Multi-terminal video display adaptive method based on hierarchical traffic prediction

Abstract

The invention discloses a hierarchical traffic prediction-based multi-terminal video display adaptive method, which comprises the following steps of: acquiring network traffic, user behaviors and equipment state information in real time by constructing a multi-dimensional feature acquisition module, and generating a dynamic feature matrix containing traffic, interaction and performance features; processing the feature matrix by using a deep learning model, and generating and dynamically adjusting a network traffic grading prediction result; formulating an optimized video content distribution strategy based on a prediction result and equipment performance parameters, and coordinating multi-terminal resource distribution; through user behavior analysis and network state monitoring, a self-adaptive video display control strategy is generated, dynamic adaptation of video resolution, frame rate and playing logic is realized, a full-link collaborative optimization module is constructed, terminal, network and application layer resource allocation is coordinated, a video transmission and display strategy is optimized, and network flow prediction and multi-terminal interaction requirements are combined to realize multi-terminal interaction. And generating a dynamic video layout and content switching strategy.

Description

Technical Field

[0001] The present invention relates to the technical field of video display, and particularly to an adaptive method for multi-terminal video display based on hierarchical traffic prediction. Background Art

[0002] With the rapid development of information technology and network communication, the demand for video display on multi-terminal devices is increasing continuously. Especially in scenarios where intelligent devices are widely used, users have put forward higher requirements for cross-terminal and seamless video experience. However, limited by the uneven distribution of network bandwidth and terminal device performance, the prior art faces many challenges and technical bottlenecks when realizing the adaptability of multi-terminal video display.

[0003] In the prior art, multi-terminal video display solutions mainly rely on static rule design and traditional bandwidth allocation methods. These methods have significant deficiencies in dynamically adapting to complex network environments and diverse user requirements. Specifically, the defects of the prior art include the following aspects: 1. Lack of accurate network traffic prediction ability: The network traffic prediction models in the prior art usually rely on simple linear analysis or rule matching, and it is difficult to accurately capture the dynamic changes of traffic in complex network environments. This results in the difficulty of ensuring the stability and quality of video display when the traffic fluctuates greatly.

[0004] 2. Uneven content distribution across terminal devices: In multi-terminal scenarios, traditional methods often use fixed bandwidth allocation strategies, ignoring the performance parameters and network conditions of different terminal devices. This method cannot effectively coordinate resources, easily leading to a decline in the display quality of some devices and affecting the overall user experience.

[0005] 3. Lack of integrated analysis of user behavior and network status: Existing solutions mostly take network bandwidth as the core indicator, ignoring user behavior characteristics and real-time interaction requirements. Such a single optimization perspective cannot meet the needs of user personalized experience, and it is also difficult to find the best balance between user behavior and network conditions.

[0006] 4. Slow response speed and insufficient adaptability: Traditional video display methods lack real-time performance and cannot quickly adjust video display parameters according to the dynamic changes of network status and user requirements. This delayed response mechanism leads to problems such as video playback stuttering and resolution reduction, affecting the display coherence of multi-terminal devices.

[0007] 5. Insufficient full-link optimization: The prior art usually focuses on single-point optimization, such as resource allocation at the terminal or network layer, while ignoring the synergy between the terminal, network, and application layers. Such a single-point optimization method is difficult to improve the overall system performance, and the effect is particularly limited in complex environments.

[0008] Therefore, how to provide an adaptive method for multi-terminal video display based on hierarchical traffic prediction is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0009] An object of the present invention is to propose an adaptive method for multi-terminal video display based on hierarchical traffic prediction. The present invention makes full use of deep learning models, dynamic optimization algorithms and real-time adaptation mechanisms, and details the dynamic adaptation of video display among multi-terminal devices through hierarchical prediction of network traffic, cross-terminal dynamic content distribution, and fusion analysis of user behavior and network status, with the advantages of high accuracy, fast response speed and strong display consistency.

[0010] An adaptive method for multi-terminal video display based on hierarchical traffic prediction according to an embodiment of the present invention includes the following steps: S1. Construct a multi-dimensional feature acquisition module to collect real-time network traffic information, user behavior data and device operation status information from multi-terminal devices, and generate a multi-dimensional dynamic feature matrix, where the multi-dimensional dynamic feature matrix includes network traffic features, user interaction features and device performance features; S2. Use a deep learning hierarchical traffic prediction model to process the multi-dimensional dynamic feature matrix, generate a network traffic hierarchical prediction result through feature extraction and time series analysis, and dynamically adjust the network traffic hierarchical prediction result according to the input data updated in real time to accurately reflect changes in the network environment; S3. Construct a cross-terminal video content distribution module, generate a dynamic content distribution strategy based on the network traffic hierarchical prediction result, terminal device performance parameters and network resource constraint conditions, and solve the dynamic content distribution strategy through an optimization algorithm to coordinate resource allocation among multi-terminal devices; S4. Fusion the output data of the user behavior analysis module and the network status monitoring module to generate an adaptive video display control strategy, where the adaptive video display control strategy realizes the dynamic adaptation of video display parameters to user needs and network conditions by dynamically adjusting video resolution, frame rate, bit rate and playback logic; S5. Construct a full-link collaborative optimization module to coordinate resource allocation among the terminal layer, network layer and application layer through a network status real-time feedback mechanism, and dynamically optimize video transmission parameters and display strategies; S6. Generate a dynamic video display layout and content switching strategy based on the real-time interaction requirements of cross-terminal devices and the network traffic hierarchical prediction result, where the dynamic content distribution strategy is used to keep the video display consistent among multi-terminal devices, and the content switching strategy optimizes the display coherence in a multi-device environment; S7. Construct a real-time video display adaptation process to dynamically update the video display strategy by combining the network traffic hierarchical prediction result, user behavior characteristics and multi-terminal interaction requirements.

