A data transmission method and system based on network adaptive switching

Abstract

The present invention belongs to the technical field of data transmission, and discloses a data transmission method and system based on network adaptive switching. The method comprises the following steps: a first data transmission device confirms a second data transmission device and configures a number of preset network transmission channels; dynamically allocates network resources according to the real-time network resource status and real-time data transmission requirements to obtain a real-time network resource dynamic allocation plan; executes the real-time network resource dynamic allocation plan, selects the optimal network transmission channel, and sends the real-time data to be transmitted to the second data transmission device based on the optimal network transmission channel; performs network quality analysis based on the real-time traffic monitoring data of the optimal network transmission channel, and performs network adaptive switching based on the obtained real-time network quality analysis results. The present invention solves the problems of rigid network selection mechanism, inflexible resource allocation, network switching delay and low data transmission reliability in the prior art.

Description

Technical Field

[0001] The present invention belongs to the technical field of data transmission, and in particular relates to a data transmission method and system based on network adaptive switching. Background Art

[0002] Data transmission technology is a fundamental component of the information technology field. Its background and development are closely tied to the advancements in computer technology, communications technology, and network technology. With the growth and widespread adoption of information technology, people's lives and work are becoming increasingly intertwined with data transmission technology, which bears a significant responsibility for information exchange in human society. Therefore, ensuring the reliability and stability of data transmission has become a key development goal in data transmission technology.

[0003] Existing data transmission technology has the following defects:

[0004] 1) Rigid network selection mechanism: Existing network selection technologies are often based on fixed rules or preset parameters, which cannot adapt well to dynamically changing network environments. Static network selection strategies cannot fully utilize resources in a multi-network environment, resulting in low transmission efficiency.

[0005] 2) Inflexible resource allocation: Existing technologies fail to consider real-time network status when allocating network resources, resulting in irrational resource allocation. Static resource allocation strategies cannot be dynamically adjusted according to actual needs, causing resource waste or shortage.

[0006] 3) Network handover delay: When network quality degrades, the existing network handover process may be delayed, affecting the user experience. The network handover mechanism is not intelligent enough to adapt to changes in network quality, resulting in untimely handover.

[0007] 4) Low data transmission reliability: In an unstable network environment, existing technologies cannot guarantee reliable data transmission, and data loss or errors are prone to occur. Summary of the Invention

[0008] In order to solve the problems of rigid network selection mechanism, inflexible resource allocation, network switching delay and low data transmission reliability in the prior art, the present invention aims to provide a data transmission method and system based on network adaptive switching.

[0009] The technical solution adopted in the present invention is:

[0010] A data transmission method based on network adaptive switching includes the following steps:

[0011] The first data transmission device as the sender identifies the second data transmission device as the receiver, and configures a plurality of preset network transmission channels between the first data transmission device and the second data transmission device;

[0012] The first data transmission device dynamically allocates network resources according to the real-time network resource status and real-time data transmission requirements of each preset network transmission channel to obtain a real-time network resource dynamic allocation plan;

[0013] The first data transmission device executes a real-time dynamic allocation scheme for network resources, selects an optimal network transmission channel, and sends the real-time data to be transmitted to the second data transmission device based on the optimal network transmission channel;

[0014] The first data transmission device performs network quality analysis based on the real-time traffic monitoring data of the optimal network transmission channel, and performs network adaptive switching based on the obtained real-time network quality analysis results.

[0015] Furthermore, the first data transmission device as the sender identifies the second data transmission device as the receiver, and configures a plurality of preset network transmission channels between the first data transmission device and the second data transmission device, including the following steps:

[0016] The first data transmission device as the sender confirms the second data transmission device as the receiver and obtains the second transmission IP address of the second data transmission device;

[0017] identifying a number of available alternative network transmission channels between a first transmission IP address of a first data transmission device and a second transmission IP address of a second data transmission device;

[0018] collecting first real-time transmission performance data of candidate network transmission channels, and sorting the candidate network transmission channels in descending order according to the first real-time transmission performance data to obtain a list of candidate network transmission channels;

[0019] The first several candidate network transmission channels in the candidate network transmission channel list are selected as preset network transmission channels, and a preset network transmission channel list including the plurality of preset network transmission channels is obtained.

