CPE router, data transmission method and device, medium and product

Abstract

The invention discloses a CPE router, a data transmission method and device, a medium and a product, and relates to the technical field of network communication and data transmission optimizing.According to the scheme, the method comprises the steps of collecting and preprocessing inlet flow and wireless channel data of the CPE router, extracting transmission urgency and a channel quality index, constructing a transmission decision model and outputting service priority classification; and meanwhile, a dynamic transmission strategy and an early warning mechanism are formulated in combination with a channel prediction model result, and network resource allocation is optimized. According to the scheme, accurate matching and intelligent scheduling of network resources are achieved by dynamically sensing the channel state and the service characteristics, key service transmission delay is effectively reduced, the throughput of high-priority data is improved, meanwhile, transmission stability in a complex environment is enhanced, data retransmission caused by channel fluctuation is reduced, and the transmission efficiency is improved. Smoother multi-service concurrent experience is provided for users, potential transmission risks are warned in advance, and network service quality and equipment operation reliability are guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of network communication and data transmission optimization, and in particular to a CPE router, a data transmission method, a device, a medium and a product. Background Art

[0002] CPE routers are terminal network devices deployed on the user side. Their core function is to convert the 4G and 5G mobile network signals, satellite signals, and wired broadband signals provided by operators into Wi-Fi and wired signals, providing Internet access for homes, offices, and industrial scenarios. This type of device achieves network access by inserting a SIM card and connecting to optical fiber. It supports simultaneous connection of smartphones, computers, and IoT devices, and has signal conversion, network management, and security protection capabilities. Typical application scenarios include network coverage in remote areas, mobile office, emergency communications, and industrial IoT data transmission. Modern 5G CPE routers also support Wi-Fi 6 technology, dual-band concurrency, and gigabit network port output, enabling high-speed and low-latency network transmission to meet the high-bandwidth requirements of 4K and 8K video streaming, online gaming, and telemedicine, becoming a key solution to replace traditional broadband.

[0003] To address resource contention conflicts and channel response lags during CPE router data transmission, existing technologies employ static priority allocation and fixed channel scheduling. However, this can lead to insufficient adaptability to bursty traffic flows and delayed responses to sudden channel state changes, resulting in high jitter rates for high-priority traffic transmission and low channel switching success rates. Traditional solutions lack a dynamic traffic classification mechanism and channel quality prediction capabilities, resulting in low measured bandwidth utilization in mixed traffic scenarios and a high standard deviation for voice service latency, severely restricting service quality assurance in multi-service concurrent scenarios. To address these issues, a CPE router, data transmission method, device, medium, and product are proposed. Summary of the Invention

[0004] The present invention aims to provide a CPE router, a data transmission method, a device, a medium and a product to solve the problems raised in the above background technology.

[0005] To solve the above technical problems, the technical solution adopted by the present invention is as follows: in a first aspect, a CPE router is provided, comprising a traffic sniffing module, a multi-band spectrum analysis module, a feature calculation unit, a quality assessment coprocessor, a transmission decision hardware module, a channel prediction module, a policy decision unit, a parameter update coprocessor, and a status detection module;

[0006] Traffic sniffing module, used for capturing and preliminarily processing CPE router ingress traffic data;

[0007] Multi-band spectrum analysis module for real-time monitoring of wireless channel status;

[0008] A feature calculation unit, used to extract transmission urgency features;

[0009] a quality assessment coprocessor for calculating a channel quality index;

[0010] Transmission decision hardware module for service priority classification;

[0011] Channel prediction module, used for channel quality prediction;

[0012] A policy decision unit, used to generate dynamic transmission policies;

[0013] A parameter update coprocessor for dynamically adjusting bandwidth allocation weight coefficients;

[0014] Status detection module, used to monitor network status in real time.

[0015] In a second aspect, a data transmission method is provided, applied to the CPE router according to any one of the first aspects, comprising the following steps:

[0016] S1. Collect and pre-process the CPE router's ingress traffic data and wireless channel time slot occupancy status data;

[0017] S2, extracting the transmission urgency and channel quality index during the data transmission process from the pre-processed ingress traffic data and wireless channel time slot occupancy status data;

[0018] S3. Based on the transmission urgency and the channel quality index, a transmission decision model and a channel prediction model are constructed to output the service priority classification results and the channel quality index prediction value respectively;

[0019] S4. Combine the service priority classification results with the channel quality index prediction value to formulate data transmission strategies and early warning mechanisms.

[0020] A further improvement of the technical solution of the present invention is that in S1, the collection and preprocessing process of the CPE router's ingress traffic data and wireless channel time slot occupancy status data includes:

[0021] A traffic sniffing module is deployed at the network interface of the CPE router. This module consists of a multi-port mirroring device and an embedded processing chip, which is physically connected to the data bus of the main control board. The wireless baseband chip integrates a multi-band spectrum analysis module, including 2.4GHz and 5GHz dual radio frequency channels and a signal processing unit.

[0022] The ingress traffic is replicated through port mirroring technology, and a protocol parser is used to extract the packet quintuple and payload features from the ingress traffic data. The packet quintuple includes the source address, destination address, protocol type, source port, and destination port. Continuous flows with a packet length of ≤512 bytes and a transmission interval of ≤20ms are detected to identify voice services. Data flows with a video encoding identifier in the payload are captured to identify video services. TCP flows containing file fragment numbers and cyclic redundancy check codes are identified to filter out file transfer services.