[0011] Optionally, S1 specifically includes: S11. Construct a network traffic collection interface to collect network traffic parameters from multiple terminal devices in real time, including the current bandwidth utilization rate, network latency, packet loss rate, and jitter parameters, and represent them as a network traffic feature vector , where is the th component of the network traffic parameter, is the network traffic feature dimension; S12. Record the real-time user behavior data through the user behavior monitoring module, including video playback operations, interaction events, and user preference features, and convert them into a user behavior feature vector , where is the th component of the user behavior feature, is the user behavior feature dimension; S13. Establish a device status monitoring module to obtain the running status information of the terminal device in real time, including the processor utilization rate, storage utilization rate, and battery status, and represent this information as a device running status feature vector , where is the th component of the device status feature, is the device running status feature dimension; S14. Perform data preprocessing on the network traffic feature vector, user behavior feature vector, and device running status feature vector to generate the preprocessed feature vectors , and ; S15. Construct a multi-dimensional dynamic feature matrix based on the preprocessed feature vectors, and input the multi-dimensional dynamic feature matrix into the deep learning hierarchical traffic prediction model to perform feature extraction and classification calculation on : ; Among them, is the multi-dimensional dynamic feature matrix, , , respectively represent the network traffic feature, user behavior feature, and device status feature after preprocessing.

[0012] Optionally, S2 specifically includes: S21. Construct a deep learning hierarchical traffic prediction model. The deep learning hierarchical traffic prediction model includes a feature extraction layer, a time series modeling layer, and a hierarchical prediction layer. The feature extraction layer uses convolutional operations to extract the multi-dimensional dynamic feature matrix The local features are modeled for time dependence by the time series modeling layer based on the long short-term memory network, and the hierarchical prediction layer generates the hierarchical prediction results of the network traffic through multi-class logistic regression; S22. In the feature extraction layer, two-dimensional convolution operations are used to extract the local features of the multi-dimensional dynamic feature matrix: ; Among them, is the extracted local feature, is the weight matrix of the th convolutional kernel, is the th feature map of the input matrix, represents the convolution operation, is the bias term, is the activation function; S23. In the time series modeling layer, the local feature is input into the long short-term memory network, and the time dependence of the feature sequence is captured through the recursive structure to generate the time series feature representation , where encodes the dynamic characteristics of the network traffic at different times; S24. In the hierarchical prediction layer, a classification model is constructed based on the time series feature and mapped to the hierarchical prediction output : ; Among them, is the hierarchical prediction result of the network traffic, is the weight matrix of the fully connected layer, is the bias term, is the total number of classification categories; S25. Based on the hierarchical prediction result , the model parameters are optimized through the backpropagation algorithm, and the weight matrices and bias terms of the feature extraction layer, the time series modeling layer, and the hierarchical prediction layer are updated. Combining the dynamic feedback mechanism, the updated model is applied to the new multi-dimensional dynamic feature matrix to continuously generate the hierarchical prediction results and dynamically adjust the model weights.

[0013] Optionally, the S3 specifically includes: S31. Construct a cross-terminal video content distribution module that receives the network traffic hierarchical prediction result , the set of terminal device performance parameters and the network resource constraint conditions as inputs; S32. According to , and Establish an optimization model for dynamic content distribution strategy: ; Among them, is the optimization objective function, is the content quality scoring function of the th terminal, is the content allocation resource constraint of the th terminal, is the content transmission delay of the th network path, and are weight factors; S33. Use the gradient descent optimization algorithm to solve the optimization model of the dynamic content distribution strategy, and generate a set of content distribution strategies , where is the video content strategy allocated to the th terminal device; S34. Based on , dynamically adjust the video content distribution and transmission strategies among multiple terminal devices, optimize the resolution, frame rate, and encoding parameters of the content, and coordinate the use of network resources.

[0014] Optionally, the specific steps of S4 include: S41. Construct a user behavior analysis module to extract a set of user behavior characteristics , where includes the user's playback preferences, interaction events, viewing time distribution, and historical behavior patterns, is the th component of the user behavior characteristics; S42. Construct a network status monitoring module to collect a set of network status characteristics in real time, where includes the current bandwidth, network delay, packet loss rate, and jitter parameters, is the th component of the network status characteristics; S43. Fuse the set of user behavior characteristics with the set of network status characteristics , and generate a fused feature vector through feature weighted fusion: ; Among them, is the fused feature vector, and are the weighted matrices for user behavior characteristics and network status characteristics respectively, is the fused bias vector; S44. Based on the fused feature vector , construct an adaptive video display control strategy optimization model: ; Among them, is the optimized display control strategy, is the display quality function based on user behavior and network status, is the penalty function for latency and packet loss, and are weight factors to dynamically balance the display control strategy between user experience and resource utilization; S45. According to the optimized display control strategy , generate a video display parameter set , including resolution, frame rate, bit rate, and adjustment parameters for the playback logic. By dynamically adjusting the video display parameter set in real time, achieve dynamic adaptation of the display effect to user needs and network status.

