Method and related device for diagnosing and early warning of home broadband network quality

Abstract

The application provides a home broadband network quality diagnosis and early warning method and related device. The home broadband network quality diagnosis and early warning method comprises the following steps: collecting network quality indexes, obtaining network quality information according to the network quality indexes, and the network quality information comprises network quality index values of at least one dimension; determining a current time perception baseline according to the network quality information of the current time and a previous time perception baseline; the perception baseline reflects a level of network quality information expected by a user; the perception baseline comprises perception baseline values of at least one dimension, and the dimension of the perception baseline is the same as the dimension of the network quality information; obtaining a current time perception significance index according to the network quality information of the current time and the current time perception baseline; determining a current time dynamic threshold according to the current time perception baseline; and determining whether to trigger an early warning according to the current time perception significance index and the current time dynamic threshold.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of network operation and maintenance, and in particular to a method and related apparatus for diagnosing and warning of home broadband network quality. Background Technology

[0002] In terms of home broadband network quality diagnosis and early warning, existing technologies mostly use static Key Quality Indicator (KQI) thresholds or static Key Performance Indicator (KPI) thresholds to obtain a fixed mapping model for perception scores. This can easily lead to inconsistencies between the early warning triggering conditions of home broadband networks and the user's actual perception experience, i.e., the early warning triggering conditions are not synchronized with the user's dynamically changing perception sensitivity. Summary of the Invention

[0003] This invention provides a method for diagnosing and warning about the quality of home broadband networks, in order to solve the problem that the warning triggering conditions of home broadband networks are not synchronized with the user's perception sensitivity to dynamic changes.

[0004] To solve the above-mentioned technical problems, the present invention is implemented as follows:

[0005] In a first aspect, embodiments of the present invention provide a method for collecting network quality indicators and obtaining network quality information based on the network quality indicators, wherein the network quality information includes network quality indicator values ​​of at least one dimension.

[0006] The current perception baseline is determined based on the network quality information at the current moment and the perception baseline at the previous moment; the perception baseline reflects the level of the network quality information expected by the user; the perception baseline includes a perception baseline value of at least one dimension, and the dimension of the perception baseline is the same as the dimension of the network quality information.

[0007] Based on the network quality information at the current moment and the perception baseline at the current moment, the perception salience index at the current moment is obtained; the perception salience index reflects the degree to which the user can perceive changes in the network quality information.

[0008] Based on the current perception baseline, determine the dynamic threshold at the current moment;

[0009] Based on the perceived saliency index at the current moment and the dynamic threshold at the current moment, determine whether to trigger an early warning.

[0010] Optionally, determining the current sensing baseline based on the current network quality information and the previous sensing baseline includes:

[0011] Based on the network quality information at the current moment and the perception baseline at the previous moment, determine whether to update the perception baseline at the previous moment;

[0012] When determining to update the perception baseline of the previous moment, the intermediate perception baseline of the current moment is calculated based on the network quality information of the current moment and the perception baseline of the previous moment.

[0013] Based on the context category of the user-driven business at the current moment, the intermediate perception baseline at the current moment is adjusted to obtain the perception baseline at the current moment.

[0014] Optionally, determining whether to update the perception baseline at the previous moment based on the network quality information at the current moment and the perception baseline at the previous moment includes:

[0015] Calculate the absolute value of the deviation between the network quality information at the current moment and the perception baseline at the previous moment in each dimension, and obtain the maximum value of the absolute value of the deviation in each dimension;

[0016] The maximum value is compared with a preset noise threshold. If the maximum value is less than the noise threshold, the perception baseline of the previous moment is determined to be the perception baseline of the current moment.

[0017] If the maximum value is greater than or equal to the noise threshold, it is determined that the perception baseline of the previous moment should be updated.

[0018] Optionally, when determining to update the perception baseline of the previous moment, calculating the intermediate perception baseline of the current moment based on the network quality information of the current moment and the perception baseline of the previous moment includes:

[0019] The intermediate perception baseline at the current moment is obtained based on the network quality information at the current moment, the perception baseline at the previous moment, and the personalized adaptation rate parameter.

[0020] The personalized adaptation rate parameter is determined offline based on the user's historical complaint records and the network quality information for the corresponding time period, and is used to control the rate at which the perception baseline at the current moment converges to the network quality information at the current moment.

[0021] Optionally, adjusting the intermediate perception baseline at the current moment based on the context category of the user-dominated service at the current moment to obtain the perception baseline at the current moment includes:

[0022] Identify the user's dominant business type and classify the dominant business into a preset scenario category to obtain the user's dominant business scenario category;

[0023] Select the corresponding context bias index according to the dominant business context category, and perform low-pass filtering on the context bias index to obtain a smoothed context bias index. The context bias index represents the adjustment of the intermediate perception baseline under the preset context category. The dimension of the context bias index is the same as the dimension of the network quality information.

[0024] The smoothed context bias index is superimposed on the intermediate perception baseline at the current moment to obtain the perception baseline at the current moment.

