is immediately activated. First, the optimized transmission path configuration information, the real-time bandwidth data of the current link, and historical adjustment records corresponding to that node are retrieved, and the optimization parameters are recalculated. For the high latency issue, the available paths are re-evaluated. If it is found that the load of an intermediate node on the current path is increased, the calculation is switched to another path. The expected latency after the backup path is established; to address the issue of excessive packet loss, the appropriate reduction value of the recording bitrate is calculated, such as from 5.5Mbps to 5Mbps, and the feasibility of supplementing the path with an additional 2Mbps of bandwidth from the backbone link is also calculated; the recalculated optimization parameters, such as the backup path node sequence after switching, the adjusted bitrate value, and the supplemented bandwidth quota, are used to generate update instructions, which are then sent to the corresponding network nodes through the system's real-time instruction channel. Among them, the path switching instruction is sent to the intermediate routing nodes and source nodes involved, the bitrate adjustment instruction is sent to the source node encoding device of the core calibration, and the bandwidth supplementation instruction is sent to the bandwidth management node of the county-level private network to ensure that each node quickly executes the update operation and promptly corrects performance data that deviates from the threshold.

[0053] Step 6.4: Based on the execution result feedback, verify the effect of parameter adjustment, and decide whether to start a new round of optimization adjustment based on the verification results to form a closed-loop control. Specifically, this includes: after the update command is issued and executed, the system continuously receives execution result data from the corresponding network nodes, such as confirmation information from intermediate routing nodes indicating that the path switching is complete, status information from source nodes indicating that the bitrate has been adjusted to 5Mbps, and resource allocation information from bandwidth management nodes indicating that 2Mbps of bandwidth has been added to the target path. At the same time, the system continuously collects real-time performance data of the destination node in the remote town / school, and observes the changes in latency and packet loss rate after adjustment. If the latency of the node decreases from 220ms to 190ms within 5 seconds after adjustment, the system will be considered. If the latency and packet loss rate decrease from 1.2% to 0.7%, both returning to the preset threshold range, then the parameter adjustment effect is verified as effective, and the current transmission quality is determined to be up to standard, without needing to initiate a new round of optimization adjustments. If the latency only decreases to 210ms after adjustment and does not return to the threshold, or the packet loss rate remains at 1.1%, then the parameter adjustment effect is verified as not meeting expectations. The system immediately re-triggers the dynamic adjustment mechanism, re-analyzes the reasons for the performance data deviation, such as whether there are other intermediate node load issues, recalculates the optimization parameters, such as further reducing the bitrate to 4.8Mbps or supplementing more bandwidth, generates new update instructions, and issues them for execution. Through the cyclical process of adjustment, verification, and re-adjustment, a complete closed-loop control is formed.

[0054] In this embodiment of the invention, by proactively pre-configuring network parameters before transmission, the strategy deployment of each node can be completed before the peak times of morning recording and evening playback, avoiding initial transmission anomalies from the source and preventing rural schools from missing the core content at the beginning of the course; during transmission, real-time data is continuously compared with quality thresholds, which can capture potential problems such as increased latency and packet loss rate in real time, breaking the limitation of relying on terminal feedback; once parameter deviation is detected, dynamic adjustment is immediately initiated and update instructions are issued, which can shorten the anomaly handling time from the current 15-20 seconds to the second level, reducing the impact of stuttering on the continuity of teaching; finally, a closed-loop control is formed through effect verification to ensure that the transmission quality returns to the standard state after adjustment.

[0055] like Figure 2 As shown, embodiments of the present invention also provide an AI-predictive-based intelligent education recording network transmission quality optimization system, comprising: The acquisition module is used to monitor and collect network performance parameters of source nodes, intermediate routing nodes and destination nodes in the recording and broadcasting transmission path in real time; The module is used to dynamically aggregate and analyze the quality parameters of multiple nodes, establish logical relationships that reflect the overall transmission status, and output a comprehensive evaluation result of the network quality change trend. The quantitative evaluation result is input into the trained AI time series prediction model to predict the probability and type of network anomalies in the future time window, and obtain transmission strategy optimization instructions including routing adjustment, bit rate adaptation and resource reallocation strategies. The calculation module is used to parse the path selection constraints and bandwidth allocation constraints based on the transmission strategy optimization instructions. Through the constraints and quality parameters, it performs transmission route optimization calculations to determine a low-latency, high-stability transmission path. It then integrates the transmission path with the adaptive rate control strategy and dynamic bandwidth reservation mechanism in the transmission strategy optimization instructions to generate a set of global transmission control strategies that the system can execute. The processing module is used to proactively pre-configure network parameters before transmission begins based on the global transmission control strategy set, and dynamically adjust the parameters based on real-time monitored quality data during transmission to achieve closed-loop control.