[0015] Optionally, the specific content of S5 includes: S51. Construct a full-link collaborative optimization module, including a terminal layer resource allocation sub-module, a network layer transmission optimization sub-module, and an application layer policy adjustment sub-module. The three-layer module realizes dynamic resource allocation and policy synchronization adjustment through a collaborative optimization mechanism; S52. At the terminal layer, collect the resource usage status set of the terminal device in real time, where includes the processing power, storage utilization rate, battery power, and current task load of the terminal, is the resource status of the S53. At the network layer, collect the network status parameter set in real time, where includes the current network bandwidth, latency, packet loss rate, and jitter, is the status parameter of the S54. Based on the terminal layer resource usage status and the network status parameters , construct a full-link resource allocation optimization model: ; Among them, is the full-link resource allocation optimization goal, is the resource allocation utility function of the th terminal device and the corresponding network path, is the full-link network resource consumption function, and is the weight factor; S55. Collect status change data from the terminal layer, network layer, and application layer through a real-time feedback mechanism, and update the resource allocation set using the feedback optimization strategy , where is the video transmission and display resource strategy allocated to the th terminal device; S56. Based on the updated resource allocation set , dynamically adjust the video transmission parameter set , where represent the resolution, frame rate, and bit rate parameters respectively, and at the same time adjust the video display logic to achieve the integrated optimization of display quality, resource utilization efficiency, and terminal synchronous display.

[0016] Optionally, the specific steps of S6 are as follows: S61. Collect and process the real-time interaction requirement set of multiple terminal devices , including the operation records, content preferences, interaction frequencies, and device priority information of terminal users, is the interaction requirement feature of the th terminal device; S62. Integrate the network traffic hierarchical prediction result with the real-time interaction requirement set to generate the input feature set , where contains network traffic features, device interaction requirement features, and display priority weights, is the integrated input feature; S63. Based on the input feature set , construct a dynamic video display layout optimization model: ; where, is the layout optimization objective, is the display consistency scoring function of the th terminal device, is the video content feature allocated to the th terminal device, is the content switching transmission delay penalty function based on the hierarchical prediction result, and are the optimization weight factors; S64. Solve the layout optimization model to generate a dynamic content allocation strategy , construct a video content plan allocated to each terminal device, is the Content distribution strategy for a terminal device; S65. Based on the dynamic content distribution strategy generate and apply a content switching strategy , including the triggering conditions for content switching, switching paths, and display logic optimization parameters, and comprehensively adjusting the video display layout and content switching logic.

[0017] Optionally, the S7 specifically includes: S71. Integrate the network traffic classification prediction results , the set of user behavior characteristics and the set of multi-terminal device interaction requirements to generate a comprehensive feature input vector , where includes dynamic network traffic characteristics, user behavior preferences, and terminal interaction requirements, is the th component of the comprehensive feature; S72. Based on the comprehensive feature input vector , construct a real-time video display adaptation optimization model: ; where is the optimization objective of the video display strategy, is the display quality function based on user behavior characteristics and interaction requirements, is the network resource consumption function caused by the network traffic classification prediction result, and are weight factors; S73. Generate a set of real-time video display strategies through the solution of the optimization model , including dynamic adjustment strategies for video resolution, frame rate, bit rate, and playback logic, is the display strategy allocated to the th terminal device; S74. Combine the real-time monitoring data of the network status and user interaction behavior, and iteratively optimize and adjust the set of video display strategies to synchronously optimize the set of video transmission parameters , where respectively represent resolution, frame rate, and bit rate parameters, and update the display logic to achieve real-time synchronization and coherence optimization of video display among multi-terminal devices.

[0018] The beneficial effects of the present invention are: (1) By combining a deep learning hierarchical traffic prediction model, a dynamic content distribution strategy optimization model, and a full-link collaborative optimization mechanism, the present invention realizes accurate prediction of dynamic changes in network traffic, efficient allocation of multi-terminal device resources, and integrated analysis of user behavior and network status, enabling the system to dynamically adapt to complex network environments and diverse user needs, effectively improving the quality and consistency of multi-terminal video displays, especially maintaining the coherence and fluency of displays under complex network conditions.

[0019] (2) Through a real-time feedback mechanism and multi-level collaborative optimization, the present invention establishes an efficient information interaction channel among the terminal, network, and application layers, significantly reducing the system response time and quickly adjusting video display parameters, thereby achieving synchronous video display and dynamic adaptation among multi-terminal devices, and significantly improving the response efficiency and adaptability of the system.

[0020] (3) By integrating a user behavior analysis module and a network status monitoring module, comprehensively considering user preferences, interaction requirements, and network resource constraints, dynamically generating video display layouts and content switching strategies, the present invention realizes personalized optimization of the user experience and rational utilization of network resources, reduces display inconsistency problems caused by resource competition, and at the same time ensures the stability and efficiency of the display effects of multi-terminal devices.