[0020] Furthermore, the first data transmission device dynamically allocates network resources according to the real-time network resource status and real-time data transmission requirements of each preset network transmission channel to obtain a real-time network resource dynamic allocation plan, including the following steps:

[0021] The first data transmission device confirms the real-time data transmission demand and collects the real-time network resource status of each preset network transmission channel in the preset network transmission channel list;

[0022] According to a number of real-time network resource states and real-time data transmission requirements, a pre-trained network resource dynamic allocation model is updated to obtain an updated network resource dynamic allocation model;

[0023] An updated network resource dynamic allocation model is used to dynamically allocate network resources and obtain a real-time network resource dynamic allocation solution.

[0024] Furthermore, the network resource dynamic allocation model is constructed based on the MPO-DQN algorithm, and the network resource dynamic allocation model includes a meta-strategy optimization module constructed based on the MPO algorithm and a reinforcement learning module constructed based on the DQN algorithm. The reinforcement learning module is provided with an intelligent agent, a deep Q network, and an experience replay pool. The intelligent agent is connected to the deep Q network, the experience replay pool, and the meta-strategy optimization module respectively, and the meta-strategy optimization module is connected to the experience replay pool.

[0025] The training method of the network resource dynamic allocation model includes the following steps:

[0026] Collect some historical data transmission requirements and some historical network resource status, set the historical data transmission requirements as the meta-policy optimization scenario, and use the MPO algorithm to build the initial meta-policy optimization module;

[0027] In several scenarios, the initial meta-strategy optimization module is trained using several historical network resource states to obtain the final meta-strategy optimization module and generate several historical meta-strategy optimization experiences.

[0028] The dynamic allocation of network resources is used as the simulation environment for the DQN algorithm. The state space of the DQN algorithm is defined based on several historical network resource states, and the action space of the DQN algorithm is defined based on several preset network resource allocation actions.

[0029] Based on the impact of the preset network resource allocation action on the historical network resource status, the reward function of the DQN algorithm is defined. The final meta-strategy optimization module is used to generate the initial deep Q network. Based on the initial deep Q network, the initial reinforcement learning module is constructed.

[0030] Based on the state space, action space, and reward function, the initial reinforcement learning module is trained using several historical network resource states to obtain the final reinforcement learning module and generate several historical reinforcement learning experiences.

[0031] Several historical meta-strategy optimization experiences and several historical reinforcement learning experiences are stored in the experience replay pool of the final reinforcement learning module, and the final meta-strategy optimization module and the final reinforcement learning module are integrated to obtain a network resource dynamic allocation model.

[0032] Furthermore, according to a number of real-time network resource states and real-time data transmission requirements, the pre-trained network resource dynamic allocation model is updated to obtain an updated network resource dynamic allocation model, including the following steps:

[0033] According to the real-time data transmission requirements, search and match are performed in the experience replay pool to obtain the real-time meta-strategy optimization experience and real-time reinforcement learning experience of successful matches;

[0034] Based on the experience of real-time meta-strategy optimization, the meta-strategy optimization module is used to update the deep Q network of the pre-trained network resource dynamic allocation model to obtain an updated deep Q network;

[0035] Generate an updated action space based on real-time reinforcement learning experience, and generate an updated state space based on the status of several real-time network resources;

[0036] By integrating the updated deep Q network, the updated action space, and the updated state space, we obtain an updated network resource dynamic allocation model.

[0037] Furthermore, using the updated network resource dynamic allocation model, network resource dynamic allocation is performed to obtain a real-time network resource dynamic allocation solution, including the following steps:

[0038] Using an agent of the updated network resource dynamic allocation model, controlling an updated deep Q-network, and generating a Q-value for each possible real-time network resource allocation action in the updated action space for each real-time network resource state in the updated state space;

[0039] According to the reward function, obtain the reward value of each possible real-time network resource allocation action for each real-time network resource state, and update the corresponding Q value according to the reward value to obtain an updated Q value;

[0040] Repeat the above Q-value updating steps until the number of iterations reaches the iteration threshold, and obtain the final Q-value of each possible real-time network resource allocation action for each real-time network resource state;

[0041] Using a greedy strategy, the possible real-time network resource allocation action corresponding to the highest final Q value is selected as the execution real-time network resource allocation action corresponding to the real-time network resource state;

[0042] Integrate all real-time network resource states to execute real-time network resource allocation actions and obtain a real-time network resource dynamic allocation plan.

