Network communication dynamic optimization method based on multi-module cooperation

Abstract

The application discloses a network communication dynamic optimization method based on multi-module cooperation, relates to the technical field of network communication, and deploys a dual-mode communication module at each network node.The network communication dynamic optimization method comprises the following steps: each network node broadcasts its own existence information and power line channel characteristics through an HPLC channel of the dual-mode communication module; each network node scans a surrounding wireless network through an HRF channel of the dual-mode communication module, and reports its own wireless channel quality information to a gateway; and a global state perception module collects all information, and constructs a global network view comprising a physical topology and a channel quality atlas.The dual-mode cooperation mechanism, the machine learning prediction model and the dynamic decision strategy of the application can adapt to complex dynamic environments such as power line noise fluctuation and wireless interference change, and can autonomously optimize communication parameters without manual intervention; meanwhile, the modular design facilitates extension to a multi-mode communication scene, and has a wide application prospect.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of network communication, in particular to a network communication dynamic optimization method based on multi-module cooperation. BACKGROUND

[0002] With the rapid development of the Internet of Things, smart grid, industrial automation and other fields, the reliability, real-time performance and quality of service (QoS) requirements of network communication are increasingly stringent. In existing network communication systems, single communication mode has significant limitations, such as relying only on power line carrier or only on wireless radio frequency;

[0003] High-speed power line carrier (HPLC) communication relies on power line transmission, which is easily affected by power line noise such as impulse noise, background noise, line attenuation, load change, etc., resulting in poor channel stability and transmission rate fluctuation; high-speed radio frequency (HRF) communication is easily affected by wireless environmental interference, obstacle shielding, etc., and the channel quality is easily affected by environmental dynamic changes, making it difficult to guarantee reliability.

[0004] Existing communication strategies are mostly designed based on static channel characteristics, lack global awareness of real-time channel state, and cannot respond to sudden changes in channel quality, such as power line burst noise and wireless signal shielding, resulting in rigid path selection and difficulty in dynamically adjusting to avoid channel degradation areas.

[0005] For high-priority services such as protection signals in smart grids and real-time instructions in industrial control, the requirements for delay, throughput and packet loss rate are stringent, but the bandwidth and stability of a single channel often cannot meet such demands; at the same time, existing technologies lack multi-channel cooperation mechanisms, making it difficult to improve QoS guarantee levels through multi-path concurrent transmission.

[0006] Some systems have deployed multi-mode communication modules, but lack unified state awareness and strategy decision mechanisms between modules, making it difficult to achieve global optimization of channel resources, making it difficult to take advantage of multi-mode, and even causing problems such as resource competition and transmission conflicts between modules.

[0007] Therefore, there is an urgent need for a method that can integrate the advantages of multi-mode communication modules, achieve global channel state awareness, and dynamically optimize communication strategies to improve the reliability, adaptability and QoS guarantee capability of network communication. SUMMARY

[0008] To solve the above technical problems, the present application provides a network communication dynamic optimization method based on multi-module cooperation. The technical scheme adopted is as follows:

[0009] The network communication dynamic optimization method based on multi-module cooperation deploys dual-mode communication modules at each network node, and the network communication dynamic optimization method comprises the following steps:

[0010] Step 1, each network node broadcasts its presence information and basic power line channel characteristics through the HPLC channel of the dual-mode communication module;

[0011] Step 2, each network node scans the surrounding wireless network through the HRF channel of the dual-mode communication module, and reports its wireless channel quality information to the gateway;

[0012] Step 3, the global state perception module collects all information and constructs a global network view containing physical topology and channel quality map;

[0013] Step 4, the intelligent strategy decision module receives the global network view from the global state perception module, learns the historical channel change rule using machine learning algorithm, predicts the trend of channel quality in the future period of time, generates dynamic path selection strategy, and generates dual-path concurrent communication selection strategy when judging that the communication content meets the preset stringent service quality QoS threshold;

[0014] Step 5, the dual-mode communication module receives the dynamic path selection strategy or the dual-path concurrent communication selection strategy, configures the working mode of the network node demodulator, and completes the channel switching.

[0015] Optionally, in step 1, after the network node device is powered on and initialized, the HPLC modem in the dual-mode communication module sends periodic beacon frames on the pre-defined broadcast channel or the default channel;

[0016] The beacon frame content includes device identifier, network identifier, supported maximum data rate, modulation mode, channel noise floor based on preliminary listening, and its own transmission power.

[0017] Optionally, after the gateway receives the beacon frame of the network node, it initiates an active probe process, and the gateway sends a channel probe request frame to the network node, and the network node replies with a channel probe response frame after receiving the request;

[0018] The response frame content includes channel attenuation, signal-to-noise ratio, impulse noise statistics, and channel transfer function.

[0019] Optionally, the channel attenuation is obtained by calculating the difference between the received signal strength indication and the known transmission power; the signal-to-noise ratio is obtained by calculating the average SNR channel attenuation on multiple subcarriers; the impulse noise statistics is the number, average amplitude and duration of the detected impulse noise in a set period of time; and the channel transfer function is the data representing the channel frequency response.

[0020] Optionally, in step 2, the HRF module in the dual-mode communication module periodically scans all supported frequency bands and channels, including listening to beacon frames from the gateway and neighboring network nodes; recording the received signal strength indication of all detectable signals, analyzing channel occupancy, and calculating channel utilization; the network node packages the scanning results into an HRF environment report and sends it to the gateway through the HRF channel just scanned.

[0021] Optionally, the HRF environment report content includes the HRF interface MAC address of the node, the BSSID address, channel, RSSI value for each scanned network node, the average interference noise level of the main channel where the node is located, and the estimated packet error rate.