[0025] Optionally, based on the network quality information at the current moment and the perception baseline at the current moment, the perception saliency index at the current moment is obtained, including:

[0026] Obtain the relative deviation of each dimension between the network quality information at the current moment and the perception baseline at the current moment;

[0027] The weighting coefficients of the relative deviation of each dimension are determined based on the dominant business scenario category at the current moment;

[0028] The relative deviations of each dimension are weighted by the weighting coefficients of the relative deviations of each dimension to obtain the weighted relative deviations of each dimension between the network quality information at the current time and the perception baseline at the current time.

[0029] The maximum value of the weighted relative deviation of each dimension is taken as the perceptual saliency index at the current moment.

[0030] Optionally, based on the current sensing baseline, the dynamic threshold for the current moment is determined, including:

[0031] Obtain the average value of each dimension of the perception baseline at the current moment;

[0032] The dynamic threshold at the current moment is calculated based on the preset basic threshold constant, the proportional coefficient, and the average value. The basic threshold constant and the proportional coefficient are set and adjusted periodically based on the statistical analysis results of user historical complaint data.

[0033] Optionally, determining whether to trigger an early warning based on the perceived saliency index at the current moment and the dynamic threshold at the current moment includes:

[0034] Determine whether the perceived saliency index at the current moment is greater than or equal to the dynamic threshold at the current moment. If the perceived saliency index at the current moment is greater than or equal to the dynamic threshold at the current moment and the direction of change of the dimension index with the maximum weighted relative deviation of each dimension is the direction of deterioration, then trigger an early warning.

[0035] The rule for determining the direction of degradation is as follows: when the dimension index with the maximum weighted relative deviation of each dimension is a positive index, if the value of the dimension that contributes the most in the perception significance index at the current moment is less than the value of the dimension corresponding to the perception baseline at the current moment, then it is considered degradation.

[0036] When the dimension index with the maximum weighted relative deviation of each dimension is a negative index, and the value of the dimension that contributes the most in the perception significance index at the current moment is greater than the value of the dimension corresponding to the perception baseline at the current moment, then it is considered deterioration.

[0037] Secondly, embodiments of the present invention provide a home broadband network quality diagnosis and early warning device, comprising:

[0038] The data acquisition module is used to collect home broadband network quality indicators and obtain network quality information at the current moment based on the home broadband network quality indicators. The network quality information includes network quality indicators of at least one dimension.

[0039] The sensing baseline calculation module is used to calculate the sensing baseline at the current moment based on the network quality information at the current moment and the sensing baseline at the previous moment.

[0040] The perception saliency index calculation module is used to obtain the perception saliency index at the current moment based on the network quality information at the current moment and the perception baseline at the current moment.

[0041] The dynamic threshold calculation module is used to determine the dynamic threshold at the current moment based on the current sensing baseline.

[0042] The early warning decision module is used to execute early warning decisions and output the results.

[0043] Thirdly, embodiments of the present invention provide an electronic device, including: a processor, a memory, and a program stored in the memory and executable on the processor, wherein when the program is executed by the processor, it implements the steps of the home broadband network quality diagnosis and early warning method as described in the first aspect above.

[0044] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the home broadband network quality diagnosis and early warning method described in the first aspect above.

[0045] Fifthly, embodiments of the present invention provide a computer program product, including computer instructions, which, when executed by a processor, implement the steps of the home broadband network quality diagnosis and early warning method described in the first aspect above.

[0046] In this embodiment of the invention, network quality information is obtained based on collected network quality indicators, the network quality information including network quality indicator values ​​of at least one dimension; a perception baseline is determined based on the network quality information at the current moment and the perception baseline at the previous moment; the perception baseline reflects the level of network quality information expected by the user; the perception baseline includes a perception baseline value of at least one dimension, the dimension of the perception baseline being the same as the dimension of the network quality information; a perception saliency index is obtained based on the network quality information at the current moment and the perception baseline at the current moment; the perception saliency index reflects the degree to which the user can perceive changes in the network quality information; a dynamic threshold is determined based on the perception baseline at the current moment; and a warning is determined based on the perception saliency index and the dynamic threshold at the current moment. In this way, by dynamically determining the perception baseline at the current moment, the system reflects the user's psychological expectations of network quality indicators in real time, ensuring the consistency between the warning triggering conditions and the user's actual psychological expectations; by dynamically determining the perception salience index at the current moment, the system reflects the degree of deviation of the network quality indicators from the user's psychological expectations of network quality indicators at the current moment and in the current context; and by determining the dynamic threshold, which is dynamically adjusted according to the user's perception state, the system improves the accuracy of home broadband network warnings in different network environments. Attached Figure Description

[0047] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0048] Figure 1 A flowchart illustrating a home broadband network quality diagnosis and early warning method provided in an embodiment of the present invention;

[0049] Figure 2 This is a schematic diagram of a home broadband network quality diagnosis and early warning device provided in an embodiment of the present invention;

[0050] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] Please refer to Figure 1 This invention provides a method for diagnosing and warning about the quality of a home broadband network. This method can be executed by a home gateway or an access network edge computing node, and includes the following steps:

[0053] Step 101: Collect network quality indicators and obtain network quality information based on the network quality indicators. The network quality information includes network quality indicator values ​​of at least one dimension.