[0023] A full-band rapid scan is performed every 100ms to obtain the occupancy status data of each wireless channel time slot, and the channel occupancy rate C and effective signal-to-noise ratio S are calculated. The calculation process is as follows:

[0024]

[0025] Among them, busy t is the occupancy state 0 and 1 of time slot t, N = 50 sampling time slots, P r is the received signal strength, P n Obtained in real time by the background noise measurement circuit;

[0026] Discard checksum error packets and illegal port data in ingress traffic data. Establish a mapping rule base and mark continuous traffic flows with discrete service labels. Voice services are labeled 1, video services are labeled 2, and file transfer services are labeled 3. A leaky bucket algorithm is implemented to control burst traffic and obtain the total amount of burst traffic data passing through the CPE router ingress per unit time.

[0027] A sliding average filter is applied to the channel occupancy rate, and a median filter is used for the effective signal-to-noise ratio. The channel occupancy rate is mapped to the interval [0,1], and the effective signal-to-noise ratio is nonlinearly compressed. When three consecutive sampling values exceed the historical mean ±3σ range, the hardware self-check program is triggered and the previous value hold strategy is adopted. At the same time, a time synchronization module is introduced to align the timestamps of the ingress traffic data and the wireless channel time slot occupancy status data.

[0028] A further improvement of the technical solution of the present invention is that in S2, the process of extracting the transmission urgency and the channel quality index during data transmission includes:

[0029] A feature calculation unit containing a dedicated register group and a floating-point operation accelerator is added to the embedded processing chip, which is connected to the pre-processed service label data bus. A quality assessment coprocessor containing a sliding window cache and a nonlinear operation unit is integrated into the signal processing unit, which is directly connected to the filtered wireless channel time slot occupancy status data.

[0030] According to the service type, the statistical period W of voice service, video service and file transfer service is set to 0.5s, 1s and 2s respectively, and the request frequency F of each service type is calculated. t , the calculation process is as follows:

[0031]

[0032] Among them, M is the number of requests for each business in the statistical period, l avg is the average length of the data packet, L ref The base length is 256 bytes for voice services, 1024 bytes for video services, and 2048 bytes for file transfer services.

[0033] When the request frequency change rate ΔF between two consecutive windows exceeds 30%, the window adaptation mechanism is activated to update the window statistical period of each service type.

[0034] Based on the normalized channel occupancy C and the effective signal-to-noise ratio S after nonlinear compression, the channel quality index Q is calculated. The calculation process is as follows:

[0035]

[0036] in, is the channel occupancy fluctuation coefficient;

[0037] When a channel switching event is detected, instantaneous quality compensation is started Among them, λ=0.1 is the gradient sensitivity coefficient, is the signal-to-noise ratio change rate;

[0038] When F t When a division by zero error occurs, set F t =0, if Q exceeds the theoretical range [0,1], the constraint function Q is activated clamp =max(0,min(1,Q)).

[0039] A further improvement of the technical solution of the present invention is that in S3, the process of constructing a transmission decision model and outputting a service priority classification result includes:

[0040] A dedicated hardware module is integrated into the embedded processing chip. The dedicated hardware module includes an attribute selector, a rule memory, and a branch predictor. The module is connected to the transmission urgency output by the feature calculation unit via a high-speed bus. The module deploys a programmable logic array, has a built-in dynamic pruning controller and a gain rate calculation unit, and receives service type labels and channel quality indexes in real time.

[0041] A two-dimensional feature vector is constructed based on service labels and transmission urgency, and continuous transmission urgency is binned by service type. The binning process includes dividing voice services into [0, 6), [6, 12), and [12, +∞), video services into [0, 9), [9, 18), and [18, +∞), and file transfers into [0, 15), [15, 30), and [30, +∞). This binning process implements real-time threshold matching through a hardware comparator array, converting service labels and transmission urgency into discrete inputs that can be processed by the improved C5.0 decision tree algorithm.

[0042] The transmission decision model is constructed using an improved C5.0 decision tree algorithm. The attribute splitting gain rate is used as the node splitting criterion to obtain the ratio of information gain to the amount of information in the split, thereby suppressing the bias problem of multi-valued attributes. The complexity of the transmission decision model is controlled through a dynamic pruning strategy. When the number of node samples is less than 1% of the total, the split is terminated. The transmission decision model training phase adopts an online incremental learning mechanism and uses the channel quality index to dynamically update the binning threshold, so that the transmission decision model can adapt to changes in network status.

[0043] A three-layer tree structure is adopted. The first layer branches according to the service type. The second layer splits the child nodes according to the transmission urgency bin threshold. The leaf nodes are mapped to priority labels, which include highest priority, medium priority and low priority. Voice service and video service are classified as the highest priority when the transmission urgency exceeds 12 and 18 respectively. Voice service, video service and file transfer service are classified as medium priority when the transmission urgency is in the range of [6,12), [9,18) and [15,+∞) respectively. Voice service, video service and file transfer service are classified as low priority when the transmission urgency is in the range of [0,6), [0,9) and [0,15) respectively.

[0044] Three groups of comparators are deployed to synchronously process each business type branch. The first level of the comparator completes business type matching, the second level performs transmission urgency threshold interval comparison, and the third level outputs a binary-coded priority label. At the same time, when the input request frequency is an illegal value, it is classified as the lowest priority. When the business label is not recognized, it is processed as a file transfer business by default.

[0045] A further improvement of the technical solution of the present invention is that in S3, the process of constructing a channel prediction model and outputting a channel quality index prediction value includes:

[0046] A trend extraction unit and a periodicity detection unit are added to the quality assessment coprocessor, including a sliding window cache and a Loess local weighted regression operator. The preprocessed channel quality index sequence is connected via a dedicated bus. A triple computing core is deployed, including horizontal component calculation cores, trend component calculation cores, and seasonal component calculation cores. A prediction execution unit, including a time-shift register group and a state transfer controller, is integrated to build a multi-step prediction pipeline.