[0021] (4) By combining a real-time video display adaptation process and a dynamic display strategy update mechanism, and dynamically adjusting video display parameters and transmission strategies through iterative optimization, the present invention can effectively cope with fluctuations in network traffic and changes in user needs, ensuring that the video display system in a multi-terminal, multi-network environment has higher robustness, real-time performance, and display performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation to the present invention. In the drawings: Figure 1 It is a system architecture block diagram of an adaptive method for multi-terminal video display based on hierarchical traffic prediction proposed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0023] Now, the present invention will be further described in detail with reference to the drawings. These drawings are all simplified schematic diagrams, only illustrating the basic structure of the present invention in a schematic manner, so they only show the components related to the present invention.

[0024] Refer to Figure 1 , an adaptive method for multi-terminal video display based on hierarchical traffic prediction, including the following steps: S1. Build a multi-dimensional feature collection module to collect real-time network traffic information, user behavior data, and device operation status information from multiple terminal devices, and generate a multi-dimensional dynamic feature matrix, where the multi-dimensional dynamic feature matrix includes network traffic features, user interaction features, and device performance features; S2. Use a deep learning hierarchical traffic prediction model to process the multi-dimensional dynamic feature matrix, generate a network traffic hierarchical prediction result through feature extraction and time series analysis, and the network traffic hierarchical prediction result is dynamically adjusted according to the real-time updated input data to accurately reflect the changes in the network environment; S3. Build a cross-terminal video content distribution module, generate a dynamic content distribution strategy based on the network traffic hierarchical prediction result, terminal device performance parameters, and network resource constraints, and solve the dynamic content distribution strategy through an optimization algorithm to coordinate the resource allocation among multiple terminal devices; S4. Integrate the output data of the user behavior analysis module and the network status monitoring module to generate an adaptive video display control strategy, where the adaptive video display control strategy realizes the dynamic adaptation of video display parameters to user needs and network conditions by dynamically adjusting video resolution, frame rate, bit rate, and playback logic; S5. Build a full-link collaborative optimization module to coordinate resource allocation among the terminal layer, network layer, and application layer through a network status real-time feedback mechanism, and dynamically optimize video transmission parameters and display strategies; S6. Generate a dynamic video display layout and content switching strategy based on the real-time interaction requirements of cross-terminal devices and the network traffic hierarchical prediction result, where the dynamic content distribution strategy is used to keep the video display consistent among multiple terminal devices, and the content switching strategy optimizes the display coherence in a multi-device environment; S7. Build a real-time video display adaptation process to dynamically update the video display strategy by combining the network traffic hierarchical prediction result, user behavior characteristics, and multi-terminal interaction requirements.

[0025] In this embodiment, S1 specifically includes: S11. Build a network traffic collection interface to collect network traffic parameters from multiple terminal devices in real time, including current bandwidth utilization rate, network latency, packet loss rate, and jitter parameters, and represent them as a network traffic feature vector , where is the th component of the network traffic parameter, is the network traffic feature dimension; S12. Record the real-time behavior data of users through the user behavior monitoring module, including video playback operations, interaction events, and user preference characteristics, and convert them into a user behavior feature vector , where is the component is the user behavior feature dimension; S13. Establish a device status monitoring module to obtain the running status information of the terminal device in real time, including the processor utilization rate, storage utilization rate, and battery status, and represent this information as a device running status feature vector , where is the th component of the device status feature, is the device running status feature dimension; S14. Perform data preprocessing on the network traffic feature vector, user behavior feature vector, and device running status feature vector to generate the preprocessed feature vectors , and ; S15. Construct a multi-dimensional dynamic feature matrix based on the preprocessed feature vectors, and input the multi-dimensional dynamic feature matrix into the deep learning hierarchical traffic prediction model to perform feature extraction and classification calculation on : ; Among them, is the multi-dimensional dynamic feature matrix, , , respectively represent the network traffic feature, user behavior feature, and device status feature after preprocessing.

[0026] In this embodiment, the S2 specifically includes: S21. Construct a deep learning hierarchical traffic prediction model. The deep learning hierarchical traffic prediction model includes a feature extraction layer, a time series modeling layer, and a hierarchical prediction layer. The feature extraction layer uses convolutional operations to extract the local features of the multi-dimensional dynamic feature matrix . The time series modeling layer performs time dependence modeling on the extracted local features based on the long short-term memory network, and the hierarchical prediction layer generates a hierarchical prediction result of the network traffic through multi-class logistic regression; S22. In the feature extraction layer, use two-dimensional convolutional operations to extract the local features of the multi-dimensional dynamic feature matrix: ; Among them, is the extracted local feature, is the weight matrix of the th convolutional kernel, is the th feature map of the input matrix, represents the convolutional operation, is the bias term, is the activation function; S23. In the time series modeling layer, input the local features into the long short-term memory network, capture the time dependence of the feature sequence through the recursive structure, and generate the time series feature representation , where encodes the dynamic characteristics of network traffic at different times; S24. In the hierarchical prediction layer, based on the time series feature construct a classification model and map it to the hierarchical prediction output : ; wherein, is the hierarchical prediction result of network traffic, is the weight matrix of the fully connected layer, is the bias term, is the total number of classification categories; S25. Based on the hierarchical prediction result , optimize the model parameters through the backpropagation algorithm, update the weight matrix and bias term of the feature extraction layer, the time series modeling layer and the hierarchical prediction layer, and combine the dynamic feedback mechanism to apply the updated model to the new multi-dimensional dynamic feature matrix , continuously generate the hierarchical prediction result, and dynamically adjust the model weight.