[0054] The aforementioned network quality indicators may include at least one of the following: downlink speed, network latency, latency jitter, packet loss rate, etc. Optionally, these network quality indicators can be obtained by deploying a lightweight data collection probe on the home gateway and continuously collecting data at a preset period. Downlink speed is a positive indicator; a higher value indicates better network quality. Network latency, latency jitter, and packet loss rate are all negative indicators; lower values ​​indicate better network quality.

[0055] In some embodiments, to facilitate unified calculation, the collected network quality indicators are normalized, that is, mapped to the [0,1] interval. For positive indicators such as downlink rate, the normalized value is the ratio of the actual downlink rate to the user's contracted bandwidth. If the actual downlink rate exceeds the contracted bandwidth, the normalized value is 1. For negative indicators such as network latency, latency jitter, and packet loss rate, the normalized value is the ratio of the actual value to a preset reference upper limit value. If the actual value exceeds the reference upper limit value, the normalized value is 1. The reference upper limit value of each indicator can be preset based on network operation and maintenance experience.

[0056] In some embodiments, the network quality information may optionally be represented in a vector format.

[0057] In some embodiments, the normalized network quality indicators can be arranged in a fixed order to form a network quality information vector. For example, the network quality information vector is represented by the following formula:

[0058] ;

[0059] in, This is the normalized value for the downlink rate. This is the normalized value of network latency. This is the normalized value for delay jitter. This is the normalized value for packet loss rate. Indicates the current moment.

[0060] By using vector form to express network quality, the subsequent evolution of the sensing baseline and the calculation of relative deviation can be transformed into operations between vectors, resulting in low computational overhead and making it suitable for deployment on home gateway devices.

[0061] Step 102: Determine the current perception baseline based on the network quality information at the current moment and the perception baseline at the previous moment; the perception baseline reflects the level of the network quality information expected by the user; the perception baseline includes a perception baseline value of at least one dimension, and the dimension of the perception baseline is the same as the dimension of the network quality information.

[0062] In some embodiments, the sensing baseline may optionally be represented as a vector.

[0063] Step 103: Based on the network quality information at the current moment and the perception baseline at the current moment, obtain the perception salience index at the current moment; the perception salience index reflects the degree to which the user can perceive changes in the network quality information.

[0064] The Perceived Saliency Index (PSI) in this embodiment of the invention is used to quantitatively assess the degree to which the current network quality deviates from the current perceived baseline, reflecting the extent to which the user is likely to perceive the deviation. A higher PSI value indicates that the user is more likely to perceive the deviation of the current network quality from the current perceived baseline.

[0065] Step 104: Determine the dynamic threshold at the current moment based on the current sensing baseline;

[0066] Dynamic threshold in the embodiments of the present invention This reflects the differentiated tolerance of users to network fluctuations under different network environments. Specifically, when users are in a consistently high-quality network environment, their tolerance for minor fluctuations is relatively high. A larger fluctuation makes it less likely to trigger an alert; conversely, when users are in a poor network environment, their tolerance for small fluctuations is relatively low. Smaller sizes are more likely to trigger warnings.

[0067] Step 105: Determine whether to trigger an early warning based on the perception saliency index at the current moment and the dynamic threshold at the current moment.

[0068] In this embodiment of the invention, network quality information is obtained based on collected network quality indicators, the network quality information including network quality indicator values ​​of at least one dimension; a perception baseline is determined based on the network quality information at the current moment and the perception baseline at the previous moment; the perception baseline reflects the level of network quality information expected by the user; the perception baseline includes a perception baseline value of at least one dimension, the dimension of the perception baseline being the same as the dimension of the network quality information; a perception saliency index is obtained based on the network quality information at the current moment and the perception baseline at the current moment; the perception saliency index reflects the degree to which the user can perceive changes in the network quality information; a dynamic threshold is determined based on the perception baseline at the current moment; and a warning is determined based on the perception saliency index and the dynamic threshold at the current moment. In this way, by dynamically determining the perception baseline at the current moment, the system reflects the user's psychological expectations of network quality indicators in real time, ensuring the consistency between the warning triggering conditions and the user's actual psychological expectations; by dynamically determining the perception salience index at the current moment, the system reflects the degree of deviation of the network quality indicators at the current moment and in the current context from the user's psychological expectations of network quality indicators; and by determining the dynamic threshold, which is dynamically adjusted according to the user's perception state, the system improves the accuracy of home broadband network warnings in different network environments.