[0047] Taking the channel quality index as input, a hybrid prediction model architecture of time series decomposition and exponential smoothing is adopted to construct a channel prediction model;

[0048] In the time series decomposition stage, the channel quality index sequence Q t Decomposed into trend component C t , seasonal component S t and the residual component R t ,The trend component is extracted using a sliding window local weighted regression algorithm with ,the window width set to 10 time slots. The seasonal component is ,separated by the periodic difference method with a preset period length of 10 time slots. The residual ,component is obtained by subtracting the trend component and the seasonal ,component from the channel quality index;

[0049] In the initialization phase of exponential smoothing forecast, the horizontal component l0=C1 and the trend component With seasonal component s -m+1:0 =S1:S 10 , dynamically update the horizontal component l by recursive calculation of the smoothing coefficient t , trend component b t With seasonal component s t , superimpose the linear trend and extrapolate the seasonality to generate the channel quality prediction value for the next three time slots A parameter adaptive adjustment mechanism is introduced to enable the channel prediction model to dynamically track changes in channel status. The calculation process is as follows:

[0050] l t =α(C t -s t-m )+(1-α)(l t-1 +b t-1 );

[0051] b t =β(l t -l t-1 )+(1-β)b t-1 ;

[0052] s t =γ(C t -l t )+(1-γ)s t-m ;

[0053]

[0054] Where k = 1, 2, 3;

[0055] Encapsulating the predicted channel quality value for the next three time slots in a fixed format and transmitting it via an independent low-latency bus, wherein the fixed format includes a timestamp, the predicted channel quality value for the three time slots, and a check code;

[0056] When the input channel quality index sequence exceeds the range [0,1], the boundary constraint is enabled. When the channel quality index sequence is interrupted for more than 3 time slots, historical data interpolation is started If the deviation between the channel quality prediction value and the measured channel quality for two consecutive times is greater than 0.2, the prediction is switched to the simple moving average prediction.

[0057] A further improvement of the technical solution of the present invention is that in S4, the process of formulating a data transmission strategy and an early warning mechanism by combining the service priority classification result and the channel quality index prediction value includes:

[0058] A policy decision unit is deployed on the main control board. It receives service priority classification results and channel quality predictions via a dual-channel bus. It integrates a transmission policy matrix corresponding to priority P∈{0,1,2}×channel quality Q'∈{low,medium,high} and a bandwidth allocation calculation core. A status monitoring module, including a transmission delay timer, a packet loss rate counter, and a channel switching trigger, is configured on the communication center side. This module is connected to the output bus of the policy execution engine. A parameter update coprocessor containing a PID control algorithm unit is added. This coprocessor receives warning signals in real time and generates weight adjustment coefficients, which are then transmitted back to the transmission decision model and channel prediction model via a reverse bus.

[0059] Combine the service priority classification results with the channel quality index prediction value to dynamically allocate bandwidth to services of different priorities. The calculation process is as follows:

[0060]

[0061] Among them, P i =0, 1, 2 are low priority, medium priority and highest priority of the business, BW total The bandwidth allocation result is written into the queue management register through the time division multiplexing controller, and the bandwidth allocation weight coefficient is refreshed every 50ms;

[0062] The maximum permissible delay for voice services is set to 30ms, the maximum permissible delay for video services is set to 80ms, and the maximum permissible delay for file transfer services is set to 500ms. When the end-to-end delay of non-low priority services exceeds the maximum permissible delay of the corresponding service type, it is determined to be a delay warning. If the gradient of the channel quality prediction value decreases for three consecutive time slots and meets the If it is judged as a channel degradation warning, a channel switching instruction is triggered, releasing 20% of the current channel resources to occupy a new channel;

[0063] Preemptive weighted round-robin scheduling is adopted. A 3-bit priority tag is written into the TOS field of the IP header during data encapsulation. The transport layer adds 2 bytes of metadata to record the channel quality prediction value.

[0064] When the channel prediction fails and switches to the moving average mode, the bandwidth allocation weight coefficient is forced to be set to 0.5, and the multipath transmission function is turned off. If the policy execution engine fails, the preset conservative policy BW is enabled. alloc =0.3×BW total ×(P+1), the data path switches to the backup bus.

[0065] In a third aspect, a data transmission device is provided, which includes a routing data acquisition module, a routing feature extraction module, a priority prediction module, and a transmission warning module;

[0066] The routing data acquisition module collects and pre-processes the ingress traffic data and wireless channel time slot occupancy status data of the CPE router;

[0067] The routing feature extraction module extracts the transmission urgency and channel quality index during the data transmission process from the pre-processed inlet traffic data and the wireless channel time slot occupancy status data;

[0068] The priority prediction module constructs a transmission decision model and a channel prediction model based on the transmission urgency and the channel quality index, and outputs the service priority classification result and the channel quality index prediction value respectively;

[0069] The transmission warning module formulates a data transmission strategy and warning mechanism by combining the service priority classification result and the channel quality index prediction value.

[0070] In a fourth aspect, a computer-readable storage medium is provided, the computer-readable storage medium comprising computer-executable instructions, which, when executed on a computer, enable the computer to execute the data transmission method described in the first aspect.

[0071] In a fifth aspect, a computer program product is also provided, which includes a computer program and instructions. When the computer instructions are run on a data transmission device, the data transmission device executes the data transmission method as described in the first aspect above.