[0027] In this embodiment, the specific steps of S3 are as follows: S31. Construct a cross-terminal video content distribution module, and receive the network traffic hierarchical prediction result , the set of terminal device performance parameters and the network resource constraint condition as inputs; S32. According to , and establish a dynamic content distribution strategy optimization model: ; wherein, is the optimization objective function, is the content quality scoring function of the th terminal, is the content distribution resource constraint of the th terminal, is the content transmission delay of the th network path, and are weight factors; S33. Use the gradient descent optimization algorithm to solve the dynamic content distribution strategy optimization model and generate the content distribution strategy set , where is the video content policy assigned to the th terminal device; S34. Based on , dynamically adjust the video content distribution and transmission policies among multiple terminal devices, optimize the resolution, frame rate, and encoding parameters of the content, and coordinate the use of network resources.

[0028] In this embodiment, the specific steps of S4 are as follows: S41. Construct a user behavior analysis module to extract the set of user behavior characteristics , where includes the user's playback preferences, interaction events, viewing time distribution, and historical behavior patterns, is the th component of the user behavior characteristics; S42. Construct a network status monitoring module to collect the set of network status characteristics in real time , where includes the current bandwidth, network latency, packet loss rate, and jitter parameters, is the th component of the network status characteristics; S43. Fuse the set of user behavior characteristics and the set of network status characteristics to generate a fused feature vector : ; where, is the fused feature vector, and are the weighting matrices for user behavior characteristics and network status characteristics respectively, is the fusion bias vector; S44. Based on the fused feature vector , construct an adaptive video display control strategy optimization model: ; where, is the optimized display control strategy, is the display quality function based on user behavior and network status, is the penalty function for delay and packet loss, and are the weight factors to dynamically balance the display control strategy between user experience and resource utilization; S45. According to the optimized display control strategy , generate the set of video display parameters , Adjustment parameters including resolution, frame rate, bit rate, and playback logic, which dynamically adapt the display effect to the user's requirements and network status by dynamically adjusting the video display parameter set in real time.

[0029] In this embodiment, S5 specifically includes: S51. Construct a full-link collaborative optimization module, including a terminal-layer resource allocation sub-module, a network-layer transmission optimization sub-module, and an application-layer policy adjustment sub-module. The three-layer modules achieve dynamic resource allocation and policy synchronization adjustment through a collaborative optimization mechanism; S52. At the terminal layer, collect the resource usage status set of the terminal device in real time , where includes the processing capacity, storage utilization rate, battery power, and current task load of the terminal, being the resource status of the th terminal device; S53. At the network layer, collect the network status parameter set in real time , where includes the current network bandwidth, delay, packet loss rate, and jitter, being the status parameters of the th network path; S54. Based on the terminal-layer resource usage status and the network status parameters , construct a full-link resource allocation optimization model: ; where is the full-link resource allocation optimization objective, is the resource allocation utility function of the th terminal device and the corresponding network path, is the full-link network resource consumption function, and are weight factors; S55. Through a real-time feedback mechanism, collect status change data from the terminal layer, network layer, and application layer, and use the feedback optimization strategy to update the resource allocation set , where is the video transmission and display resource policy allocated to the th terminal device; S56. Based on the updated resource allocation set , dynamically adjust the video transmission parameter set , where respectively represent the resolution, frame rate, and bit rate parameters, and at the same time adjust the video display logic to achieve an integrated optimization of display quality, resource utilization efficiency, and terminal synchronous display.

[0030] In this embodiment, S6 specifically includes: S61. Collect and process the real-time interaction requirement sets of multiple terminal devices , including the operation records, content preferences, interaction frequencies, and device priority information of terminal users, which are the interaction requirement characteristics of the th terminal device; S62. Integrate the network traffic classification prediction results with the real-time interaction requirement sets to generate an input feature set , where it includes network traffic characteristics, device interaction requirement characteristics, and display priority weights, which are the integrated input features; S63. Based on the input feature set , construct a dynamic video display layout optimization model: ; where is the layout optimization objective, is the display consistency scoring function of the th terminal device, is the video content feature assigned to the th terminal device, is the content switching transmission delay penalty function based on the classification prediction results, and are the optimization weight factors; S64. Solve through the layout optimization model to generate a dynamic content allocation strategy , construct a video content plan assigned to each terminal device, which is the content allocation strategy of the th terminal device; S65. Based on the dynamic content allocation strategy , generate and apply a content switching strategy , including the triggering conditions, switching paths, and display logic optimization parameters of content switching, and comprehensively adjust the video display layout and content switching logic.

[0031] In this embodiment, S7 specifically includes: S71. Integrate the network traffic classification prediction results , the user behavior feature set and the multi-terminal device interaction requirement sets to generate a comprehensive feature input vector , where including dynamic network traffic characteristics, user behavior preferences, and terminal interaction requirements, is the th component of the comprehensive feature; S72. Based on the comprehensive feature input vector , construct a real-time video display adaptation optimization model: ; Among them, is the optimization goal of the video display strategy, is the display quality function based on user behavior characteristics and interaction requirements, is the network resource consumption function caused by the network traffic classification prediction result, and are weight factors; S73. Generate a set of real-time video display strategies , including dynamic adjustment strategies for video resolution, frame rate, bit rate, and playback logic, is the display strategy allocated to the th terminal device; S74. Combine the real-time monitoring data of the network status and user interaction behavior, and iteratively optimize and adjust the set of video display strategies , synchronously optimize the set of video transmission parameters , where respectively represent the resolution, frame rate, and bit rate parameters, and update the display logic to achieve real-time synchronization and coherence optimization of video display among multiple terminal devices.