[0069] In some embodiments, optionally, determining the current sensing baseline based on the current network quality information and the previous sensing baseline includes:

[0070] Step 201: Based on the network quality information at the current moment and the perception baseline at the previous moment, determine whether to update the perception baseline at the previous moment;

[0071] Specifically, the network environment contains numerous transient and irregular micro-fluctuations, such as millisecond-level latency jitter caused by sudden traffic bursts and rate fluctuations caused by transient interference with wireless signals. These transient and irregular micro-fluctuations are usually not noticeable to users. If the perception baseline is frequently updated to follow these transient fluctuations, it will lead to instability in the perception baseline value, thereby affecting the reliability of subsequent early warning judgments. Therefore, based on the network quality information at the current moment and the perception baseline at the previous moment, it is determined whether the network fluctuation at the current moment is a transient and irregular micro-fluctuation. If it is such a transient and irregular micro-fluctuation, the perception baseline is not updated; if the network fluctuation at the current moment is a continuous and significant network fluctuation, the perception baseline is updated.

[0072] Step 202: When determining to update the perception baseline of the previous moment, calculate the intermediate perception baseline of the current moment based on the network quality information of the current moment and the perception baseline of the previous moment.

[0073] Specifically, the intermediate perception baseline is calculated jointly based on the network quality information vector at the current moment and the perception baseline at the previous moment, causing the perception baseline to be gradually adjusted towards the network quality index at the current moment. This gradual adjustment simulates the user's habituation process to the network environment, meaning that the user's psychological expectations will not change drastically immediately due to a single network fluctuation, but will gradually transition to a new level.

[0074] Step 203: Adjust the intermediate perception baseline at the current moment according to the scenario category of the user-led business at the current moment to obtain the perception baseline at the current moment.

[0075] Specifically, since different user-dominated business types have significantly different sensitivities to various dimensions of network quality information vector indicators, and users have different psychological expectations of network quality when engaging in different activities, the intermediate perception baseline at the current moment is adjusted according to the context category of the user-dominated business at the current moment to obtain the perception baseline at the current moment.

[0076] In this embodiment of the invention, by ignoring instantaneous and irregular minor network fluctuations, the stability of the perception baseline is effectively ensured, and the reliability of subsequent early warning judgments is improved, thus avoiding frequent changes in the perception baseline due to such minor fluctuations. Simultaneously, by progressively adjusting the intermediate perception baseline towards the current network state, the rhythm of change in the perception baseline is kept consistent with the evolution of user expectations, avoiding drastic fluctuations in the perception baseline caused by a single network fluctuation. Furthermore, the intermediate perception baseline is differentiated according to the scenario category of the user's dominant business at the current moment, ensuring that the perception baseline matches the perception requirements of the actual business scenario, thereby improving the accuracy of subsequent early warning judgments.

[0077] In some embodiments, optionally, based on the network quality information at the current moment and the perception baseline at the previous moment, it is determined whether to update the perception baseline at the previous moment. This determination can be performed through a noise filtering layer, and the specific steps include:

[0078] Step 301: Calculate the absolute value of the deviation between the network quality information at the current moment and the perception baseline at the previous moment in each dimension, and obtain the maximum value of the absolute value of the deviation in each dimension.

[0079] The above maximum value is denoted as Specifically, this can be achieved using the following formula:

[0080] ;

[0081] In the formula, Represents the sensing baseline vector In the The numerical value of the dimension. Represents a network quality information vector In the The numerical value of the dimension. `max` indicates taking... The maximum value in.

[0082] Step 302: Compare the maximum value with a preset noise threshold. If the maximum value is less than the noise threshold, determine that the perception baseline of the previous moment is used as the perception baseline of the current moment. If the maximum value is greater than or equal to the noise threshold, determine that the perception baseline of the previous moment is updated.

[0083] Preset noise threshold This indicates that its value is determined by referencing the fluctuation statistics of historical network quality indicators collected by home gateways or edge computing nodes of access networks.

[0084] In some implementations, the noise threshold The specific value can be selected from the high quantile of the minute-level change statistics of each dimension of the network quality information vector during the initial period of user network access.

[0085] when If the network fluctuation at the current moment is considered small, such small fluctuations are usually not noticeable to the user and are considered transient noise. The noise filtering layer will filter out this type of transient noise, so the perception baseline will not be updated in this instance. Proceed to step 103 to complete the calculation of the Perceived Significance Index (PSI).

[0086] when If the current network fluctuations are deemed to have reached a level worthy of attention, the perception baseline will be updated.

[0087] In this embodiment of the invention, numerous transient and irregular micro-fluctuations exist in the network environment, such as millisecond-level latency jitter caused by sudden traffic surges and rate fluctuations caused by transient interference with wireless signals. The noise filtering layer filters out these transient noises that do not cause explicit user perception, ensuring that the perception baseline only responds to network fluctuations with a certain degree of persistence and significance. This makes the evolution of the perception baseline more stable and ensures the reliability of subsequent early warning judgments. Specifically, by introducing a discrimination method based on maximum dimensionality deviation and a noise threshold, effective filtering of transient noise is achieved. By setting a noise threshold to distinguish between transient fluctuations and network fluctuations of significant importance, if the network fluctuation amplitude is less than the noise threshold, the perception baseline is not updated. This effectively suppresses the interference of transient fluctuations on the perception baseline, ensuring that the perception baseline only responds to network fluctuation changes with substantial impact. This improves the stability and robustness of the perception baseline evolution and provides a reliable reference benchmark for the subsequent calculation of the perception significance index.