[0022] Optionally, in step 3, the global network view construction includes the following steps:

[0023] Step 31, the global state awareness module receives and processes all HPLC channel characteristics and HRF environment reports from the nodes;

[0024] Step 32, parse the device ID and neighbor information in the report, and construct a bidirectional network topology graph with devices as vertices and communication links as edges;

[0025] For HPLC: calculate the communication capability index for the power line link between any two nodes, which is a function of signal-to-noise ratio, attenuation, and noise level;

[0026] For HRF: calculate the stability score HRF_S for the wireless link based on RSSI and channel utilization. HRF_S = g(RSSI, Channel_U);

[0027] g is the stability score function, and Channel_U is the channel utilization.

[0028] Optionally, step 4 includes the following sub-steps:

[0029] Step 41, the intelligent strategy decision module extracts historical time series data from the network state database as input features for the machine learning model;

[0030] Step 42, the machine learning model predicts the predicted values of the HPLC communication capability index and the HRF stability score of each link in the future time window;

[0031] Step 43, based on the current state and the predicted state, calculate the cost of all possible paths for each data flow;

[0032] Step 44, select the path with the lowest cost as the result of the dynamic path selection strategy.

[0033] Optionally, the formula for calculating the cost in step 43 is:

[0034]

[0035] wherein is the cost value, is the total end-to-end time experienced by the data packet from the original network node to the destination network node, is the amount of valid user data successfully transmitted per unit time through the path, is the ratio of the number of lost data packets to the total number of transmitted data packets during transmission, are weight coefficients, respectively, for adjusting the importance of delay, throughput and packet loss rate in the total cost calculation.

[0036] Optionally, in step 4, the quality of service QoS requirement of the service data stream is obtained; the real-time state information of the high-speed power line carrier HPLC channel and the high-speed radio frequency HRF channel is obtained; when the QoS requirement exceeds a preset threshold, the following steps are executed:

[0037] The data packets of the service data stream are divided into N data fragments; based on the real-time state information, a fragment allocation strategy is generated, which specifies for each data fragment to be transmitted through the HPLC channel or the HRF channel; the HPLC communication module and the HRF communication module are controlled to concurrently execute the transmission of the N data fragments; at the receiving end, the data fragments from the HPLC and HRF channels are received and recombined into the complete data packet.

[0038] In summary, the present application includes at least one of the following beneficial technical effects:

[0039] The present application can provide a network communication dynamic optimization method based on multi-module cooperation, through the cooperative design of HPLC and HRF dual-mode communication modules, the complementary characteristics of the two channels are utilized, avoiding the inherent defects of single communication mode. When the quality of a channel deteriorates, it can quickly switch to another channel through dynamic path selection, or realize data redundancy through dual-path concurrent transmission, significantly reducing the risk of transmission interruption caused by sudden channel degradation.

[0040] The global state perception module collects the HPLC channel characteristics and HRF environment information of each node to construct a global network view containing physical topology and channel quality map, overcoming the limitations of local perception in the prior art. Combined with machine learning algorithm to learn historical channel variation law and predict future trend, it can identify the trend of channel quality change in advance, making the communication strategy adjustment more forward-looking and effectively avoiding potential channel degradation risk.

[0041] ​The intelligent strategy decision module calculates the optimal path through a cost function based on real-time and predicted channel states, and realizes dynamic optimization of the path; for services meeting preset stringent QoS thresholds, data packets are allocated to the HPLC and HRF channels for parallel transmission through a data fragmentation and dual-path concurrent transmission strategy, the bandwidth superposition effect of the dual channels is utilized to improve the throughput, and the end-to-end delay is reduced through fragmentation and recombination, so that the QoS requirements of high-priority services are met.

[0042] The global network view supports unified scheduling of HPLC and HRF channel resources, and the dynamic path selection strategy can allocate communication paths in real time according to channel quality, avoiding idle or excessive competition for resources; at the same time, the dual-path concurrent strategy fully utilizes the hardware resources of the dual-mode module while meeting high QoS requirements, improving the overall resource utilization and data transmission efficiency of the network.

[0043] The dual-mode cooperative mechanism, machine learning prediction model and dynamic decision strategy of the application can adapt to complex dynamic environments such as power line noise fluctuations and wireless interference changes, and can autonomously optimize communication parameters without human intervention; at the same time, the modular design facilitates expansion to multi-mode communication scenarios, and has wide application prospects. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 is a flowchart of the network communication dynamic optimization method based on multi-module cooperation of the application. DETAILED DESCRIPTION

[0045] The application will be further described in detail below with reference to the accompanying drawings.

[0046] The application embodiment discloses a network communication dynamic optimization method based on multi-module cooperation.

[0047] Reference Figure 1 , embodiment 1, the network communication dynamic optimization method based on multi-module cooperation, a dual-mode communication module is deployed at each network node, and the network communication dynamic optimization method comprises the following steps:

[0048] Step 1: Each network node broadcasts its existence information and basic power line channel characteristics through the HPLC channel of the dual-mode communication module;

[0049] Step 2: Each network node scans the surrounding wireless network through the HRF channel of the dual-mode communication module, and reports its wireless channel quality information to the gateway;

[0050] Step 3: The global state perception module collects all information and constructs a global network view containing physical topology and channel quality map;

[0051] Step 4, the intelligent policy decision module receives the global network view of the global state awareness module, learns the historical channel change rule by using a machine learning algorithm, predicts the trend of the channel quality in the future period of time, generates a dynamic path selection policy, and generates a dual-path concurrent communication selection policy when judging that the communication content meets the preset stringent quality of service (QoS) threshold.