[0043] Furthermore, the first data transmission device executes a real-time network resource dynamic allocation scheme, selects an optimal network transmission channel, and sends the real-time data to be transmitted to the second data transmission device based on the optimal network transmission channel, including the following steps:

[0044] The first data transmission device allocates network resources to all preset network transmission channel lists in the preset network transmission channel list according to all execution real-time network resource allocation actions of the real-time network resource dynamic allocation scheme, thereby obtaining a plurality of running network transmission channels;

[0045] collecting second real-time transmission performance data of the operating network transmission channel, and selecting an optimal network transmission channel from among the plurality of operating network transmission channels based on the second real-time transmission performance data;

[0046] Confirm the real-time data to be transmitted, and send the real-time data to be transmitted to the second data transmission device based on the optimal network transmission channel.

[0047] Furthermore, the first data transmission device performs network quality analysis based on the real-time traffic monitoring data of the optimal network transmission channel, and performs network adaptive switching based on the obtained real-time network quality analysis results, including the following steps:

[0048] A first data transmission device collects real-time traffic monitoring data of the optimal network transmission channel and inputs the real-time traffic monitoring data into a pre-trained network quality analysis model;

[0049] Based on real-time traffic monitoring data, network quality analysis is performed using a network quality analysis model to obtain real-time network quality analysis results.

[0050] If the real-time network quality analysis result shows that the network quality is poor, the optimal network transmission channel is switched to the next preset network transmission channel in the preset network transmission channel list.

[0051] Furthermore, the network quality analysis model is constructed based on the RF-MLP algorithm, and the network quality analysis model includes a key feature screening module constructed based on the RF algorithm and a network quality analysis module constructed based on the MLP algorithm, which are connected in sequence.

[0052] A data transmission system based on network adaptive switching is used to implement a data transmission method. The system is arranged in a data transmission device and includes a network transmission channel configuration unit, a network resource dynamic allocation unit, a data transmission unit and a network adaptive switching unit connected in sequence.

[0053] The beneficial effects of the present invention are:

[0054] The present invention provides a data transmission method and system based on network adaptive switching, which can dynamically select the optimal network transmission channel according to the real-time network resource status and data transmission requirements, thereby improving the flexibility and efficiency of network selection; through real-time monitoring of the network status, dynamic allocation of network resources is achieved, resource utilization is optimized, and the overall performance of data transmission is improved; network adaptive switching is achieved, the delay of network switching is reduced, and it can quickly switch to a backup network channel when the network quality degrades, thereby ensuring the continuity and stability of data transmission; through intelligent network quality analysis, the reliability of data transmission is improved, data loss and errors are reduced, and data integrity is ensured.

[0055] Other beneficial effects of the present invention will be further described in the specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a flowchart of the data transmission method based on network adaptive switching in the present invention.

[0057] Figure 2 It is a structural block diagram of the data transmission system based on network adaptive switching in the present invention. DETAILED DESCRIPTION

[0058] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments.

[0059] Example 1:

[0060] like Figure 1 As shown, this embodiment provides a data transmission method based on network adaptive switching, including the following steps:

[0061] S1: A first data transmission device as a sender identifies a second data transmission device as a receiver and configures a plurality of preset network transmission channels between the first data transmission device and the second data transmission device, including the following steps:

[0062] S1-1: The first data transmission device as the sender identifies the second data transmission device as the receiver and obtains the second transmission IP address of the second data transmission device;

[0063] S1-2: Identify several available alternative network transmission channels between the first transmission IP address of the first data transmission device and the second transmission IP address of the second data transmission device;

[0064] Network transmission channels include Wi-Fi, 4G, 5G, wired network, Bluetooth, and Ethernet;

[0065] S1-3: Collecting first real-time transmission performance data of candidate network transmission channels, and sorting the candidate network transmission channels in descending order according to the first real-time transmission performance data to obtain a list of candidate network transmission channels;

[0066] Transmission performance data includes performance parameters such as bandwidth, latency, packet loss rate, signal strength, signal stability, and data transmission rate of the network transmission channel. These parameters are used to evaluate the reliability and stability of the network transmission channel for data transmission. These data are inherent and static characteristics of the network transmission channel, reflecting its transmission capacity and limitations.