[0088] In some embodiments, optionally, when determining to update the perception baseline of the previous time step, the intermediate perception baseline of the current time is calculated based on the network quality information of the current time step and the perception baseline of the previous time step. The calculation of the intermediate perception baseline of the current time step can be performed through the basic adaptation layer, specifically including:

[0089] The intermediate perception baseline at the current moment is obtained based on the network quality information at the current moment, the perception baseline at the previous moment, and the personalized adaptation rate parameter.

[0090] The personalized adaptation rate parameter is determined offline based on the user's historical complaint records and the network quality information for the corresponding time period, and is used to control the rate at which the perception baseline at the current moment converges to the network quality information at the current moment.

[0091] The aforementioned intermediate sensing baseline can be represented in vector form, denoted as . ;

[0092] In some implementations... The exponential smoothing method can be used for calculation, specifically expressed by the following formula:

[0093] ;

[0094] in, This is the sensing baseline vector from the previous moment; This is the network quality information vector at the current moment; The deviation vector of the network quality information vector at the current moment from the perceived baseline vector at the previous moment represents the gap between the objective network quality at the current moment and the network quality expected by the user at the current moment. The personalized adaptation rate parameter takes a real number between 0 and 1. The smaller the value, the slower the perceived baseline changes, indicating that the user is slower to adapt to network fluctuations, meaning they are more sensitive to changes in network quality. The higher the value, the faster the perceived baseline changes, indicating that the user is more likely to accept the new network level as normal.

[0095] In other implementations, Alternatively, the moving average method can be used to obtain the data, and by adjusting the size of the sliding window, the sensing baseline can be gradually converged to the network quality index at the current moment.

[0096] In some implementations, the personalized adaptation rate parameter This can be determined through the following steps:

[0097] Step 401: For each complaint, trace back the network quality information vector for a period of time prior to the complaint;

[0098] Step 402: Set a candidate The values ​​are used to simulate the evolution of the perception baseline during this time period and to calculate the perception saliency index at each time point.

[0099] Step 403: With the objective of minimizing the time difference between the moment when the perceived significance index exceeds the dynamic threshold and the moment when the user actually files a complaint, a parameter optimization method is used to determine the user's optimal [time / time]. For new users with no historical complaint records, the global statistical average can be used as the initial value, and personalized adjustments can be made after sufficient data has been accumulated.

[0100] In this embodiment of the invention, a basic adaptation layer is used to gradually adjust the evolution of the current perception baseline towards the actual changes in the network quality index at the current moment. Specifically, an exponential smoothing method is used to gradually converge the perception baseline towards the network quality index at the current moment, avoiding abrupt changes in the perception baseline due to network fluctuations, and simulating the gradual adjustment process of user psychological expectations as the network environment changes. The optimal time difference between the moment when the model determines that the perception significance index exceeds the dynamic threshold and the moment when the user actually complains is minimized, and the optimal time difference is determined by combining parameter optimization methods. The value is adjusted so that the convergence speed of the perception baseline matches the user's sensitivity to network fluctuations, thereby improving the accuracy of the perception baseline in representing the user's subjective experience.

[0101] In some embodiments, optionally, the step of adjusting the intermediate perception baseline at the current moment based on the context category of the user-dominated service at the current moment to obtain the perception baseline at the current moment can be performed by a context modulation layer, and the specific steps include:

[0102] Step 501: Identify the user's dominant business type and classify the dominant business into a preset scenario category to obtain the user's dominant business scenario category;

[0103] The identification methods for the dominant services of the aforementioned users may include: identifying them by analyzing the protocol characteristics, connection modes, and packet length distribution of the current traffic based on the application identification capabilities of the home gateway or access network device; inferring them by analyzing the time characteristics of the traffic, such as connection duration, data packet arrival interval distribution, and uplink / downlink traffic ratio; and identifying them by using application information reported by the user terminal.

[0104] The aforementioned preset scenario categories can include real-time interactive, streaming media, and best-effort scenarios. Different scenario categories have different sensitivities to the various dimensions of the network quality information vector. Among them, the real-time interactive scenario category has higher requirements for network latency and latency jitter, the streaming media scenario category has certain requirements for downlink speed but a relatively high tolerance for latency jitter, and the best-effort scenario category has relatively relaxed requirements for the various indicators of the network quality information vector.

[0105] Step 502: Select the corresponding scenario bias index according to the dominant business scenario category, and perform low-pass filtering on the scenario bias index to obtain a smoothed scenario bias index. The scenario bias index represents the adjustment of the intermediate perception baseline under the preset scenario category. The dimension of the scenario bias index is the same as the dimension of the network quality information.