[0052] Step 5, the dual-mode communication module receives the dynamic path selection policy or the dual-path concurrent communication selection policy, configures the working mode of the modulator of the network node, and completes channel switching.

[0053] In step 1 of embodiment 2, after the network node device is powered on and initialized, the HPLC modem in the dual-mode communication module sends a periodic beacon frame on a predefined broadcast channel or a default channel.

[0054] The beacon frame content includes a device identifier, a network identifier, a supported maximum data rate, a modulation mode, a channel noise floor based on preliminary listening, and a self-transmission power.

[0055] In embodiment 3, after the gateway receives the beacon frame of the network node, an active probe process is initiated. The gateway sends a channel probe request frame to the network node, and the network node replies with a channel probe response frame after receiving the request.

[0056] The response frame content includes channel attenuation, signal-to-noise ratio (SNR), impulse noise statistics, and channel transmission function.

[0057] In embodiment 4, the channel attenuation is obtained by calculating the difference between the received signal strength indication and the known transmission power; the SNR is obtained by calculating the average SNR channel attenuation on multiple subcarriers; the impulse noise statistics are the number, average amplitude, and duration of the detected impulse noise in a set time period; and the channel transmission function is data representing the channel frequency response.

[0058] By adopting the above technical solution, after the network node is powered on and initialized, its HPLC modem periodically sends a beacon frame on a predefined or default channel. The frame contains a device identifier, a network identifier, a supported maximum data rate, a modulation mode, a channel noise floor based on preliminary listening, and a self-transmission power.

[0059] The periodic design of the beacon frame ensures that new nodes and fault recovery nodes in the network can be quickly discovered; the predefined channel avoids channel conflicts between nodes and ensures the stability of information interaction.

[0060] Automatic node discovery and network initialization are achieved: through the device ID and network ID in the beacon frame, the gateway and surrounding nodes can quickly identify the identity of the newly connected node and the network to which it belongs, providing initial basis for dynamic updating of the network topology.

[0061] Provide basic communication capability reference: the parameters of maximum data rate, modulation mode, etc. directly reflect the hardware performance and communication potential of the node, helping the gateway to preliminarily judge whether the node can meet the transmission requirements of a specific service, such as matching a high data rate node for a high bandwidth service.

[0062] Establish channel benchmark data: noise floor and transmit power provide the original benchmark for subsequent calculation of key indicators such as channel attenuation and signal-to-noise ratio. For example, the calculation of channel attenuation = received signal strength - transmit power in embodiment 4 relies on this and is the starting point for channel quality assessment.

[0063] After receiving the beacon frame of the node, the gateway initiates an active probing process: sends a channel probing request frame to the node, triggers the node to reply a channel probing response frame; the response frame contains channel attenuation, signal-to-noise ratio, impulse noise statistics, and channel transmission function.

[0064] This process is a supplement to the passive broadcast of the beacon frame: the beacon frame only provides preliminary information, and the active probing interacts through the gateway request to the network node response, forcing to obtain more in-depth channel characteristics, solving the problem of missing key information that may be missed by the passive broadcast of the node.

[0065] Improve the depth and initiative of channel perception: the beacon frame only contains the basic information of preliminary listening, while the content of the response frame is the fine measurement result of the node on the channel, greatly improving the comprehensiveness of the channel state description, from basic capability to real-time interference and frequency response.

[0066] Realize the overall control of the gateway to the channel: through active probing, the gateway can obtain the channel characteristics of specific nodes as needed, such as high-frequency probing of nodes with large communication quality fluctuations, avoiding the lag of relying on node self-reporting and enhancing the real-time control ability of the channel state.

[0067] Standardize the calculation method of key parameters in the response frame:

[0068] Channel attenuation: calculated by the difference between received signal strength indication (RSSI) and known transmit power, i.e. the loss amount of the signal from transmission to reception;

[0069] Signal-to-noise ratio (SNR): calculate the average SNR on multiple subcarriers, which more truly reflects the overall channel's anti-noise capability;

[0070] Impulse noise statistics: quantify the number of impulse noises (i.e. interference frequency), average amplitude (i.e. interference intensity), and duration (i.e. interference duration) within a certain period of time, and specifically depict the common bursty impulse interference in power line channels;

[0071] Channel transfer function: data form to characterize the channel attenuation / gain characteristics of different frequency signals (i.e. frequency response), reflecting the transmission impact of the channel on the multi-carrier signal.

[0072] Ensure the uniformity and comparability of parameters: standardized calculation method avoids parameter deviation caused by different nodes due to measurement logic difference, so that the gateway can compare the channel characteristics of all nodes horizontally, such as the signal-to-noise ratio of node A and node B, which can be directly compared, providing a unified data basis for the calculation of communication capability index in the global network view.

[0073] Multi-subcarrier average SNR overcomes the contingency of single subcarrier and is closer to the actual communication scenario, because HPLC usually uses multi-carrier modulation;

[0074] Multi-dimensional statistics of impulse noise accurately capture the typical interference characteristics of power line channel, providing basis for subsequent anti-interference strategy;

[0075] Channel transfer function provides fine guidance in frequency domain for the working mode configuration of node modem, improving the stability of signal transmission.

[0076] From passive sensing to active detection, it realizes the upgrade from preliminary description to fine characterization of channel state; standardized parameters and calculation methods ensure the reliability and comparability of data, providing high-quality input for the global network view of physical topology + channel quality map constructed by the global state perception module;

[0077] Finally, it provides accurate channel quality basis for path selection and dual-path concurrent strategy generation of intelligent strategy decision module, ensuring the scientificity and effectiveness of network communication dynamic optimization from the bottom.