[0067] S1-4: Selecting the first several candidate network transmission channels in the candidate network transmission channel list as preset network transmission channels, and obtaining a preset network transmission channel list including the plurality of preset network transmission channels;

[0068] S2: The first data transmission device dynamically allocates network resources according to the real-time network resource status and real-time data transmission requirements of each preset network transmission channel to obtain a real-time network resource dynamic allocation plan, including the following steps:

[0069] S2-1: The first data transmission device confirms the real-time data transmission demand and collects the real-time network resource status of each preset network transmission channel in the preset network transmission channel list;

[0070] Data transmission requirements include the size, type, format, and delay threshold of the data to be transmitted. Network resource status includes the data flow allocation status of the network transmission channel, transmission node status, network resource occupancy status, and remaining network resource status.

[0071] S2-2: updating the pre-trained network resource dynamic allocation model according to a number of real-time network resource states and real-time data transmission requirements to obtain an updated network resource dynamic allocation model;

[0072] The dynamic network resource allocation model is based on the Meta-Policy Optimization (MPO)-Deep Q Network (DQN) algorithm. It includes a meta-policy optimization module based on the MPO algorithm and a reinforcement learning module based on the DQN algorithm. The reinforcement learning module includes an intelligent agent, a deep Q network, and an experience replay pool. The intelligent agent is connected to the deep Q network, the experience replay pool, and the meta-policy optimization module, and the meta-policy optimization module is connected to the experience replay pool.

[0073] The meta-policy optimization module is used to optimize the parameters of the Deep Q Network in the reinforcement learning module so that these parameters can quickly adapt to new, unseen network resource states, improving the model's generalization ability. Even under unseen network resource states, the Deep Q Network can be updated based on previous learning experience. Through meta-policy optimization, the DQN algorithm can converge faster, reducing the amount of data and time required for training, significantly improving learning efficiency. The reinforcement learning module uses previous learning experience to output the appropriate network resource allocation action based on the real-time network resource state.

[0074] The training method of the network resource dynamic allocation model includes the following steps:

[0075] A-1: Collect some historical data transmission requirements and some historical network resource status, set the historical data transmission requirements as the meta-policy optimization scenario, and use the MPO algorithm to build the initial meta-policy optimization module;

[0076] A-2: Use several historical network resource states in several scenarios to train the initial meta-policy optimization module, obtain the final meta-policy optimization module, and generate several historical meta-policy optimization experiences;

[0077] A-3: Use the dynamic network resource allocation problem as the simulation environment for the DQN algorithm. Define the state space of the DQN algorithm based on several historical network resource states. Define the action space of the DQN algorithm based on several preset network resource allocation actions.

[0078] A-4: Define the reward function of the DQN algorithm based on the impact of the preset network resource allocation action on the historical network resource status. Use the final meta-policy optimization module to generate the initial deep Q network. Based on the initial deep Q network, build the initial reinforcement learning module.

[0079] A-5: Based on the state space, action space, and reward function, use several historical network resource states to train the initial reinforcement learning module to obtain the final reinforcement learning module and generate several historical reinforcement learning experiences;

[0080] A-6: Store several historical meta-strategy optimization experiences and several historical reinforcement learning experiences in the experience replay pool of the final reinforcement learning module, and integrate the final meta-strategy optimization module and the final reinforcement learning module to obtain a dynamic network resource allocation model;

[0081] Based on a number of real-time network resource states and real-time data transmission requirements, a pre-trained network resource dynamic allocation model is updated to obtain an updated network resource dynamic allocation model, including the following steps:

[0082] S2-2-1: Based on the real-time data transmission requirements, search and match in the experience replay pool to obtain the real-time meta-strategy optimization experience and real-time reinforcement learning experience that are successfully matched;

[0083] S2-2-2: Based on the real-time meta-strategy optimization experience, use the meta-strategy optimization module to update the pre-trained deep Q network of the network resource dynamic allocation model to obtain an updated deep Q network;

[0084] S2-2-3: Generate an updated action space based on real-time reinforcement learning experience ,in, For the updated Action value, is the action indicator, is the total number of action space dimensions, and generates an updated state space based on several real-time network resource states ,in, For the updated Status value, is the status indicator, is the total number of state space dimensions;

[0085] S2-2-4: Integrate the updated deep Q network, the updated action space, and the updated state space to obtain an updated network resource dynamic allocation model;

[0086] S2-3: Using the updated network resource dynamic allocation model, dynamically allocate network resources to obtain a real-time network resource dynamic allocation solution, including the following steps:

[0087] S2-3-1: Use the agent of the updated network resource dynamic allocation model to control the updated deep Q network to generate the Q value of each possible real-time network resource allocation action in the updated action space for each real-time network resource state in the updated state space;

[0088] S2-3-2: According to the reward function, obtain the reward value of each possible real-time network resource allocation action for each real-time network resource state, and update the corresponding Q value according to the reward value to obtain an updated Q value;

[0089] The formula is:

[0090]

[0091] Where, The updated status value and updated action value The corresponding updated Q value; Status value and action value The corresponding Q value; is the learning rate; is the highest predicted Q value; It is a comprehensive indicator; To update the parameters; To use the updated action value The status value Transition to the updated state value The reward value;

[0092] S2-3-3: Repeat the above Q-value updating steps until the number of iterations reaches the iteration threshold, and obtain the final Q-value of each possible real-time network resource allocation action for each real-time network resource state;

[0093] S2-3-4: Using a greedy strategy, select the possible real-time network resource allocation action corresponding to the highest final Q value as the execution real-time network resource allocation action corresponding to the real-time network resource state;

[0094] S2-3-5: Integrate all real-time network resource states to execute real-time network resource allocation actions and obtain a real-time network resource dynamic allocation plan;

[0095] The dynamic allocation scheme of network resources includes data flow allocation adjustment actions, transmission node replacement actions, and network resource allocation actions;

[0096] S3: The first data transmission device executes a real-time network resource dynamic allocation scheme, selects an optimal network transmission channel, and sends the real-time data to be transmitted to the second data transmission device based on the optimal network transmission channel, including the following steps:

[0097] S3-1: The first data transmission device allocates network resources to all preset network transmission channels in the preset network transmission channel list according to all execution real-time network resource allocation actions of the real-time network resource dynamic allocation scheme, and obtains a plurality of running network transmission channels;

[0098] S3-2: collecting second real-time transmission performance data of the operating network transmission channel, and selecting an optimal network transmission channel from among the plurality of operating network transmission channels based on the second real-time transmission performance data;

[0099] S3-3: confirming the real-time data to be transmitted, and sending the real-time data to be transmitted to the second data transmission device based on the optimal network transmission channel;

[0100] S4: The first data transmission device performs network quality analysis based on the real-time traffic monitoring data of the optimal network transmission channel, and performs network adaptive switching based on the obtained real-time network quality analysis results, including the following steps:

[0101] S4-1: The first data transmission device collects real-time traffic monitoring data of the optimal network transmission channel and inputs the real-time traffic monitoring data into a pre-trained network quality analysis model;

[0102] The network quality analysis model is built based on the Random Forest (RF)-Attention-Multilayer Perceptron (MLP) algorithm, and includes a key feature screening module based on the RF algorithm and a network quality analysis module based on the MLP algorithm, which are connected in sequence.

[0103] The key feature screening module uses an internal Classification and Regression Tree (CART) to screen the feature components of the input traffic monitoring data. It can process a large number of feature components, generate a key feature importance score for each feature component, and select the most stable and discriminative key feature components based on the key feature importance score. After the key feature screening module has been trained, it can directly screen the newly input traffic monitoring data based on the selected key features to obtain the corresponding key features. The MLP network, as a fully connected network, can accurately and efficiently predict labels based on the key features.

[0104] The training method of the network quality analysis model includes the following steps:

[0105] B-1: Use the RF-MLP algorithm to build an initial network quality analysis model; the initial network quality analysis model includes an initial key feature screening module and an initial network quality analysis module;

[0106] B-2: Collect a number of historical traffic monitoring data and train the initial key feature screening module based on the historical traffic monitoring data to obtain a number of key feature indicators, a number of historical key features of each historical traffic monitoring data, and a final key feature screening module, including the following steps:

[0107] B-2-1: Collect a number of historical flow monitoring data, and pre-process the historical flow monitoring data to obtain a number of pre-processed historical flow monitoring data;

[0108] B-2-2: Use the RF structure in the key feature screening module to extract the feature contribution of M candidate features in the pre-processed historical traffic monitoring data;

[0109] The formula is:

[0110]

[0111] Where, For the Feature contribution of candidate features; For the The candidate features are in the random forest The feature contribution of each tree; is the CART tree indicator; is an alternative characteristic indicator; is the total number of CARTs;

[0112]