[0106] The aforementioned context bias metrics can be expressed in vector form, i.e., context bias vector. The indicators for each dimension in the context bias vector represent the adjustments made to the intermediate perception baseline under the user-dominated business context category at the current moment. A negative value indicates that the user has higher expectations for that dimension, and the value of that dimension in the intermediate perception baseline should be lowered; a positive value indicates that the user has higher tolerance for that dimension, and the value of that dimension in the intermediate perception baseline can be increased.

[0107] The above-mentioned selection of corresponding scenario bias indicators based on the dominant business scenario categories specifically involves pre-setting a network quality information vector for each dominant business scenario category. The context bias vector template of the same dimension can be set based on the specific values ​​of each dimension in the context bias vector template, either by setting it according to operation and maintenance experience or by collecting network satisfaction evaluation data of users in different dominant business scenarios and performing parameter fitting.

[0108] The scenario bias index after low-pass filtering is expressed in vector form, denoted as the smoothing bias vector. Specifically, it is calculated using the following formula:

[0109] ;

[0110] in, The smoothing coefficient ranges from 0 to 1; t-1 represents the previous time step. This is the current context bias vector.

[0111] Step 503: Superimpose the smoothed context bias index onto the intermediate perception baseline at the current moment to obtain the perception baseline at the current moment.

[0112] The resulting perception baseline vector is described above. Specifically, it can be calculated using the following formula:

[0113] ;

[0114] in, This is the intermediate sensing baseline vector. Let t be the smooth bias vector, and t be the current time.

[0115] In this embodiment of the invention, since different service types have significantly different sensitivities to various dimensions of the network, and users' psychological expectations of network quality also differ when engaging in different activities, a context modulation layer is set up to adjust the perception baseline in conjunction with the current user's dominant service context. This allows the perception baseline to reflect the user's psychological expectations of network quality under different service contexts. Specifically, the intermediate perception baseline is adjusted using the context bias vector corresponding to the user's dominant service type at the current moment. That is, considering the differences in sensitivity of different service context categories to various dimensions of the network quality information vector, the intermediate perception baseline value is lowered for dimensions with higher requirements in the network quality information vector to improve detection sensitivity, and the intermediate perception baseline value is raised for dimensions with higher tolerance to relax the judgment criteria. This achieves dynamic adaptation of the perception baseline to different user-dominated service categories, solving the problem that a single perception baseline cannot take into account multiple user-dominated service categories. In addition, by performing low-pass filtering on the selected context bias vector, the change of the perception baseline during service context switching is made smoother, avoiding abrupt changes in the perception baseline value when switching dominant service contexts, making the warning judgment result more consistent with the user's experience in the actual service context.

[0116] In some embodiments, optionally, the perception saliency index at the current moment is obtained based on the network quality information at the current moment and the perception baseline at the current moment, including:

[0117] Obtain the relative deviation of each dimension between the network quality information at the current moment and the perception baseline at the current moment;

[0118] The weighting coefficients of the relative deviation of each dimension are determined based on the dominant business scenario category at the current moment;

[0119] The relative deviations of each dimension are weighted by the weighting coefficients of the relative deviations of each dimension to obtain the weighted relative deviations of each dimension between the network quality information at the current time and the perception baseline at the current time.

[0120] The maximum value of the weighted relative deviation of each dimension is taken as the perceptual saliency index at the current moment.

[0121] Specifically, the perceived significance index PSI can be expressed by the following formula:

[0122] ;

[0123] in, Represents the network quality information vector at the current moment. In the The value that a dimension can take; Represents the current sensing baseline vector In the The value that a dimension can take; Indicates the first The absolute deviation between the state of the network above the perception baseline; It is a very small positive real number, used to prevent the formula from having a denominator. The case where it is zero; Indicates the first The importance weight of each dimension; This means that the final perceived significance index is determined by aggregating the maximum values, i.e., taking the maximum value of the weighted relative deviations of all dimensions.

[0124] In some embodiments, weight The relative weights are determined based on the current user-driven service scenario. For example, in real-time interactive scenarios, network latency and latency jitter can be assigned higher weights, while downlink speed can be assigned lower weights. The specific numerical calibration can be determined by optimization based on expert experience or data analysis.

[0125] In some embodiments, the maximum value aggregation method can be replaced by weighted summation or Euclidean distance. Different aggregation strategies are selected according to the actual application scenario, and the sensitivity to multi-dimensional synchronous degradation will vary.

[0126] In this embodiment of the invention, the calculation of the perceived saliency index uses a relative deviation, reflecting the idea of ​​Weber's Law, which states that the relative rather than the absolute amount of deviation is the main factor determining user perceptibility. For the same absolute deviation value, the relative deviation is smaller when the network quality is better, and larger when the network quality is poor. This allows the perceived saliency index to more realistically reflect changes in users' subjective feelings, improving the accuracy of perceptual change identification. Dynamically adapting weights based on the context category of the dominant business highlights key sensitive dimensions in specific business contexts, improving the consistency between the perceived saliency index and the user's actual experience. Calculating the perceived saliency index using maximum value aggregation prevents significantly degraded dimensions from being ignored by other non-degraded dimensions, ensuring effective capture of significantly degraded dimensions. Introducing a minimal positive real number ε to prevent it from splitting to zero ensures numerical stability and robustness in actual engineering deployments, applicable to the full range of values ​​for various network quality parameters.