[0032] Example 1: To verify the feasibility of the present invention, the present invention is applied to the multi-terminal video display system of a technology company B that provides global video streaming services. The streaming service operated by Company B covers millions of users, and they watch videos through devices such as smartphones, tablets, smart TVs, and personal computers. Due to the complex network environment, diverse user needs, and differences in device performance, the existing video display solutions of Company B face significant technical bottlenecks. Users frequently feedback problems such as video playback stuttering, image quality degradation, and device display asynchronization during peak hours. This not only has an adverse impact on the user experience but also increases the maintenance pressure on the technical team. Especially in high network load scenarios, the existing technology lacks the ability to dynamically adapt to the network status and cannot achieve real-time adjustment of video playback parameters, resulting in low resource allocation efficiency.

[0033] To solve the above problems, Company B decides to deploy the multi-terminal video display adaptive method based on hierarchical traffic prediction proposed by the present invention. By constructing a deep learning hierarchical traffic prediction model and combining a user behavior analysis module, a network status monitoring module, and a full-link collaborative optimization module, the system can monitor the network status and user behavior in real time, and dynamically generate a video display strategy based on the hierarchical traffic prediction result, so as to achieve video display synchronization and optimization among multi-terminal devices.

[0034] In a typical home scenario, a family of four watches video content simultaneously through a smart TV, two tablets, and a smartphone. The smart TV plays high-definition TV dramas, the tablets are used by children to watch cartoons, and the smartphone user browses short videos through a mobile network outdoors. In such a complex environment with multiple scenarios and diverse requirements, traditional video distribution strategies often struggle to ensure balanced resource utilization and display consistency. Through the method of the present invention, the system first monitors the network status of each terminal in real time, including the current bandwidth, latency, and packet loss rate, while analyzing user behavior characteristics such as interaction frequency, content priority, and device usage patterns. Combining this data, the deep learning model generates a hierarchical traffic prediction result and dynamically adjusts the video resolution, frame rate, and bit rate based on the cross-terminal optimization model. Finally, the smart TV maintains smooth playback in 4K high-definition quality, and there is no lag in the playback of the tablets and smartphones, resulting in a significant improvement in the user feedback experience.

[0035] To verify the optimization effect of the system, Company B selects the peak period of a quarter for data comparison and analysis. The performance indicators of the system before and after optimization are recorded in the following table.

[0036] Table 1 Data on the Optimization Effect of Company B's Multi-Terminal Video Display System

[0037] It can be seen from the data that the method of the present invention shows significant advantages in solving multi-terminal video display problems. Before the system was deployed, the number of video lags reported by users was as high as 986 times per month, which decreased to 36 times after optimization. The video playback synchronization rate of multi-terminal devices increased significantly from 73.4% before optimization to 97.8%. Especially during peak hours, due to the inability of the traditional solution to accurately allocate network resources, the resolution fluctuated frequently, while the method of the present invention successfully reduced the number of resolution adjustments from 158 times per month to 12 times per month, and the user satisfaction score for the video playback experience also increased from 3.6 points to 4.8 points.

[0038] In an actual scenario, the present invention also demonstrates extremely strong dynamic adaptation capabilities. For example, during a peak evening rush hour, a user watches video content simultaneously on a smart TV, a tablet, and a smartphone. Insufficient network bandwidth causes traditional solutions to fail to meet the high-quality playback requirements of all devices. The picture quality of the smart TV deteriorates, and the smartphone playback lags. However, the present invention adjusts the bitrate of the smartphone in real time and preferentially allocates bandwidth to the smart TV. The smart TV maintains high-definition picture quality, the smartphone plays smoothly, and the display quality of the tablet is also guaranteed. During the entire process, the response time of the system is only 1.5 seconds, and the user does not notice any abnormal playback.

[0039] In addition, the optimized system achieves efficient utilization of bandwidth, and the average utilization rate of network resources increases from 65% to 91%. In a specific test, when a user performs video-on-demand through a smart TV, the system monitors that the home network bandwidth fluctuates greatly. Through the hierarchical traffic prediction and dynamic content distribution strategy, the system preferentially adjusts the bitrates of the tablet and the smartphone to ensure that the video playback quality of the smart TV is not affected, and the tablet and the smartphone still play smoothly. This optimization strategy makes the video experience in the entire home scenario more harmonious and smooth.

[0040] The application of the present invention in Company B shows that through accurate traffic prediction, multi-dimensional data fusion, and real-time dynamic optimization, the performance of the video display system has been significantly improved. Especially in complex network environments and multi-terminal device scenarios, the present invention can achieve video display synchronization and efficient resource allocation among multi-terminal devices, greatly enhancing the user experience.