[0056] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for optimizing the transmission quality of smart education recording and broadcasting networks based on AI prediction, characterized in that, The method includes: Step 1: Monitor and collect network performance parameters of source nodes, intermediate routing nodes and destination nodes in the recording and broadcasting transmission path in real time; Step 2: Perform dynamic aggregation analysis on the quality parameters of multiple nodes, establish logical relationships reflecting the overall transmission status, and output a comprehensive evaluation result on the trend of network quality changes. Step 3: Input the quantitative evaluation results into the trained AI time series prediction model to predict the probability and type of network anomalies in the future time window, and obtain transmission strategy optimization instructions including routing adjustment, bit rate adaptation and resource reallocation strategies. Step 4: Based on the transmission strategy optimization instructions, the path selection constraints and bandwidth allocation constraints are parsed out. Through the constraints and quality parameters, the transmission route optimization calculation is performed to determine a low-latency and high-stability transmission path. Step 5: Integrate the adaptive rate control strategy and dynamic bandwidth reservation mechanism in the transmission path and transmission strategy optimization instructions to generate a set of global transmission control strategies that the system can execute. Step 6: Based on the global transmission control strategy set, network parameters are proactively pre-configured before transmission begins, and parameters are dynamically adjusted based on real-time monitored quality data during transmission to achieve closed-loop control.

2. The method for optimizing the transmission quality of smart education recording and broadcasting networks based on AI prediction as described in claim 1, characterized in that, Real-time monitoring and collection of network performance parameters of source nodes, intermediate routing nodes, and destination nodes in the recording and broadcasting transmission path, including: Step 1.1: Send performance data collection commands to the source node, intermediate routing nodes, and destination node; Step 1.2: Receive network performance parameters periodically reported by each node in response to instructions. These network performance parameters include latency, jitter, packet loss rate, and available bandwidth.

3. The method for optimizing the transmission quality of smart education recording and broadcasting networks based on AI prediction as described in claim 2, characterized in that, Dynamic aggregation analysis of multi-node quality parameters is performed to establish logical relationships reflecting the overall transmission status, and a comprehensive evaluation result of network quality change trends is output, including: Step 2.1: Based on the quality parameters of each node, and according to predefined aggregation rules and weight configuration, perform weighted aggregation calculation on performance parameters of different types from different nodes to generate an aggregated index that characterizes the global quality status of the entire transmission path. Step 2.2: Based on the aggregation index and the spatiotemporal correlation between the parameters of each node, construct a global transmission quality map data with nodes as vertices and transmission link quality relationships as edges; Step 2.3: Based on the key performance values ​​extracted from the aggregated indicators and the link relationship weights in the global transmission quality map data, calculate the deviation of the key performance values ​​from the historical baseline data, and combine the link relationship weights to generate the real-time quality attenuation coefficient of each link. Step 2.4: Perform time series smoothing on the quality decay coefficient, and calculate the slope and fluctuation expectation of quality change in the short term based on the processed series. Combine the deviation, quality decay coefficient, slope and fluctuation expectation, and generate a quantitative comprehensive evaluation result including quality score and trend level through a weighted decision algorithm.

4. The method for optimizing the transmission quality of smart education recording and broadcasting networks based on AI prediction as described in claim 3, characterized in that, The quantitative evaluation results are input into a trained AI time-series prediction model to predict the probability and type of network anomalies within future time windows, resulting in transmission strategy optimization instructions that include routing adjustments, bitrate adaptation, and resource reallocation strategies. Step 3.1: Based on the quantitative comprehensive evaluation results, perform feature extraction and standardization preprocessing on the results; Step 3.2: Input the preprocessed feature data into the AI ​​time series prediction module to analyze the evolution trend of network quality parameters within a specific future time window; Step 3.3: Based on the evolution trend, calculate the probability of occurrence of various network anomalies and identify the anomaly types; Step 3.4: Based on the occurrence probability and anomaly type, generate corresponding routing adjustment strategies, bitrate adaptation schemes, and resource reallocation strategies; Step 3.5: Integrate the routing adjustment strategy, bitrate adaptation scheme, and resource reallocation strategy into a system-executable transmission strategy optimization instruction.