[0127] In some embodiments, optionally, determining the dynamic threshold at the current moment based on the current sensing baseline includes:

[0128] Obtain the average value of each dimension of the perception baseline at the current moment;

[0129] The dynamic threshold at the current moment is calculated based on the preset basic threshold constant, the proportional coefficient, and the average value. The basic threshold constant and the proportional coefficient are set and adjusted periodically based on the statistical analysis results of user historical complaint data.

[0130] The above dynamic threshold This can be expressed by the following formula:

[0131] ;

[0132] in, It is the basic threshold constant, used to represent the minimum relative deviation perceived by the user under poor network quality conditions; The arithmetic mean of the values ​​of each dimension of the current perceived baseline vector reflects the overall network quality level that the user is accustomed to at the current moment. The larger the value, the higher the overall network quality for the user; This is a scaling factor used to control the influence of the arithmetic mean of the values ​​of each dimension of the current perception vector on the current dynamic threshold.

[0133] In this embodiment of the invention, the dynamic threshold changes in real time with the perception baseline vector, positively correlated with the user's current accustomed network quality level. Higher network quality results in a higher dynamic threshold, indicating greater user tolerance for network fluctuations; conversely, lower network quality leads to a lower dynamic threshold, making users more sensitive to network fluctuations. This mechanism of real-time dynamic threshold change with the perception baseline avoids false alarms in high-quality network environments and missed alarms in low-quality network environments, ensuring that the warning triggering conditions match the user's actual psychological expectations and improving the accuracy of network quality warnings.

[0134] In some embodiments, optionally, determining whether to trigger an early warning based on the perceived saliency index at the current moment and the dynamic threshold at the current moment includes:

[0135] Determine whether the perceived saliency index at the current moment is greater than or equal to the dynamic threshold at the current moment. If the perceived saliency index at the current moment is greater than or equal to the dynamic threshold at the current moment and the direction of change of the dimension index with the maximum weighted relative deviation of each dimension is the direction of deterioration, then trigger an early warning.

[0136] The rules for determining the direction of degradation are as follows: when the dimension index with the maximum weighted relative deviation of each dimension is a positive index, the value of the dimension that contributes the most to the perception saliency index at the current moment is less than the value of the dimension corresponding to the perception baseline at the current moment, then it is considered degradation; when the dimension index with the maximum weighted relative deviation of each dimension is a negative index, the value of the dimension that contributes the most to the perception saliency index at the current moment is greater than the value of the dimension corresponding to the perception baseline at the current moment, then it is considered degradation.

[0137] The specific rules for judging the direction of degradation are as follows: For positive indicators such as downlink rate, if the perceived significance index at the current moment is less than the perceived baseline at the current moment, then it is considered degradation; for negative indicators such as network latency, latency jitter, and packet loss rate, if the perceived significance index at the current moment is greater than the perceived baseline at the current moment, then it is considered degradation.

[0138] If an alert is triggered, the alert system generates an alert record, which includes the current timestamp, user ID, PSI value, and dynamic threshold. Numerical value, degradation dimension name, currently identified business scenario category, and exceeding the PSI threshold. The system classifies warning levels based on their severity, and the generated warning records are pushed to the operation and maintenance management system or customer service system for operation and maintenance personnel to proactively intervene and handle.

[0139] If no warning is triggered, all data within this collection period, including the network quality information vector at the current moment, the perception vector at the current moment, the perception saliency index at the current moment, the dynamic threshold at the current moment, the business scenario category at the current moment, and the weight coefficients of each dimension under the current scenario, will be written into the log storage for subsequent offline analysis and parameter optimization.

[0140] In this embodiment of the invention, the early warning judgment method based on the comparison of the perceived saliency index and the dynamic threshold, combined with the degradation direction of the dimension that contributes the most to the perceived saliency index, ensures that the early warning only targets negative changes in network quality, avoiding false alarms due to network quality improvements; by recording the name of the degradation dimension, the fault type is accurately located, shortening the operation and maintenance response time; by dividing the early warning levels beyond the threshold, the hierarchical scheduling of operation and maintenance resources is supported; in addition, the log storage of data that does not trigger an early warning establishes a data closed loop, providing data support for subsequent offline analysis and optimization of dynamic parameters, and realizing the system's continuous adaptability to complex network environments.

[0141] Please refer to Figure 2 This invention also provides a home broadband network quality diagnosis and early warning device 210, comprising:

[0142] The data acquisition module 211 is used to collect home broadband network quality indicators and obtain network quality information at the current time based on the home broadband network quality indicators. The network quality information includes network quality indicators of at least one dimension.