[0113] Where, CART tree node of random forest m ,node and nodes The Gini index; CART tree node m Medium Category the proportion; is the total number of categories; m 、 、 is the node indicator; is the category indicator;

[0114] B-2-3: Normalize the feature contributions of the M candidate features to obtain corresponding normalized feature contributions;

[0115] The formula is:

[0116]

[0117] Where, is the contribution of the feature after normalization; J is the total number of candidate features;

[0118] B-2-4: Generate feature selection standard values ​​for several candidate features based on the normalized feature contribution;

[0119] The formula is:

[0120]

[0121] Where, For the Feature selection criterion values ​​for candidate features; For the The contribution of the candidate features after normalization; is an alternative characteristic indicator;

[0122] B-2-5: Sort the candidate features in descending order according to the feature selection standard value, and select the feature indices corresponding to the first M candidate features as key feature indices to obtain M key feature indices;

[0123] B-2-6: Based on the M key feature indicators, perform key feature screening on each pre-processed historical traffic monitoring data to obtain several corresponding historical key features;

[0124] B-2-6: Traverse all pre-processed historical traffic monitoring data to obtain the final key feature screening module;

[0125] B-3: Based on several historical key features of all pre-processed historical traffic monitoring data, the initial network quality analysis module is trained to obtain the final network quality analysis module;

[0126] B-4: Integrate the final key feature screening module and the final network quality analysis module to obtain the final network quality analysis model;

[0127] S4-2: Based on the real-time traffic monitoring data, use the network quality analysis model to perform network quality analysis and obtain real-time network quality analysis results, including the following steps:

[0128] S4-2-1: Input real-time traffic monitoring data into the network quality analysis model;

[0129] Traffic monitoring data includes traffic packet data of the optimal network transmission channel, traffic transmission frequency data, and transmission node distribution traffic data. Traffic monitoring data is a dynamic characteristic of the network transmission channel. Compared with transmission performance data, traffic monitoring data reflects the real-time status of the network transmission channel.

[0130] S4-2-2: Based on the key feature indicators obtained through training, a key feature screening module is used to extract several real-time key features of the real-time traffic monitoring data;

[0131] Key characteristic indicators include traffic packet characteristics, transmission frequency characteristics, distribution traffic address characteristics, traffic source characteristics, etc.

[0132] S4-2-3: Based on several real-time key features, use the network quality analysis module to perform network quality analysis and obtain real-time network quality analysis results;

[0133] Network quality analysis results include poor network quality, good network quality, excellent network quality, the presence of network attacks, and network attack types;

[0134] S4-3: If the real-time network quality analysis result shows that the network quality is poor, the optimal network transmission channel is switched to the next preset network transmission channel in the preset network transmission channel list.

[0135] Example 2:

[0136] like Figure 2 As shown, this embodiment provides a data transmission system based on network adaptive switching, which is used to implement a data transmission method. The system is set in a data transmission device, and the system includes a network transmission channel configuration unit, a network resource dynamic allocation unit, a data transmission unit, and a network adaptive switching unit connected in sequence;

[0137] A network transmission channel configuration unit is configured to, at a first data transmission device as a sender, identify a second data transmission device as a receiver, and configure a plurality of preset network transmission channels between the first data transmission device and the second data transmission device;

[0138] A network resource dynamic allocation unit is used to dynamically allocate network resources according to the real-time network resource status and real-time data transmission requirements of each preset network transmission channel to obtain a real-time network resource dynamic allocation plan;

[0139] a data transmission unit, configured to execute a real-time dynamic allocation scheme for network resources, select an optimal network transmission channel, and send the real-time data to be transmitted to a second data transmission device based on the optimal network transmission channel;

[0140] The network adaptive switching unit is used to perform network quality analysis based on the real-time traffic monitoring data of the optimal network transmission channel, and perform network adaptive switching based on the obtained real-time network quality analysis results.

[0141] The present invention provides a data transmission method and system based on network adaptive switching, which can dynamically select the optimal network transmission channel according to the real-time network resource status and data transmission requirements, thereby improving the flexibility and efficiency of network selection; through real-time monitoring of the network status, dynamic allocation of network resources is achieved, resource utilization is optimized, and the overall performance of data transmission is improved; network adaptive switching is achieved, the delay of network switching is reduced, and it can quickly switch to a backup network channel when the network quality degrades, thereby ensuring the continuity and stability of data transmission; through intelligent network quality analysis, the reliability of data transmission is improved, data loss and errors are reduced, and data integrity is ensured.