5. The method for optimizing the transmission quality of smart education recording and broadcasting networks based on AI prediction as described in claim 4, characterized in that, Based on the path selection constraints and bandwidth allocation constraints parsed from the transmission policy optimization instructions, and using these constraints and quality parameters, transmission route optimization calculations are performed to determine low-latency, high-stability transmission paths, including: Step 4.1: Parse the transmission strategy optimization instructions, extract the path selection constraints and bandwidth allocation constraints contained therein, obtain the current network topology data and real-time quality parameters, and construct a multi-objective optimization function based on the constraints; Step 4.2: Based on the multi-objective optimization function, iterative calculations are performed in the set of optional transmission paths to evaluate the delay and stability indices of each path and obtain the evaluation results; Step 4.3: Based on the evaluation results, select the optimized transmission path that simultaneously meets the requirements of low latency and high stability, and obtain the path selection instruction containing detailed configuration information of the optimized transmission path.

6. The method for optimizing the transmission quality of smart education recording and broadcasting networks based on AI prediction as described in claim 5, characterized in that, The adaptive rate control strategy and dynamic bandwidth reservation mechanism in the transmission path and transmission strategy optimization instructions are fused to generate a global transmission control strategy set that the system can execute, including: Step 5.1: Based on the path selection instruction and the transmission strategy optimization instruction, parse the path selection instruction to obtain the optimized transmission path configuration information, and extract the adaptive bit rate control strategy parameters and dynamic bandwidth reservation mechanism parameters from the transmission strategy optimization instruction. Step 5.2: Establish the policy mapping relationship between path configuration information, rate control strategy and bandwidth reservation mechanism, and perform multi-policy collaborative optimization calculation; Step 5.3: Based on the collaborative optimization calculation results, generate a three-in-one control strategy that includes transmission path configuration, bit rate adaptive adjustment rules and bandwidth dynamic allocation scheme, and encapsulate the three-in-one control strategy into a global transmission control strategy set that the system can recognize and execute.

7. The method for optimizing the transmission quality of smart education recording and broadcasting networks based on AI prediction as described in claim 6, characterized in that, Based on the global transmission control strategy set, network parameters are proactively pre-configured before transmission begins, and dynamically adjusted based on real-time monitored quality data during transmission to achieve closed-loop control, including: Step 6.1: Generate the corresponding network device configuration instruction set according to the global transmission control policy set. Before the transmission starts, send the configuration instruction set to the source node, intermediate routing node and destination node to complete the active pre-configuration of network parameters. Step 6.2: During the transmission process, continuously receive real-time network performance data reported by each node and compare and analyze it with the preset quality threshold. Step 6.3: When real-time performance data is detected to deviate from the preset threshold, the dynamic adjustment mechanism is activated, the optimization parameters are recalculated, the recalculated optimization parameters are used to generate update instructions, and the instructions are sent to the corresponding network nodes for execution. Step 6.4: Based on the feedback of the execution results, verify the effect of parameter adjustment, and decide whether to start a new round of optimization and adjustment based on the verification results to form closed-loop control.

8. An AI-predictive-based intelligent education recording and broadcasting network transmission quality optimization system, wherein the system implements the method as described in any one of claims 1 to 7, characterized in that, include: The acquisition module is used to monitor and collect network performance parameters of source nodes, intermediate routing nodes and destination nodes in the recording and broadcasting transmission path in real time; The module is used to dynamically aggregate and analyze the quality parameters of multiple nodes, establish logical relationships that reflect the overall transmission status, and output a comprehensive evaluation result of the network quality change trend. The quantitative evaluation result is input into the trained AI time series prediction model to predict the probability and type of network anomalies in the future time window, and obtain transmission strategy optimization instructions including routing adjustment, bit rate adaptation and resource reallocation strategies. The calculation module is used to parse the path selection constraints and bandwidth allocation constraints based on the transmission strategy optimization instructions. Through the constraints and quality parameters, it performs transmission route optimization calculations to determine a low-latency, high-stability transmission path. It then integrates the transmission path with the adaptive rate control strategy and dynamic bandwidth reservation mechanism in the transmission strategy optimization instructions to generate a set of global transmission control strategies that the system can execute. The processing module is used to proactively pre-configure network parameters before transmission begins based on the global transmission control strategy set, and dynamically adjust the parameters based on real-time monitored quality data during transmission to achieve closed-loop control.

9. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images