[0072] It should be noted that the above-mentioned computer instructions can be stored in whole or in part on a computer-readable storage medium, wherein the computer-readable storage medium can be packaged together with the processor of the data transmission device, or can be packaged separately from the processor of the data transmission device. This embodiment of the present application does not limit this.

[0073] The description of the third, fourth and fifth aspects of this application can refer to the detailed description of the first and second aspects.

[0074] In the embodiments of the present application, the name of the above-mentioned data transmission device does not constitute a limitation on the device and functional module itself. In actual implementation, these devices and functional modules may appear with other names, as long as the functions of each device and functional module are similar to those of the present application and fall within the scope of the claims of the present application and their equivalent technologies.

[0075] Due to the adoption of the above technical solution, the present invention has the following technical advancements compared to the prior art:

[0076] 1. The present invention provides a data transmission method. By deploying a multi-port mirroring device and a spectrum analysis module to collaboratively collect inlet traffic and channel status data, combined with protocol parsing and multi-dimensional feature extraction technology, it can achieve accurate quantitative evaluation of data transmission urgency and channel quality in multiple business scenarios, effectively improving the transmission efficiency and real-time guarantee capability of high-priority services.

[0077] 2. The present invention provides a data transmission method that adopts an improved decision tree and hybrid prediction model architecture, generates interpretable priority classification rules and multi-step channel quality prediction values based on dynamic binning of service types and channel state time series decomposition technology, and improves network resource utilization through dynamic bandwidth allocation and preemptive scheduling mechanism, while reducing the transmission jitter rate of low-latency services.

[0078] 3. The present invention provides a data transmission method, constructs a closed-loop feedback control system, and uses a channel switching early warning trigger mechanism and a weight adaptive algorithm to correct transmission strategy parameters in real time, maintain high service availability when channel quality suddenly changes and business traffic surges, and reduce bandwidth waste caused by static resource configuration in traditional solutions.

[0079] The beneficial effects of the first, second, third, fourth and fifth aspects of this application can all be referred to the analysis of the above beneficial effects, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] In order to more clearly illustrate the embodiments of the present application and the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0081] Figure 1 A flowchart of a data transmission method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0082] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0083] It should be noted that in the embodiments of this application, words such as "exemplary" and "for example" are used to indicate examples, illustrations, and descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" and "for example" is intended to present the relevant concepts in a concrete manner.

[0084] In order to facilitate a clear description of the technical solutions of the embodiments of the present application, in the embodiments of the present application, words such as "first" and "second" are used to distinguish between identical items and similar items with basically the same functions and effects. Those skilled in the art can understand that words such as "first" and "second" do not limit the quantity and execution order.

[0085] Example 1, as Figure 1 As shown, the present invention provides a CPE router, including a traffic sniffing module, a multi-band spectrum analysis module, a feature calculation unit, a quality assessment coprocessor, a transmission decision hardware module, a channel prediction module, a policy decision unit, a parameter update coprocessor and a status detection module;

[0086] Traffic sniffing module, used for capturing and preliminarily processing CPE router ingress traffic data;

[0087] Multi-band spectrum analysis module for real-time monitoring of wireless channel status;

[0088] A feature calculation unit, used to extract transmission urgency features;

[0089] a quality assessment coprocessor for calculating a channel quality index;

[0090] Transmission decision hardware module for service priority classification;

[0091] Channel prediction module, used for channel quality prediction;

[0092] A policy decision unit, used to generate dynamic transmission policies;

[0093] A parameter update coprocessor for dynamically adjusting bandwidth allocation weight coefficients;

[0094] Status detection module, used to monitor network status in real time.

[0095] Example 2, as Figure 1 As shown, based on Example 1, the present invention provides a technical solution: a data transmission method applied to a CPE router, comprising the following steps:

[0096] S1. Collect and pre-process the ingress traffic data and wireless channel time slot occupancy status data of the CPE router, deploy a traffic sniffing module at the network interface of the CPE router, the traffic sniffing module consists of a multi-port mirroring device and an embedded processing chip, is physically connected to the data bus of the main control board, and a multi-band spectrum analysis module is integrated in the wireless baseband chip, including 2.4GHz and 5GHz dual radio frequency channels and a signal processing unit. The ingress traffic is replicated through port mirroring technology, and a protocol parser is used to extract the packet quintuple and payload characteristics from the ingress traffic data. The packet quintuple includes the source address, destination address, protocol type, source port, and destination port. Continuous streams with a packet length of ≤512 bytes and a sending interval of ≤20ms are detected to identify voice services, capture data streams with a video coding identifier in the payload, identify video services, identify TCP streams with file fragment sequence numbers and cyclic redundancy check codes, and filter out file transfer services. A full-band rapid scan is performed every 100ms to obtain the time slot occupancy status data of each wireless channel, and calculate the channel occupancy rate C and the effective signal-to-noise ratio S. The calculation process is as follows:

[0097]

[0098]

[0099] Among them, busy t is the occupancy state 0 and 1 of time slot t, N = 50 sampling time slots, P r is the received signal strength, P nThe noise floor measurement circuit acquires data in real time, discarding checksum error packets and illegal port data in the ingress traffic data. A mapping rule base is established, and continuous traffic flows are marked with discrete service labels, where voice services are marked as 1, video services are marked as 2, and file transfer services are marked as 3. A leaky bucket algorithm is implemented to control burst traffic, obtaining the total amount of burst traffic data passing through the CPE router ingress per unit time. A sliding average filter is applied to the channel occupancy rate, and a median filter is used for the effective signal-to-noise ratio. The channel occupancy rate is mapped to the [0,1] interval, and the effective signal-to-noise ratio is nonlinearly compressed. When three consecutive sampling values exceed the historical mean ±3σ range, the hardware self-test program is triggered and the previous value hold strategy is adopted. At the same time, a time synchronization module is introduced to align the timestamps of the ingress traffic data with the wireless channel time slot occupancy status data.