[0142] The present invention is not limited to the above optional embodiments. Anyone can derive various other forms of products based on the teachings of the present invention. The above specific embodiments should not be construed as limiting the scope of protection of the present invention. The scope of protection of the present invention shall be based on the scope defined in the claims, and the description can be used to interpret the claims.

Claims

1. A data transmission method based on network adaptive switching, characterized by: The steps include: The first data transmission device as the sender identifies the second data transmission device as the receiver, and configures a plurality of preset network transmission channels between the first data transmission device and the second data transmission device; The first data transmission device dynamically allocates network resources according to the real-time network resource status and real-time data transmission requirements of each preset network transmission channel to obtain a real-time network resource dynamic allocation plan, including the following steps: The first data transmission device confirms the real-time data transmission demand and collects the real-time network resource status of each preset network transmission channel in the preset network transmission channel list; According to a number of real-time network resource states and real-time data transmission requirements, a pre-trained network resource dynamic allocation model is updated to obtain an updated network resource dynamic allocation model; The network resource dynamic allocation model is constructed based on the MPO-DQN algorithm, and the network resource dynamic allocation model includes a meta-strategy optimization module constructed based on the MPO algorithm and a reinforcement learning module constructed based on the DQN algorithm. The reinforcement learning module is provided with an intelligent agent, a deep Q network, and an experience replay pool. The intelligent agent is respectively connected to the deep Q network, the experience replay pool, and the meta-strategy optimization module, and the meta-strategy optimization module is connected to the experience replay pool. The training method of the network resource dynamic allocation model includes the following steps: Collect some historical data transmission requirements and some historical network resource status, set the historical data transmission requirements as the meta-policy optimization scenario, and use the MPO algorithm to build the initial meta-policy optimization module; In several scenarios, the initial meta-strategy optimization module is trained using several historical network resource states to obtain the final meta-strategy optimization module and generate several historical meta-strategy optimization experiences. The dynamic allocation of network resources is used as the simulation environment for the DQN algorithm. The state space of the DQN algorithm is defined based on several historical network resource states, and the action space of the DQN algorithm is defined based on several preset network resource allocation actions. Based on the impact of the preset network resource allocation action on the historical network resource status, the reward function of the DQN algorithm is defined. The final meta-strategy optimization module is used to generate the initial deep Q network. Based on the initial deep Q network, the initial reinforcement learning module is constructed. Based on the state space, action space, and reward function, the initial reinforcement learning module is trained using several historical network resource states to obtain the final reinforcement learning module and generate several historical reinforcement learning experiences. Store several historical meta-strategy optimization experiences and several historical reinforcement learning experiences in the experience replay pool of the final reinforcement learning module, and integrate the final meta-strategy optimization module and the final reinforcement learning module to obtain a network resource dynamic allocation model; Based on a number of real-time network resource states and real-time data transmission requirements, a pre-trained network resource dynamic allocation model is updated to obtain an updated network resource dynamic allocation model, including the following steps: According to the real-time data transmission requirements, search and match are performed in the experience replay pool to obtain the real-time meta-strategy optimization experience and real-time reinforcement learning experience of successful matches; Based on the experience of real-time meta-strategy optimization, the meta-strategy optimization module is used to update the deep Q network of the pre-trained network resource dynamic allocation model to obtain an updated deep Q network; Generate an updated action space based on real-time reinforcement learning experience, and generate an updated state space based on the status of several real-time network resources; Integrate the updated deep Q network, the updated action space, and the updated state space to obtain an updated network resource dynamic allocation model; Using the updated network resource dynamic allocation model, dynamic allocation of network resources is performed to obtain a real-time network resource dynamic allocation solution, including the following steps: Using an agent of the updated network resource dynamic allocation model, controlling an updated deep Q-network, and generating a Q-value for each possible real-time network resource allocation action in the updated action space for each real-time network resource state in the updated state space; According to the reward function, obtain the reward value of each possible real-time network resource allocation action for each real-time network resource state, and update the corresponding Q value according to the reward value to obtain an updated Q value; Repeat the above Q-value updating steps until the number of iterations reaches the iteration threshold, and obtain the final Q-value of each possible real-time network resource allocation action for each real-time network resource state; Using a greedy strategy, the possible real-time network resource allocation action corresponding to the highest final Q value is selected as the execution real-time network resource allocation action corresponding to the real-time network resource state; Integrate all real-time network resource status to execute real-time network resource allocation actions and obtain real-time network resource dynamic allocation plan; The first data transmission device executes a real-time dynamic allocation scheme for network resources, selects an optimal network transmission channel, and sends the real-time data to be transmitted to the second data transmission device based on the optimal network transmission channel; The first data transmission device performs network quality analysis based on the real-time traffic monitoring data of the optimal network transmission channel, and performs network adaptive switching based on the obtained real-time network quality analysis results.