[0078] In example 5, step 2, the HRF module in the dual-mode communication module periodically scans all supported frequency bands and channels, including listening to beacon frames from the gateway and neighboring network nodes; records the received signal strength indication of all detectable signals, analyzes the channel occupation, and calculates the channel utilization; the network node packages the scanning results into an HRF environment report and sends it to the gateway through the HRF channel just scanned.

[0079] In example 6, the HRF environment report includes the HRF interface MAC address of the node, the BSSID address, channel, RSSI value of each scanned network node, the average interference noise level of the main channel where the node is located, and the estimated packet error rate.

[0080] By adopting the above technical solution, the HRF module in the dual-mode communication module performs scanning on all frequency bands and channels supported by it at a fixed period: the scanning content includes actively listening to beacon frames from the gateway and neighbor network nodes for identifying devices that can communicate in the periphery;

[0081] Record the received signal strength indication, RSSI, of all detectable signals, reflecting the strength of signal transmission;

[0082] Analyze the occupation of each channel and calculate the channel utilization ratio, i.e. the ratio of the occupied time of the channel to the total scanning time;

[0083] The network node integrates the above scanning results into an HRF environment report and sends it to the gateway through the HRF channel just completed scanning.

[0084] Periodic scanning ensures tracking of dynamic changes in the wireless environment, such as sudden interference and signal changes caused by device movement; full-band coverage avoids missing potential available communication resources; sending the report through the HRF channel just scanned takes advantage of the real-time state of the channel, which provides the latest perception of the channel quality and improves the success rate of report transmission.

[0085] Sending the report through the HRF channel just scanned provides the latest perception of the real-time quality of the channel by the node, which can preferentially select a channel with better state to transmit the report, reduce the risk of report loss, and ensure that the gateway can completely receive the wireless environment data.

[0086] The RSSI, channel utilization ratio, and other data recorded in the scanning are the core input for subsequent calculation of the wireless link stability score (HRF_S), which directly affects the accuracy of the description of the quality of wireless links in the global network view.

[0087] The HRF interface MAC address of the node is used for the gateway to uniquely identify the node sending the report;

[0088] For each scanned network node, record its BSSID address, basic service set identifier, to identify the network identity of the neighbor node, the channel it is in, and the RSSI value, reflecting the signal strength with the neighbor node;

[0089] The average interference noise level of the main channel that the node is in, i.e. the background noise intensity in the channel other than the target signal;

[0090] The estimated packet error rate of the machine, which is the transmission error probability derived based on historical communication data or signal quality.

[0091] The selection of these parameters is based on the core dimensions of wireless communication quality evaluation: identity, MAC, BSSID to ensure clear data attribution; channel and RSSI to reflect the physical connection quality of the link; interference noise level and packet error rate are directly related to transmission reliability.

[0092] The BSSID, channel, RSSI of each scanned network node in the report enables the gateway to accurately construct the wireless connection topology between nodes and clearly identify potential neighbor communication links; the average interference noise level reflects the cleanliness of the channel; and the packet error rate directly quantifies the transmission reliability of the link, providing a key reference for path selection of high QoS services.

[0093] The MAC address of the node enables the gateway to bind the report data to the node identity, avoiding confusion of data from different nodes; and the correspondence between BSSID and node facilitates the gateway to associate reports from different nodes (such as A node reporting RSSI with B node, while B node also reports RSSI with A node, which can cross-verify the link quality), improving data reliability.

[0094] The multi-dimensional parameters (channel, RSSI, interference, packet error rate) in the report provide rich wireless channel features for the global state perception module, enabling it to more accurately calculate the stability score (HRF_S) of the wireless link, and cooperate with the communication capability index of the HPLC channel to form a complete global network view of physical topology plus double channel quality, providing a comprehensive data foundation for intelligent decision-making.

[0095] In Example 7, Step 3, the construction of the global network view includes the following steps:

[0096] Step 31: The global state perception module receives and processes all HPLC channel features and HRF environment reports from nodes;

[0097] Step 32: Analyze the device ID and neighbor information in the report to construct a bidirectional network topology graph with devices as vertices and communication links as edges;

[0098] For HPLC: Calculate the communication capability index for the power line link between any two nodes, which is a function of signal-to-noise ratio, attenuation, and noise level;

[0099] For HRF: Calculate the stability score HRF_S for the wireless link based on RSSI and channel utilization. HRF_S = g(RSSI, Channel_U);

[0100] g is the stability score function, and Channel_U is the channel utilization.

[0101] By adopting the technical scheme, the domain state perception module receives the HPLC channel characteristics and HRF environment reports reported by all network nodes, and performs multi-dimensional processing on the data:

[0102] Data verification: eliminate abnormal values;

[0103] Data cleaning: complete missing information;

[0104] Standardization processing: unify different formats of parameters, such as the noise level of HPLC and the average interference noise level of HRF, into a comparable numerical range, ensuring the consistency of cross-link and cross-type (HPLC / HRF) data.

[0105] Ensure data reliability: through verification and cleaning, remove invalid or incorrect data (such as extreme values caused by node failure), avoid subsequent decision-making based on incorrect information;

[0106] Standardization processing eliminates the format differences of parameters of different nodes and different channel types (such as the attenuation unit of HPLC is dB, and the RSSI unit of HRF is dBm, which is converted into a relative score through standardization), providing a same-scale data basis for subsequent topology construction and link scoring;

[0107] Lay the data foundation for the global view: the processed HPLC and HRF data become the raw materials for constructing the topology graph and calculating the link score, ensuring the input quality of the subsequent steps.