[0100] S2. Extract the transmission urgency and channel quality index during the data transmission process from the pre-processed ingress traffic data and the wireless channel time slot occupancy status data. Add a feature calculation unit containing a dedicated register group and a floating-point operation accelerator to the embedded processing chip, access the pre-processed service label data bus, integrate a quality assessment coprocessor containing a sliding window cache and a nonlinear operation unit in the signal processing unit, directly connect to the filtered wireless channel time slot occupancy status data, set the statistical period W of voice service, video service and file transfer service to 0.5s, 1s and 2s respectively according to the service type, and calculate the request frequency F of each service type. t , the calculation process is as follows:

[0101]

[0102] Among them, M is the number of requests for each business in the statistical period, L avg is the average length of the data packet, L ref The benchmark length is 256 bytes for voice services, 1024 bytes for video services, and 2048 bytes for file transfer services. When the request frequency change rate ΔF of two consecutive windows exceeds 30%, the window adaptation mechanism is started to update the window statistical period of each service type. Based on the normalized channel occupancy C and the effective signal-to-noise ratio S after nonlinear compression, the channel quality index Q is calculated. The calculation process is as follows:

[0103]

[0104] in, The channel occupancy fluctuation coefficient is used to start instantaneous quality compensation when a channel switching event is detected. Among them, λ=0.1 is the gradient sensitivity coefficient, is the signal-to-noise ratio change rate, when Ft When a division by zero error occurs, set F t =0, if Q exceeds the theoretical range [0,1], the constraint function Q is activated clamp =max(0,min(1,Q));

[0105] S3. Based on the transmission urgency and channel quality index, a transmission decision model and a channel prediction model are constructed to output the service priority classification result and the channel quality index prediction value respectively. A dedicated hardware module is integrated in the embedded processing chip. The dedicated hardware module includes an attribute selector, a rule memory and a branch predictor. The transmission urgency output by the feature calculation unit is connected via a high-speed bus. A programmable logic array is deployed with a built-in dynamic pruning controller and a gain rate calculation unit. The service type label and the channel quality index are received in real time. A two-dimensional feature is constructed based on the service label and the transmission urgency. The continuous transmission urgency is binned by service type. The binning process includes voice services being divided into [0, 6), [6, 12), and [12, +∞), video services being divided into [0, 9), [9, 18), and [18, +∞), and file transfers being divided into [0, 15), [15, 30), and [30, +∞). This binning process uses a hardware comparator array to implement real-time threshold matching, converting service labels and transmission urgency into discrete inputs that can be processed by the improved C5.0 decision tree algorithm. The improved C5.0 decision tree algorithm is used.0 decision tree algorithm is used to construct the transmission decision model. The attribute splitting gain rate is used as the node splitting criterion to obtain the ratio of information gain to split information amount, suppress the multi-value attribute bias problem, and control the complexity of the transmission decision model through dynamic pruning strategy. When the number of node samples is less than 1% of the total, the split is terminated. The online incremental learning mechanism is adopted in the training stage of the transmission decision model. The channel quality index is used to dynamically update the binning threshold so that the transmission decision model can adapt to changes in network status. A three-layer tree structure is adopted. The first layer branches according to the service type, and the second layer splits the child nodes according to the transmission urgency binning threshold. The leaf nodes are mapped to priority labels. The priority labels include highest priority, medium priority and low priority. Voice service and video service are classified as the highest priority when the transmission urgency exceeds 12 and 18 respectively. Voice service, video service and file transfer service are classified as medium priority when the transmission urgency is in the range of [6,12), [9,18) and [15,+∞) respectively. Voice service, video service and file transfer service are classified as medium priority when the transmission urgency is in the range of [0,6 ), [0,9), and [0,15), it is classified as low priority. Three groups of comparators are deployed to synchronously process each service type branch. The first level of the comparator completes service type matching, the second level performs transmission urgency threshold interval comparison, and the third level outputs a binary-coded priority label. At the same time, when the input request frequency is an illegal value, it is classified as the lowest priority. When the service label is not recognized, it is processed as a file transfer service by default. A trend extraction unit and a periodicity detection unit are added to the quality assessment coprocessor, including a sliding window cache and a Loess local weighted regression operator. The preprocessed channel quality index sequence is connected via a dedicated bus. A triple computing core is deployed, including a horizontal component calculation core, a trend component calculation core, and a seasonal component calculation core. A prediction execution unit including a time-shift register group and a state transfer controller is integrated to construct a multi-step prediction pipeline. The channel quality index is used as input and a hybrid prediction model architecture of time series decomposition and exponential smoothing is adopted to build a channel prediction model. In the time series decomposition stage, the channel quality index sequence Q is converted into a channel prediction model. t Decomposed into trend component C t , seasonal component S t and the residual component R t The trend component is extracted by sliding window local weighted regression algorithm, the window width is set to 10 time slots, the seasonal component is separated by periodic difference method, and the period length is preset to 10 time slots. The residual component is obtained by subtracting the trend component and seasonal component from the channel quality index. In the initialization stage of exponential smoothing prediction, the horizontal component l0=C1 and the trend component With seasonal component s -m+1:0 =S1:S 10 , dynamically update the horizontal component l by recursive calculation of the smoothing coefficient t , trend component b tWith seasonal component s t , superimpose the linear trend and extrapolate the seasonality to generate the channel quality prediction value for the next three time slots A parameter adaptive adjustment mechanism is introduced to enable the channel prediction model to dynamically track changes in channel status. The calculation process is as follows:

[0106] l t =α(C t -s t-m )+(1-α)(l t-1 +b t-1 );

[0107] b t =β(l t -l t-1 )+(1-β)b t-1 ;

[0108] s t =γ(C t -l t )+(1-γ)s t-m ;

[0109]

[0110] Where k = 1, 2, or 3. The channel quality prediction values for the next three time slots are encapsulated in a fixed format and transmitted via an independent low-latency bus. The fixed format includes a timestamp, a three-time slot channel quality prediction value, and a check code. When the input channel quality index sequence exceeds the range [0, 1], the boundary constraint is enabled. When the channel quality index sequence is interrupted for more than 3 time slots, historical data interpolation is started If the deviation between the channel quality prediction value and the measured channel quality for two consecutive times is greater than 0.2, the prediction is switched to the simple moving average prediction.

[0111] S4. Combine the service priority classification results with the channel quality index prediction value to formulate a data transmission strategy and early warning mechanism. Deploy a policy decision unit on the main control board, receive the service priority classification results and channel quality prediction value via a dual-channel bus, integrate the transmission strategy matrix corresponding to priority P∈{0,1,2}×channel quality Q'∈{low,medium,high} and the bandwidth allocation calculation core. Configure a status monitoring module on the communication center side, including a transmission delay timer, a packet loss rate counter, and a channel switching trigger, and connect it to the output bus of the policy execution engine. Add a parameter update coprocessor containing a PID control algorithm unit to receive early warning signals in real time and generate weight adjustment coefficients. These are transmitted back to the transmission decision model and channel prediction model via the reverse bus. Combine the service priority classification results with the channel quality index prediction value to dynamically allocate bandwidth for services of different priorities. The calculation process is as follows:

[0112]

[0113] Among them, P i =0, 1, 2 are low priority, medium priority and highest priority of the business, BW total The bandwidth allocation result is written into the queue management register through the time division multiplexing controller, and the bandwidth allocation weight coefficient is refreshed every 50ms. The maximum allowable delay of voice service is set to 30ms, the maximum allowable delay of video service is set to 80ms, and the maximum allowable delay of file transfer service is set to 500ms. When the end-to-end delay of non-low priority service exceeds the maximum allowable delay of the corresponding service type, it is judged as a delay warning. If the gradient of the channel quality prediction value of three consecutive time slots meets It is determined to be a channel degradation warning, triggering a channel switching instruction, releasing 20% of the current channel resources to seize a new channel, using preemptive weighted polling scheduling, writing a 3-bit priority tag in the IP header TOS field during data encapsulation, and adding 2 bytes of metadata to the transport layer to record the channel quality prediction value. When the channel prediction fails and switches to the moving average mode, the bandwidth allocation weight coefficient is forced to be set to 0.5, and the multipath transmission function is turned off. If the policy execution engine fails, the preset conservative policy BW is enabled. alloc =0.3×BW total ×(P+1), the data path switches to the backup bus.

[0114] The above mainly introduces the solution provided by the embodiment of the present application from the perspective of method. In order to realize the above functions, it includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should easily appreciate that, in combination with the units and algorithm steps of each example described in the embodiments disclosed herein, the present application can be implemented in the form of a combination of hardware and computer software. Whether a function is executed in a hardware or computer software driven hardware manner depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0115] In the embodiment of the present application, the data transmission device can be divided into functional modules according to the above method example. For example, each functional module can be divided according to each function, or two or more functions can be integrated into one processing module. The above integrated modules can be implemented in the form of hardware or software functional modules. Optionally, the division of modules in the embodiment of the present application is schematic and is only a logical functional division. In actual implementation, there may be other division methods.

[0116] Example 3, as Figure 1As shown, based on embodiment 1-2, the present invention provides a technical solution: a data transmission device, including a routing data acquisition module, a routing feature extraction module, a priority prediction module and a transmission warning module;

[0117] Routing data collection module, collects and pre-processes the CPE router's ingress traffic data and wireless channel time slot occupancy status data;

[0118] The routing feature extraction module extracts the transmission urgency and channel quality index during the data transmission process from the pre-processed ingress traffic data and wireless channel time slot occupancy status data;

[0119] The priority prediction module builds a transmission decision model and a channel prediction model based on transmission urgency and channel quality index, and outputs service priority classification results and channel quality index prediction values respectively;

[0120] The transmission warning module combines the service priority classification results with the channel quality index prediction value to formulate data transmission strategies and warning mechanisms.

[0121] An embodiment of the present application further provides a computer-readable storage medium, which includes computer-executable instructions. When the computer-executable instructions are executed on a computer, the computer executes the data transmission method provided in the above embodiment.

[0122] An embodiment of the present application also provides a computer program, which can be directly loaded into a memory and contains software code. After being loaded and executed by a computer, the computer program can implement the data transmission method provided in the above embodiment.

[0123] Those skilled in the art will appreciate that, in one or more of the examples above, the functions described herein can be implemented using hardware, software, firmware, or any combination thereof. When implemented using software, these functions can be stored in a computer-readable medium or transmitted as one or more instructions and codes on the computer-readable medium. Computer-readable media include computer-readable storage media and communication media, wherein communication media includes any medium that facilitates the transmission of computer programs from one location to another. The storage medium can be any available medium that can be accessed by general-purpose and special-purpose computers.