2. The data transmission method based on network adaptive switching according to claim 1, characterized in that: A first data transmission device as a sender identifies a second data transmission device as a receiver, and configures a plurality of preset network transmission channels between the first data transmission device and the second data transmission device, including the following steps: The first data transmission device as the sender confirms the second data transmission device as the receiver and obtains the second transmission IP address of the second data transmission device; identifying a number of available alternative network transmission channels between a first transmission IP address of a first data transmission device and a second transmission IP address of a second data transmission device; collecting first real-time transmission performance data of candidate network transmission channels, and sorting the candidate network transmission channels in descending order according to the first real-time transmission performance data to obtain a list of candidate network transmission channels; The first several candidate network transmission channels in the candidate network transmission channel list are selected as preset network transmission channels, and a preset network transmission channel list including the plurality of preset network transmission channels is obtained.

3. The data transmission method based on network adaptive switching according to claim 2, characterized in that: The first data transmission device executes a real-time network resource dynamic allocation scheme, selects an optimal network transmission channel, and sends real-time data to be transmitted to a second data transmission device based on the optimal network transmission channel, including the following steps: The first data transmission device allocates network resources to all preset network transmission channel lists in the preset network transmission channel list according to all execution real-time network resource allocation actions of the real-time network resource dynamic allocation scheme, thereby obtaining a plurality of running network transmission channels; collecting second real-time transmission performance data of the operating network transmission channel, and selecting an optimal network transmission channel from among the plurality of operating network transmission channels based on the second real-time transmission performance data; Confirm the real-time data to be transmitted, and send the real-time data to be transmitted to the second data transmission device based on the optimal network transmission channel.

4. The data transmission method based on network adaptive switching according to claim 3, characterized in that: The first data transmission device performs network quality analysis based on real-time traffic monitoring data of the optimal network transmission channel, and performs network adaptive switching based on the obtained real-time network quality analysis results, including the following steps: A first data transmission device collects real-time traffic monitoring data of the optimal network transmission channel and inputs the real-time traffic monitoring data into a pre-trained network quality analysis model; Based on real-time traffic monitoring data, network quality analysis is performed using a network quality analysis model to obtain real-time network quality analysis results. If the real-time network quality analysis result shows that the network quality is poor, the optimal network transmission channel is switched to the next preset network transmission channel in the preset network transmission channel list.

5. The data transmission method based on network adaptive switching according to claim 4, characterized in that: The network quality analysis model is constructed based on the RF-MLP algorithm, and the network quality analysis model includes a key feature screening module constructed based on the RF algorithm and a network quality analysis module constructed based on the MLP algorithm, which are connected in sequence.

6. A data transmission system based on network adaptive switching, used to implement the data transmission method according to any one of claims 1 to 5, characterized in that: The system is arranged in a data transmission device, and the system comprises a network transmission channel configuration unit, a network resource dynamic allocation unit, a data transmission unit and a network adaptive switching unit connected in sequence; A network transmission channel configuration unit is configured to, at a first data transmission device as a sender, identify a second data transmission device as a receiver, and configure a plurality of preset network transmission channels between the first data transmission device and the second data transmission device; A network resource dynamic allocation unit is used to dynamically allocate network resources according to the real-time network resource status and real-time data transmission requirements of each preset network transmission channel to obtain a real-time network resource dynamic allocation plan; a data transmission unit, configured to execute a real-time dynamic allocation scheme for network resources, select an optimal network transmission channel, and send the real-time data to be transmitted to a second data transmission device based on the optimal network transmission channel; The network adaptive switching unit is used to perform network quality analysis based on the real-time traffic monitoring data of the optimal network transmission channel, and perform network adaptive switching based on the obtained real-time network quality analysis results.

Images