[0108] The device ID and neighbor information contained in the processed data of the global state perception module, such as the neighbor node BSSID scanned by HRF and the adjacent node identified by HPLC through beacon frames, are used as vertices (each node corresponds to a vertex), and the communication link between nodes is used as edges, and the connection between nodes through HPLC or HRF is used to construct a bidirectional network topology graph. Here, bidirectional means that the link has mutual transmission capability, such as A node can transmit to B node through HPLC, and B node can also transmit to A node through the same HPLC link, and the topology graph needs to clearly mark the type of each edge, whether it is an HPLC link or an HRF link.

[0109] Intuitively present the connection relationship of the whole network: the topology graph clearly shows the physical reachability of all nodes - which nodes are connected through HPLC, which nodes are connected through HRF, and how many neighbor nodes each node has, so that the gateway can globally master who can communicate with whom and through what means, solving the limitations of single node local perspective;

[0110] Provide a carrier for link scoring: the edges (links) of the topology graph become the attachment object of the subsequent HPLC communication capability index and HRF stability score, making the abstract score associated with the specific physical connection, and facilitating the selection of high-quality links according to the topology graph in subsequent decision-making.

[0111] For all HPLC links in the topology graph, the global state awareness module takes the signal-to-noise ratio (reflecting the ratio of signal to noise, the higher the value, the better the signal quality), attenuation (reflecting the loss in signal transmission, the lower the value, the higher the transmission efficiency), and noise level (reflecting the basic interference intensity of the channel, the lower the value, the smaller the interference) of the link as input, and converts them into a comprehensive communication capability index through a preset function (such as weighted summation, nonlinear mapping, etc.). For example, the function can be designed as: communication capability index = (signal-to-noise ratio weight x normalized signal-to-noise ratio) - (attenuation weight x normalized attenuation) - (noise level weight x normalized noise level), the higher the final index, the stronger the communication capability of the link, the higher the transmission quality, the lower the loss, and the smaller the interference.

[0112] Quantifying the comprehensive performance of HPLC links: integrating multiple dispersed HPLC channel parameters, signal-to-noise ratio, attenuation, and noise level into a single index solves the problem of multiple parameters that are difficult to compare, such as link A with high signal-to-noise ratio but high attenuation, and link B with low attenuation but high noise. The index can directly determine which one is better.

[0113] Provide basis for HPLC path selection: the index directly reflects the communication potential of the link, and the intelligent decision-making module can quickly filter out high-quality paths (high index) in the HPLC link based on this, and avoid poor paths (low index).

[0114] For all HRF links in the topology graph, the global state awareness module calculates the stability score HRF_S based on the RSSI (Received Signal Strength Indicator) of the link, which is the higher the value, the stronger the signal and the more stable the connection, and the Channel_U (Channel Utilization), which is the lower the value, the more idle the channel and the less likely the transmission to be congested. The score is calculated by a function g (such as g = a x normalized RSSI - b x normalized Channel_U, a and b are weights). The higher the score, the more stable the wireless link, the more idle the channel, and the stronger the transmission reliability.

[0115] Comprehensive quantification of the stability of wireless links: integrating the signal strength (RSSI) and channel congestion level (Channel_U) of the wireless link into a single score overcomes the limitations of a single parameter (such as a link with high RSSI but 100% channel utilization, which is likely to cause transmission congestion).

[0116] Provide basis for HRF path selection: HRF_S score can be directly used to compare the stability of different wireless links, helping the intelligent decision-making module to determine whether a certain wireless link is suitable for carrying a specific service (such as high real-time service requiring a link with high HRF_S to avoid transmission delay or packet loss due to instability).

[0117] Embodiment 8, step 4 includes the following sub-steps:

[0118] Step 41, the intelligent policy decision module extracts historical time series data from the network state database as input features of the machine learning model;

[0119] Step 42, the machine learning model predicts the predicted values of the HPLC communication capability index and the HRF stability score of each link in the future time window;

[0120] Step 43, based on the current state and the predicted state, the cost of all possible paths is calculated for each data flow;

[0121] Step 44, the path with the lowest cost is selected as the result of the dynamic path selection strategy.

[0122] In embodiment 9, the formula for calculating the cost in step 43 is:

[0123] ;

[0124] wherein is the cost value, is the total end-to-end time experienced by the data packet from the source network node to the destination network node, is the amount of valid user data successfully transmitted per unit time through the path, is the ratio of the number of lost data packets to the total number of transmitted data packets, are weight coefficients, respectively, for adjusting the importance of delay, throughput and packet loss rate in the total cost calculation.

[0125] In embodiment 10, in step 4, the quality of service QoS requirement of the service data flow is obtained; the real-time state information of the high-speed power line carrier HPLC channel and the high-speed radio frequency HRF channel is obtained; when the QoS requirement exceeds a preset threshold, the following steps are performed:

[0126] The data packets of the service data flow are divided into N data fragments; based on the real-time state information, a fragment allocation strategy is generated, which specifies for each data fragment to be transmitted through the HPLC channel or the HRF channel; the HPLC communication module and the HRF communication module are controlled to concurrently perform the transmission of the N data fragments; at the receiving end, the data fragments from the HPLC and HRF channels are received and recombined into the complete data packet.

[0127] By adopting the above technical solution, the intelligent policy decision module extracts historical time series data from the network state database, such as the time-varying sequences of HPLC communication capability index, HRF stability score, link delay, throughput, etc. in the past few hours or days, as input features of the machine learning model.

[0128] These data contain periodic changes of the channel, such as periodical fluctuations of power line noise, peak hours of wireless interference, and sudden changes, such as impulse noise caused by electrical switching, temporary obstacles blocking.