[0124] Through the description of the above implementation methods, technical personnel in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0125] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic, and the division of the modules and units is merely a logical function division. There may be other division methods in actual implementation. For example, multiple units and components can be combined or integrated into another device, and some features can be ignored or not executed. Another point is that the mutual coupling and direct coupling and communication connection shown and discussed can be through some interfaces, indirect coupling and communication connection of devices and units, which can be electrical, mechanical and other forms. The units described as separate components may or may not be physically separated, and the components shown as units may be one physical unit or multiple physical units, that is, they may be located in one place, or they may be distributed in multiple different places. Some or all of the units can be selected according to actual needs to achieve the purpose of the present embodiment.

[0126] In addition, the functional units in the various embodiments of the present application can be integrated into a processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit. If the integrated unit is implemented in the form of a software functional unit and sold and used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application is essentially the part that contributes to the general technology and all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for enabling a device and a processor to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage media include: U disk, mobile hard disk, ROM, RAM, disk and CD-ROM, various media that can store program code.

[0127] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes and substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A CPE router, characterized by: The CPE router includes a traffic sniffing module, a multi-band spectrum analysis module, a feature calculation unit, a quality assessment coprocessor, a transmission decision hardware module, a channel prediction module, a policy decision unit, a parameter update coprocessor and a status detection module; The traffic sniffing module is used to capture and preliminarily process the CPE router inlet traffic data; The multi-band spectrum analysis module is used to monitor the status of wireless channels in real time; The feature calculation unit is used to extract the transmission urgency feature; The quality assessment coprocessor is used to calculate the channel quality index; The transmission decision hardware module is used for service priority classification; The channel prediction module is used for channel quality prediction; The policy decision unit is used to generate a dynamic transmission policy; The parameter update coprocessor is used to dynamically adjust the bandwidth allocation weight coefficient; The status detection module is used to monitor the network status in real time.

2. A data transmission method, characterized in that: Applied to CPE routers, the following steps are included: S1. Collect and pre-process the CPE router's ingress traffic data and wireless channel time slot occupancy status data; S2, extracting the transmission urgency and channel quality index during the data transmission process from the pre-processed ingress traffic data and wireless channel time slot occupancy status data; S3. Based on the transmission urgency and the channel quality index, a transmission decision model and a channel prediction model are constructed to output the service priority classification results and the channel quality index prediction value respectively; S4. Combine the service priority classification results with the channel quality index prediction value to formulate data transmission strategies and early warning mechanisms.

3. A data transmission method according to claim 2, characterized in that: In S1, the collection and preprocessing process of the CPE router's ingress traffic data and wireless channel time slot occupancy status data includes: A traffic sniffing module is deployed at the network interface of the CPE router. This module consists of a multi-port mirroring device and an embedded processing chip, which is physically connected to the data bus of the main control board. The wireless baseband chip integrates a multi-band spectrum analysis module, including 2.4GHz and 5GHz dual radio frequency channels and a signal processing unit. The ingress traffic is replicated through port mirroring technology, and a protocol parser is used to extract the packet quintuple and payload characteristics from the ingress traffic data, thereby identifying voice services, video services, and file transfer services. The packet quintuple includes the source address, destination address, protocol type, source port, and destination port. Perform a full-band fast scan every 100ms to obtain the occupancy status data of each wireless channel time slot and calculate the channel occupancy rate C and effective signal-to-noise ratio S; A mapping rule base is established to mark continuous traffic flows with discrete service labels, where voice services are marked as 1, video services are marked as 2, and file transfer services are marked as 3.

4. A data transmission method according to claim 3, characterized in that: In S2, the process of extracting the transmission urgency and the channel quality index during data transmission includes: A feature calculation unit containing a dedicated register group and a floating-point operation accelerator is added to the embedded processing chip, which is connected to the pre-processed service label data bus. A quality assessment coprocessor containing a sliding window cache and a nonlinear operation unit is integrated into the signal processing unit, which is directly connected to the filtered wireless channel time slot occupancy status data. According to the service type, the statistical period W of voice service, video service and file transfer service is set to 0.5s, 1s and 2s respectively, and the request frequency F of each service type is calculated. t , where the voice service benchmark length is 256 bytes, the video service benchmark length is 1024 bytes, and the file transfer service benchmark length is 2048 bytes; When the request frequency change rate ΔF of two consecutive windows exceeds 30%, the window adaptation mechanism is activated to update the window statistical period W' of each service type; Calculate the channel quality index Q based on the normalized channel occupancy C and the effective signal-to-noise ratio S after nonlinear compression; When a channel switching event is detected, instantaneous quality compensation Q is started adj ; When F t When a division by zero error occurs, set F t =0, if Q exceeds the theoretical range [0,1], the constraint function Q is activated clamp .