[0041] The above is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution of the present invention and its inventive concept, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. An adaptive method for multi-terminal video display based on hierarchical traffic prediction, characterized in that It includes the following steps: S1. Construct a multi-dimensional feature acquisition module to collect real-time network traffic information, user behavior data, and device operation status information from multiple terminal devices, and generate a multi-dimensional dynamic feature matrix, where the multi-dimensional dynamic feature matrix includes network traffic features, user interaction features, and device performance features; S2. Use a deep learning hierarchical traffic prediction model to process the multi-dimensional dynamic feature matrix, generate a network traffic hierarchical prediction result through feature extraction and time series analysis, and the network traffic hierarchical prediction result is dynamically adjusted according to the real-time updated input data to accurately reflect the changes in the network environment; S3. Construct a cross-terminal video content distribution module, generate a dynamic content distribution strategy based on the network traffic hierarchical prediction result, terminal device performance parameters, and network resource constraints, and solve the dynamic content distribution strategy through an optimization algorithm to coordinate the resource allocation among multiple terminal devices; S4. Integrate the output data of the user behavior analysis module and the network status monitoring module to generate an adaptive video display control strategy, where the adaptive video display control strategy realizes the dynamic adaptation of video display parameters to user needs and network conditions by dynamically adjusting video resolution, frame rate, bit rate, and playback logic; S5. Construct a full-link collaborative optimization module to coordinate resource allocation among the terminal layer, network layer, and application layer through a network status real-time feedback mechanism, and dynamically optimize video transmission parameters and display strategies; S6. Generate a dynamic video display layout and content switching strategy based on the real-time interaction requirements of cross-terminal devices and the network traffic hierarchical prediction result, where the dynamic content distribution strategy is used to keep the video display consistent among multiple terminal devices, and the content switching strategy optimizes the display coherence in a multi-device environment; S7. Construct a real-time video display adaptation process to dynamically update the video display strategy by combining the network traffic hierarchical prediction result, user behavior characteristics, and multi-terminal interaction requirements.

2. The adaptive method for multi-terminal video display based on hierarchical traffic prediction according to claim 1, characterized in that The specific content of S1 includes: S11. Construct a network traffic collection interface to collect network traffic parameters from multiple terminal devices in real time, including the current bandwidth utilization rate, network latency, packet loss rate, and jitter parameters, and represent them as a network traffic feature vector , where is the th component of the network traffic parameter, is the network traffic feature dimension; S12. Record the real-time user behavior data through the user behavior monitoring module, including video playback operations, interaction events, and user preference characteristics, and convert them into user behavior feature vectors , where is the th component of the user behavior feature, and is the user behavior feature dimension; S13. Establish a device status monitoring module to obtain the operating status information of the terminal device in real time, including the processor utilization rate, storage utilization rate, and battery status, and represent this information as a device operating status feature vector , where is the th component of the device status feature, and is the dimension of the device operating status feature; S14. Perform data preprocessing on the network traffic feature vector, user behavior feature vector, and device operation status feature vector to generate the preprocessed feature vector , and ; S15. Construct a multi-dimensional dynamic feature matrix based on the preprocessed feature vectors , and input the multi-dimensional dynamic feature matrix into the deep learning hierarchical traffic prediction model to perform feature extraction and classification calculations on : ； Among them, is a multi-dimensional dynamic feature matrix, , , respectively represent the network traffic features, user behavior features, and device status features after preprocessing.

3. An adaptive method for multi-terminal video display based on hierarchical traffic prediction according to claim 1, characterized in that, The specific content of S2 includes: S21. Construct a deep learning hierarchical traffic prediction model. The deep learning hierarchical traffic prediction model includes a feature extraction layer, a time series modeling layer, and a hierarchical prediction layer. The feature extraction layer extracts the local features of a multi-dimensional dynamic feature matrix using convolutional operations. The time series modeling layer models the time dependence of the extracted local features based on a long short-term memory network. The hierarchical prediction layer generates a hierarchical prediction result of network traffic through multi-class logistic regression. S22. In the feature extraction layer, use two-dimensional convolution operations to extract the local features of the multi-dimensional dynamic feature matrix: ； Among them, is the extracted local feature, is the weight matrix of the -th convolutional kernel, is the -th feature map of the input matrix, represents the convolution operation, is the bias term, is the activation function; S23. In the timing modeling layer, the local features are input into a long short-term memory network to capture the temporal dependence of the feature sequence through a recursive structure, generating a time series feature representation , where encodes the dynamic characteristics of network traffic at different times; S24. In the hierarchical prediction layer, based on the time series features construct a classification model and map it to the hierarchical prediction output : ； Among them, is the hierarchical prediction result of network traffic, is the weight matrix of the fully connected layer, is the bias term, is the total number of classification categories; S25. Based on the hierarchical prediction results Optimize the model parameters through the backpropagation algorithm, update the weight matrices and bias terms of the feature extraction layer, the time series modeling layer, and the hierarchical prediction layer, and combine the dynamic feedback mechanism to apply the updated model to the new multi-dimensional dynamic feature matrix Continuously generate hierarchical prediction results and dynamically adjust the model weights.

4. The adaptive method for multi-terminal video display based on hierarchical traffic prediction according to claim 1, wherein The specific content of S3 includes: S31. Construct a cross-terminal video content distribution module and receive the network traffic classification prediction result , the set of terminal device performance parameters and network resource constraint conditions as inputs; S32. Establish a dynamic content distribution policy optimization model based on , and : ； Among them, is the optimization objective function, is the content quality scoring function of the th terminal, is the content allocation resource constraint of the th terminal, is the content transmission delay of the th network path, and are weight factors; S33. Solve the dynamic content allocation strategy optimization model using the gradient descent optimization algorithm to generate a content allocation strategy set , where is the video content strategy allocated to the th terminal device; S34. Based on , dynamically adjust the video content distribution and transmission strategy among multiple terminal devices, optimize the resolution, frame rate and encoding parameters of the content, and coordinate the use of network resources.