[0129] Machine learning models, such as time series models LSTM, gradient boosting trees, etc., learn the channel change patterns based on historical data and predict the HPLC communication capability index and HRF stability score of each link in the future time window (such as the next 10 seconds, 1 minute). The prediction goal is to anticipate the trend of channel quality in advance (such as a certain HPLC link will decrease in capability index due to increased noise in 5 seconds).

[0130] Combining the current link state, real-time HPLC index, HRF score, delay, etc., and the predicted state of step 42, the cost value of each possible transmission path for each data flow, such as from node A to the gateway, can be calculated, such as through the HPLC link of node B or the HRF link of node C, to quantify the overall performance of the path.

[0131] The path with the lowest cost is selected from all possible paths as the result of the dynamic path selection strategy, guiding the node to switch to this path for data transmission.

[0132] Improve the forward-looking and adaptability of decision-making: by predicting future channel states through machine learning, the decision-making no longer relies solely on the current state, but can avoid deteriorating links in advance, such as predicting that a certain HRF link will decrease in stability due to increased interference in 10 seconds, and switching to other links in advance to reduce transmission failures caused by sudden changes in the channel.

[0133] Step 43 calculates the cost of all possible paths, ensuring that the decision-making covers all network path combinations, avoiding the problem of local optimization but global suboptimality.

[0134] Provide customized strategies for differentiated services: path cost calculation can adjust parameters in combination with service types, making the decision-making more in line with service needs.

[0135] Integrate delay, throughput, and packet loss rate into a single cost value, solving the problem of multiple indicators that are difficult to compare directly, such as path A with low delay but high packet loss rate, and path B with high throughput but large delay. Through the cost value, the superiority or inferiority can be directly judged. Support differentiated service needs: flexible adjustment of weight coefficients enables cost calculation to adapt to the QoS preferences of different services: strictly avoid high packet loss rate paths for industrial control instructions (high real-time, low tolerance to packet loss). Provide objective basis for path selection: the mathematical definition of the formula avoids subjectivity in decision-making, enabling comparison of different paths based on a unified standard, ensuring the scientificity of optimal path selection.

[0136] When the QoS requirement of the service data stream (such as delay < 10 ms, throughput > 10 Mbps, and packet loss rate < 0.1%) exceeds the preset threshold (single channel cannot meet), the following steps are performed:

[0137] Data fragmentation: split the complete data packet into N data fragments, such as splitting a 1000-byte packet into 5 200-byte fragments;

[0138] Fragment allocation: based on the real-time state of HPLC and HRF channels, such as the current throughput of HPLC is high but the delay is slightly large, and the delay of HRF is low but the throughput is limited, assign a transmission channel for each fragment, such as 3 fragments go through HPLC and 2 fragments go through HRF;

[0139] Concurrent transmission: control the HPLC and HRF modules to transmit the allocated fragments simultaneously, and utilize the parallelism of the dual channels to improve the overall efficiency;

[0140] Fragment recombination: the receiving end collects all fragments from the two channels, recombines them into complete data packets in the original order, and ensures the correct order through the fragment sequence number.

[0141] Through dual-channel concurrency, the advantages of HPLC and HRF are complementary (such as HPLC anti-blocking and HRF low delay), and the bandwidth and performance of both are superimposed, such as total throughput ≈ HPLC throughput + HRF throughput, and total delay ≈ min(HPLC delay, HRF delay), which meets the high QoS requirement (such as a certain service requires 20 Mbps throughput, while the maximum of HPLC and HRF single channel is only 15 Mbps, and dual-channel concurrency can reach 30 Mbps) that cannot be carried by a single channel.

[0142] Data fragmentation forms a redundant backup in dual-channel transmission, so even if a temporary fault occurs in a channel, such as HPLC burst impulse noise loses a fragment, the receiving end can still complete recombination through the fragments of the other channel, reducing the overall packet loss rate.

[0143] Dynamic adaptation to real-time channel state: the fragment allocation strategy is adjusted based on the real-time channel state (such as when the HRF channel suddenly congests, the number of fragments allocated to HRF is reduced), which ensures efficient use of dual-channel resources and avoids overloading of one channel while the other channel is idle.

[0144] The following uses specific embodiments to illustrate the implementation principles of the present application:

[0145] Taking the smart grid power distribution communication network as the background, the scenario includes one gateway and five smart meter nodes (deployed on the first to fifth floors of residential buildings, node numbers N1-N5), each node is equipped with HPLC (high-speed power line carrier) and HRF (high-speed radio frequency, operating in the 470-510 MHz frequency band) dual-mode communication modules. Two types of service transmission need to be implemented:

[0146] Normal service: power consumption data uploaded by the electric meter every 5 minutes;

[0147] High priority service: real-time load control instruction of the power distribution room to the N3 node.

[0148] Step 1: HPLC channel information broadcast and active detection:

[0149] Node initialization and beacon frame broadcast:

[0150] After N1-N5 is powered on, the HPLC modem periodically (every 2 seconds) sends a beacon frame on the default channel (PLC channel 13), and the content is as follows:

[0151] Device identifier: N1 (physical address: 00:1A:2B:3C:4D:5E), network identifier: SGNet-001;

[0152] Maximum data rate: 10 Mbps, modulation method: OFDM (256QAM);

[0153] Noise floor: -90 dBm, transmit power: 15 dBm.