5. A data transmission method according to claim 4, characterized in that: In S3, the process of constructing a transmission decision model and outputting a service priority classification result includes: A two-dimensional feature vector is constructed based on service labels and transmission urgency, and continuous transmission urgency is binned by service type. The binning process includes dividing voice services into [0, 6), [6, 12), and [12, +∞), video services into [0, 9), [9, 18), and [18, +∞), and file transfers into [0, 15), [15, 30), and [30, +∞). This binning process implements real-time threshold matching through a hardware comparator array, converting service labels and transmission urgency into discrete inputs that can be processed by the improved C5.0 decision tree algorithm. The transmission decision model is constructed using an improved C5.0 decision tree algorithm. The attribute splitting gain rate is used as the node splitting criterion to obtain the ratio of information gain to the amount of information in the split. A dynamic pruning strategy is used to control the complexity of the transmission decision model. Splitting is terminated when the number of node samples is less than 1% of the total. An online incremental learning mechanism is used in the training phase of the transmission decision model, and the channel quality index is used to dynamically update the binning threshold. A three-layer tree structure is adopted. The first layer branches according to the service type. The second layer splits the child nodes according to the transmission urgency bin threshold. The leaf nodes are mapped to priority labels, which include highest priority, medium priority and low priority. Voice service and video service are classified as the highest priority when the transmission urgency exceeds 12 and 18 respectively. Voice service, video service and file transfer service are classified as medium priority when the transmission urgency is in the range of [6,12), [9,18) and [15,+∞) respectively. Voice service, video service and file transfer service are classified as low priority when the transmission urgency is in the range of [0,6), [0,9) and [0,15) respectively. Three groups of comparators are deployed to synchronously process each business type branch. The first level of the comparator completes business type matching, the second level performs transmission urgency threshold interval comparison, and the third level outputs a binary-coded priority label. When the input request frequency is an illegal value, it is classified as the lowest priority. When the business label is not recognized, it is processed as a file transfer business by default.

6. A data transmission method according to claim 5, characterized in that: In S3, the process of constructing a channel prediction model and outputting a channel quality index prediction value includes: Taking the channel quality index as input, a hybrid prediction model architecture of time series decomposition and exponential smoothing is adopted to construct a channel prediction model; In the time series decomposition stage, the channel quality index sequence Q t Decomposed into trend component C t , seasonal component S t and the residual component R t ,The trend component is extracted using a sliding window local weighted regression algorithm with ,the window width set to 10 time slots. The seasonal component is ,separated by the periodic difference method with a preset period length of 10 time slots. The residual ,component is obtained by subtracting the trend component and the seasonal ,component from the channel quality index; In the initialization phase of exponential smoothing forecast, the horizontal component l0=C1 and the trend component With seasonal component s -m+1:0 =S1:S 10 , dynamically update the horizontal component l by recursive calculation of the smoothing coefficient t , trend component b t With seasonal component s t , superimpose the linear trend and extrapolate the seasonality to generate the channel quality prediction value for the next three time slots And introduce the parameter adaptive adjustment mechanism to enable the channel prediction model to dynamically track the changes in channel status; Encapsulating the predicted channel quality value for the next three time slots in a fixed format and transmitting it via an independent low-latency bus, wherein the fixed format includes a timestamp, the predicted channel quality value for the three time slots, and a check code; When the input channel quality index sequence exceeds the range [0,1], the boundary constraint is enabled. When the channel quality index sequence is interrupted for more than 3 time slots, the historical data interpolation Q t If the deviation between the channel quality prediction value and the measured channel quality for two consecutive times is greater than 0.2, the prediction is switched to the simple moving average prediction.

7. A data transmission method according to claim 6, characterized in that: In S4, the process of formulating a data transmission strategy and an early warning mechanism by combining the service priority classification result and the channel quality index prediction value includes: A policy decision unit is deployed on the main control board. It receives service priority classification results and channel quality prediction values through a dual-channel bus, and integrates the transmission policy matrix corresponding to priority P∈{0,1,2}×channel quality Q'∈{low,medium,high} with the bandwidth allocation calculation core. Combined with the service priority classification results and the channel quality index prediction value, the bandwidth of different priority services is dynamically allocated. The bandwidth allocation results are written into the queue management register through the time division multiplexing controller, and the bandwidth allocation weight coefficient is refreshed every 50ms. The maximum permissible delay for voice services is set to 30ms, the maximum permissible delay for video services is set to 80ms, and the maximum permissible delay for file transfer services is set to 500ms. When the end-to-end delay of non-low priority services exceeds the maximum permissible delay of the corresponding service type, it is determined to be a delay warning. If the gradient of the channel quality prediction value decreases for three consecutive time slots and meets the If it is judged as a channel degradation warning, a channel switching instruction is triggered, releasing 20% of the current channel resources to occupy a new channel; Preemptive weighted round-robin scheduling is adopted. A 3-bit priority tag is written into the TOS field of the IP header during data encapsulation. The transport layer adds 2 bytes of metadata to record the channel quality prediction value. When the channel prediction fails and switches to the moving average mode, the bandwidth allocation weight coefficient is forcibly set to 0.5, and the multi-path transmission function is disabled. If the policy execution engine fails, the preset conservative policy is enabled and the data path is switched to the backup bus.

8. A data transmission device, characterized in that: The data transmission device includes a routing data acquisition module, a routing feature extraction module, a priority prediction module and a transmission early warning module; The routing data acquisition module collects and pre-processes the ingress traffic data and wireless channel time slot occupancy status data of the CPE router; The routing feature extraction module extracts the transmission urgency and channel quality index during the data transmission process from the pre-processed inlet traffic data and the wireless channel time slot occupancy status data; The priority prediction module constructs a transmission decision model and a channel prediction model based on the transmission urgency and the channel quality index, and outputs the service priority classification result and the channel quality index prediction value respectively; The transmission warning module formulates a data transmission strategy and warning mechanism by combining the service priority classification result and the channel quality index prediction value.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium includes computer-executable instructions, and when the computer-executable instructions are executed on a computer, the computer is enabled to execute the data transmission method according to any one of claims 5 to 7.

10. A computer program product, characterized in that The computer program product comprises: a computer program and instructions, and when the computer program and instructions are run on a computer, the computer is enabled to execute the data transmission method according to any one of claims 5 to 7.

Images