5. A multi - terminal video display adaptive method based on hierarchical traffic prediction according to claim 1, characterized in that, The specific content of S4 includes: S41. Construct a user behavior analysis module to extract the set of user behavior characteristics , where include the user's playback preferences, interaction events, viewing time distribution, and historical behavior patterns, is the th component of the user behavior characteristics; S42. Construct a network status monitoring module to collect the network status feature set in real time , where it includes the current bandwidth, network delay, packet loss rate, and jitter parameter, which is the th component of the network status feature; S43. Integrate the set of user behavior characteristics with the set of network status characteristics , and generate a fused feature vector through weighted feature fusion : ； Among them, is the fused feature vector, and are the weighted matrices for user behavior features and network status features respectively, is the fused bias vector; S44. Based on the fused feature vector , construct an optimized model for the adaptive video display control strategy: ； Among them, is the optimized display control strategy, is the display quality function based on user behavior and network status, is the penalty function for latency and packet loss, and is the weight factor to dynamically balance the display control strategy between user experience and resource utilization; S45. According to the optimized display control strategy , generate a set of video display parameters , including adjustment parameters for resolution, frame rate, bit rate, and playback logic. By dynamically adjusting the set of video display parameters in real time, the dynamic adaptation of the display effect to the user's needs and network status is achieved.

6. The adaptive method for multi - terminal video display based on hierarchical traffic prediction according to claim 1, wherein, The specific content of S5 includes: S51. Construct a full-link collaborative optimization module, including a terminal layer resource allocation sub-module, a network layer transmission optimization sub-module, and an application layer policy adjustment sub-module. The three-layer modules realize dynamic resource allocation and policy synchronization adjustment through a collaborative optimization mechanism; S52. At the terminal layer, the resource usage status set of the terminal device is collected in real time , where includes the processing capacity of the terminal, storage utilization rate, battery power, and current task load, is the resource status of the th terminal device; S53. At the network layer, collect the set of network status parameters in real time , where include the current network bandwidth, latency, packet loss rate, and jitter, is the status parameter of the th network path; S54. Based on the resource usage status of the terminal layer and network status parameters , construct an optimization model for full-link resource allocation: ； Among them, is the optimization objective of full-link resource allocation, is the resource allocation utility function of the th terminal device and the corresponding network path, is the full-link network resource consumption function, and are weight factors; S55. Collect status change data from the terminal layer, network layer, and application layer through a real-time feedback mechanism, and update the resource allocation set using a feedback optimization strategy , where is the video transmission and display resource policy allocated to the th terminal device; S56. Based on the updated resource allocation set , dynamically adjust the video transmission parameter set , where respectively represent the resolution, frame rate, and bit rate parameters, and at the same time adjust the video display logic to achieve an integrated optimization of display quality, resource utilization efficiency, and terminal synchronous display.

7. A multi - terminal video display adaptive method based on hierarchical traffic prediction according to claim 1, characterized in that, The specific content of S6 includes: S61. Collect and process the set of real-time interaction requirements of multiple terminal devices , including the operation records, content preferences, interaction frequencies, and device priority information of the end-users, which are the interaction requirement characteristics of the th terminal device; S62. Integrated network traffic classification prediction results and the real-time interaction requirement set to generate an input feature set wherein it includes network traffic features, device interaction requirement features, and display priority weights, which are the integrated input features; S63. Based on the input feature set , construct a dynamic video display layout optimization model: ； Among them, is the layout optimization objective, is the display consistency scoring function for the th terminal device, is the video content feature assigned to the th terminal device, is the content switching transmission delay penalty function based on the hierarchical prediction result, and are the optimization weight factors; S64. Solve the generation of a dynamic content distribution strategy through layout optimization of the model , Construct a video content plan allocated to each terminal device, for the content distribution strategy of the th terminal device; S65. Based on the dynamic content distribution strategy , generate and apply a content switching strategy , including the triggering conditions for content switching, switching paths, and display logic optimization parameters, and comprehensively adjust the video display layout and content switching logic.

8. A multi-terminal video display adaptive method based on hierarchical traffic prediction according to claim 1, characterized in that, The specific content of S7 includes: S71. Integrated network traffic classification prediction results , user behavior feature set and multi-terminal device interaction requirement set to generate a comprehensive feature input vector , where includes dynamic network traffic features, user behavior preferences, and terminal interaction requirements, is the th component of the comprehensive feature; S72. Based on the comprehensive feature input vector , construct a real-time video display adaptation optimization model: ； Among them, is the optimization goal of the video display strategy, is the display quality function based on user behavior characteristics and interaction requirements, is the network resource consumption function caused by the network traffic classification prediction result, and is the weight factor; S73. Generate a set of real-time video display strategies by optimizing the solution of the model , including dynamic adjustment strategies for video resolution, frame rate, bit rate, and playback logic, which is the display strategy allocated to the th terminal device; S74. Combining the real-time monitoring data of network status and user interaction behavior, iteratively optimize and adjust the video display strategy set , and synchronously optimize the video transmission parameter set , where respectively represent the resolution, frame rate, and bit rate parameters, and update the display logic to achieve real-time synchronization and coherence optimization of video display among multiple terminal devices.

Images