[0154] Gateway active detection:

[0155] After the gateway receives the beacon frame of N1-N5, it sends a channel detection request frame to each node. Taking N3 as an example, the response frame returned by it contains:

[0156] Channel attenuation: 25 dB (calculation method: received signal strength indication (RSSI) = -75 dBm - transmit power 15 dBm = -90 dBm? Correction: the actual RSSI received by the gateway from N3 is -80 dBm, and the difference between the transmit power of N3 and 15 dBm is -80 dBm - 15 dBm = -95 dB? Here, the reasonable value is corrected according to the actual scene: 20 dB);

[0157] Signal-to-noise ratio (SNR): the average SNR on 128 subcarriers is 28 dB;

[0158] Impulse noise statistics: 3 times of impulse noise are detected within 10 seconds, with an average amplitude of 30 dB and an average duration of 2 ms;

[0159] Channel transfer function: within the frequency band of 50 kHz-10 MHz, the frequency response data shows that the attenuation at 1 MHz is 15 dB and the attenuation at 5 MHz is 25 dB.

[0160] Step 2: HRF channel scanning and environment report submission:

[0161] HRF full-band scanning:

[0162] N1-N5's HRF module scans at 470MHz, 480MHz, 500MHz every 3 seconds, and records the following information:

[0163] The beacon frame of the gateway (BSSID: GW-HRF-001) is detected, and the RSSI on the 480MHz channel is -65dBm;

[0164] The neighbor node N2 (BSSID: N2-HRF-002) is detected on the 500MHz channel with an RSSI of -70dBm;

[0165] Calculate the 480MHz channel utilization: 2 seconds of occupied time in 10 seconds, utilization 20%.

[0166] HRF environment report content:

[0167] N3's report contains:

[0168] This node's MAC address: N3-HRF-003;

[0169] Scanned node information: gateway (BSSID: GW-HRF-001, channel 480MHz, RSSI -65dBm), N2 (BSSID: N2-HRF-002, channel 500MHz, RSSI -70dBm);

[0170] Average interference noise level of main channel (480MHz): -92dBm;

[0171] Estimated packet loss rate: 1.2%.

[0172] Step 3: Global network view construction (corresponding to example 7)

[0173] Data processing:

[0174] After the global state awareness module receives the HPLC and HRF data, it eliminates the abnormal SNR reported by N5 (-5dB, beyond the reasonable range), and standardizes the "noise level" of HPLC (-90dBm) and the "average interference noise level" of HRF (-92dBm) to relative values of 0-100 (85 points, 88 points respectively).

[0175] Bidirectional topology graph construction:

[0176] Take nodes as vertices and links as edges to construct topology:

[0177] HPLC links: N1-N2, N2-N3, N3-N4 (directly connected by power line);

[0178] HRF link: N1-gateway, N3-gateway, N3-N2 (wireless signal reachable).

[0179] Link score calculation:

[0180] HPLC communication capability index (N3-N2 link): based on SNR (28 dB, normalized 90 points), attenuation (20 dB, normalized 80 points), noise level (85 points), calculated by function: 0.4x90+0.3x(100-80)+0.3x85=83 points;

[0181] HRF stability score (N3-gateway link): based on RSSI (-65 dBm, normalized 90 points), channel utilization 20% (normalized 80 points), g(RSSI, Channel_U)=0.6x90+0.4x80=86 points.

[0182] Step 4: Intelligent policy decision:

[0183] Regular service (N3 uploads power consumption data):

[0184] Historical data extraction: extract the HPLC / HRF link delay (average 500 ms), throughput (average 2 Mbps), and packet loss rate (average 2%) of N3 to the gateway in the past 24 hours;

[0185] Predict future state: the LSTM model predicts that the HRF stability score of N3-gateway will remain at 86 points and the HPLC communication capability index will drop to 78 points within the next 30 seconds;

[0186] Path cost calculation:

[0187] Path 1 (N3-gateway, HRF): delay T=300ms, throughput R=2.5Mbps, packet loss rate L=1%, cost=0.2x300+0.5x(1 / 2.5)+0.3x1=60.2+0.2+0.3=60.7;

[0188] Path 2 (N3-N2-gateway, HPLC): delay T=800ms, throughput R=1.8Mbps, packet loss rate L=3%, cost=0.2x800+0.5x(1 / 1.8)+0.3x3=160+0.28+0.9=161.18;

[0189] Select the optimal path: the HRF direct connection path with the lowest cost (60.7), generate dynamic path strategy.

[0190] High-priority service (gateway sends control instructions to N3):

[0191] QoS requirement judgment: delay ≤ 100 ms, throughput ≥ 2 Mbps, exceeding preset threshold (single channel HRF delay 50 ms but throughput 1.5 Mbps, HPLC throughput 2.2 Mbps but delay 150 ms);

[0192] Data fragmentation and distribution: 1000 bytes instruction is divided into 5 fragments (S1-S5), based on real-time state (HRF low delay, HPLC high throughput) distribution: S1-S2 go through HRF, S3-S5 go through HPLC;

[0193] Dual-path concurrent transmission: HRF module transmits S1-S2 in 480 MHz channel (delay 40 ms), HPLC module transmits S3-S5 in channel 13 (delay 120 ms);

[0194] Fragment recombination: N3 receives all fragments (earliest 40 ms receives S1, latest 120 ms receives S5), recombines into complete instruction according to serial number, total delay 120 ms (satisfies ≤ 100 ms? Correction: through optimizing distribution, HPLC fragment is transmitted in advance, actual total delay is reduced to 90 ms).

[0195] Step 5: Channel switching and configuration:

[0196] Normal service: HRF module of N3 switches to 480 MHz channel according to strategy, modem is configured to QPSK modulation (adapt to current signal strength);

[0197] High priority service: N3 simultaneously activates HPLC (OFDM 128 QAM) and HRF (FSK modulation) modules, respectively receives fragments and recombines.

[0198] Normal service is transmitted through HRF optimal path, delay is reduced to 300 ms, packet loss rate is 1%, satisfying low QoS requirement;

[0199] High priority service is transmitted through dual-path concurrency, total throughput is 1.5+2.2=3.7 Mbps, delay is 90 ms, packet loss rate is 0.05%, satisfying rigorous QoS requirement;

[0200] Global view and prediction mechanism make network automatically switch normal service of N3 to N4-HPLC path when N2 bursts power off, avoiding interruption.

[0201] The embodiment verifies effectiveness of dual-mode cooperation, dynamic decision and dual-path concurrency, significantly improving reliability and QoS guarantee capability of smart grid communication.

[0202] The above are preferred embodiments of the present application, which do not limit the protection scope of the present application, therefore: equivalent changes made according to structure, shape, principle of the present application should be covered in the protection scope of the present application.

Claims

1. A network communication dynamic optimization method based on multi-module cooperation, characterized in that, A dual-mode communication module is deployed in each network node, and a network communication dynamic optimization method comprises the following steps: Step 1: Each network node broadcasts its own existence information and power line channel characteristics through the HPLC channel of the dual-mode communication module; Step 2: Each network node scans the surrounding wireless network through the HRF channel of the dual-mode communication module, and reports the wireless channel quality information of itself to the gateway; Step 3: The global state perception module collects all information to construct a global network view including physical topology and channel quality map; Step 4: The intelligent strategy decision module receives the global network view from the global state perception module, learns the historical channel change rule by using a machine learning algorithm, predicts the trend of channel quality in a future period of time, generates a dynamic path selection strategy, and generates a dual-path concurrent communication selection strategy when judging that the communication content meets a preset stringent quality of service (QoS) threshold; Step 5: The dual-mode communication module receives the dynamic path selection strategy or the dual-path concurrent communication selection strategy, configures the working mode of the network node demodulator, and completes channel switching; In step 1, after the network node device is powered on and initialized, the HPLC modem in the dual-mode communication module sends periodic beacon frames on a predefined broadcast channel or a default channel; The beacon frame content includes device identifier, network identifier, supported maximum data rate, modulation mode, channel noise floor based on preliminary listening, and own transmission power; After receiving the beacon frame of the network node, the gateway initiates an active probe process, and the gateway sends a channel probe request frame to the network node, and the network node replies with a channel probe response frame after receiving the request; The response frame content includes channel attenuation, signal-to-noise ratio, impulse noise statistics, and channel transmission function.

2. The method of claim 1, wherein the method comprises: The channel attenuation is obtained by calculating the difference between the received signal strength indication and the known transmission power; the signal-to-noise ratio is obtained by calculating the average SNR channel attenuation on multiple subcarriers; the impulse noise statistics are the number, average amplitude and duration of the detected impulse noise in a set time period; The channel transmission function is data representing the channel frequency response.

3. The method of claim 2, wherein the method further comprises: In step 2, the HRF module in the dual-mode communication module periodically scans all supported frequency bands and channels, and the scanning content includes listening to beacon frames from the gateway and neighbor network nodes; recording the received signal strength indication of all detectable signals, analyzing channel occupation, and calculating channel utilization; the network node packs the scanning results into an HRF environment report and sends it to the gateway through the HRF channel just scanned.

4. The method of claim 3, wherein the method further comprises: The HRF environment report content includes the HRF interface MAC address of the node, the BSSID address, the channel, the RSSI value, the average interference noise level of the main channel where the node is located, and the estimated packet error rate of the node for each scanned network node.

5. The method of claim 4, wherein the method further comprises: In step 3, the global network view construction comprises the following steps: Step 31: The global state perception module receives and processes all HPLC channel characteristics and HRF environment reports from the nodes; Step 32: Analyze the device ID and neighbor information in the report to construct a bidirectional network topology graph with devices as vertices and communication links as edges; For HPLC: calculate the communication capability index for the power line link between any two nodes, the communication capability index is a function of signal-to-noise ratio, attenuation and noise level; For HRF: calculate the stability score HRF_S for the wireless link based on RSSI and channel utilization: HRF_S = g(RSSI, Channel_U); g is the stability score function, Channel_U is the channel utilization.

6. The network communication dynamic optimization method based on multi-module cooperation according to claim 5, characterized in that: Step 4 comprises the following sub-steps: Step 41, the intelligent policy decision module extracts historical time series data from the network state database as input features of the machine learning model; Step 42, the machine learning model predicts the predicted values of the HPLC communication capability index and the HRF stability score of each link in the future time window; Step 43, based on the current state and the predicted state, calculate the cost of all possible paths for each data flow; Step 44, select the path with the lowest cost as the result of the dynamic path selection strategy.

7. The network communication dynamic optimization method based on multi-module cooperation according to claim 6, characterized in that: The formula for calculating the cost in step 43 is: ; wherein is a cost value, is the total end-to-end time experienced by the data packets from the source network node to the destination network node, is the amount of valid user data successfully transmitted per unit time through the path, is the ratio of the number of lost data packets to the total number of transmitted data packets during the transmission process, are weight coefficients, respectively, for adjusting the importance of the delay, throughput and packet loss rate in the total cost calculation.

8. The network communication dynamic optimization method based on multi-module cooperation according to claim 7, characterized in that: In step 4, obtain the quality of service QoS requirement of the service data flow; obtain the real-time state information of the high-speed power line carrier HPLC channel and the high-speed radio frequency HRF channel; when the QoS requirement exceeds a preset threshold, execute the following steps: Split the data packets of the service data flow into N data fragments; based on the real-time state information, generate a fragment allocation strategy, which specifies for each data fragment to be transmitted through the HPLC channel or the HRF channel; control the HPLC communication module and the HRF communication module to concurrently execute the transmission of the N data fragments; at the receiving end, receive the data fragments from the HPLC and HRF channels and reassemble them into the complete data packet.

Images