[0143] The sensing baseline calculation module 212 is used to calculate the sensing baseline at the current moment based on the network quality information at the current moment and the sensing baseline at the previous moment.

[0144] The saliency index calculation module 213 is used to obtain the perception saliency index at the current time based on the network quality information at the current time and the perception baseline at the current time.

[0145] Threshold calculation module 214 is used to determine the dynamic threshold at the current moment based on the current sensing baseline;

[0146] The early warning decision module 215 is used to execute early warning decisions and output the results.

[0147] In this embodiment of the invention, network quality information is obtained based on collected network quality indicators, the network quality information including network quality indicator values ​​of at least one dimension; a perception baseline is determined based on the network quality information at the current moment and the perception baseline at the previous moment; the perception baseline reflects the level of network quality information expected by the user; the perception baseline includes a perception baseline value of at least one dimension, the dimension of the perception baseline being the same as the dimension of the network quality information; a perception saliency index is obtained based on the network quality information at the current moment and the perception baseline at the current moment; the perception saliency index reflects the degree to which the user can perceive changes in the network quality information; a dynamic threshold is determined based on the perception baseline at the current moment; and a warning is determined based on the perception saliency index and the dynamic threshold at the current moment. In this way, by dynamically determining the perception baseline at the current moment, the system reflects the user's psychological expectations of network quality indicators in real time, ensuring the consistency between the warning triggering conditions and the user's actual psychological expectations; by dynamically determining the perception salience index at the current moment, the system reflects the degree of deviation of the network quality indicators from the user's psychological expectations of network quality indicators at the current moment and in the current context; and by determining a dynamic threshold, which is dynamically adjusted according to the user's perception state, the system improves the accuracy of home broadband network warnings in different network environments.

[0148] Optionally, determining the current sensing baseline based on the current network quality information and the previous sensing baseline includes:

[0149] Based on the network quality information at the current moment and the perception baseline at the previous moment, determine whether to update the perception baseline at the previous moment;

[0150] When determining to update the perception baseline of the previous moment, the intermediate perception baseline of the current moment is calculated based on the network quality information of the current moment and the perception baseline of the previous moment.

[0151] Based on the context category of the user-driven business at the current moment, the intermediate perception baseline at the current moment is adjusted to obtain the perception baseline at the current moment.

[0152] Optionally, determining whether to update the perception baseline at the previous moment based on the network quality information at the current moment and the perception baseline at the previous moment includes:

[0153] Calculate the absolute value of the deviation between the network quality information at the current moment and the perception baseline at the previous moment in each dimension, and obtain the maximum value of the absolute value of the deviation in each dimension;

[0154] The maximum value is compared with a preset noise threshold. If the maximum value is less than the noise threshold, the perception baseline of the previous moment is determined to be the perception baseline of the current moment.

[0155] If the maximum value is greater than or equal to the noise threshold, it is determined that the perception baseline of the previous moment should be updated.

[0156] Optionally, when determining to update the perception baseline of the previous moment, calculating the intermediate perception baseline of the current moment based on the network quality information of the current moment and the perception baseline of the previous moment includes:

[0157] The intermediate perception baseline at the current moment is obtained based on the network quality information at the current moment, the perception baseline at the previous moment, and the personalized adaptation rate parameter.

[0158] The personalized adaptation rate parameter is determined offline based on the user's historical complaint records and the network quality information for the corresponding time period, and is used to control the rate at which the perception baseline at the current moment converges to the network quality information at the current moment.

[0159] Optionally, adjusting the intermediate perception baseline at the current moment based on the context category of the user-dominated service at the current moment to obtain the perception baseline at the current moment includes:

[0160] Identify the user's dominant business type and classify the dominant business into a preset scenario category to obtain the user's dominant business scenario category;

[0161] Select the corresponding context bias index according to the dominant business context category, and perform low-pass filtering on the context bias index to obtain a smoothed context bias index. The context bias index represents the adjustment of the intermediate perception baseline under the preset context category. The dimension of the context bias index is the same as the dimension of the network quality information.

[0162] The smoothed context bias index is superimposed on the intermediate perception baseline at the current moment to obtain the perception baseline at the current moment.

[0163] Optionally, based on the network quality information at the current moment and the perception baseline at the current moment, the perception saliency index at the current moment is obtained, including:

[0164] Obtain the relative deviation of each dimension between the network quality information at the current moment and the perception baseline at the current moment;

[0165] The weighting coefficients of the relative deviation of each dimension are determined based on the dominant business scenario category at the current moment;

[0166] The relative deviations of each dimension are weighted by the weighting coefficients of the relative deviations of each dimension to obtain the weighted relative deviations of each dimension between the network quality information at the current time and the perception baseline at the current time.

[0167] The maximum value of the weighted relative deviation of each dimension is taken as the perceptual saliency index at the current moment.

[0168] Optionally, based on the current sensing baseline, the dynamic threshold for the current moment is determined, including:

[0169] Obtain the average value of each dimension of the perception baseline at the current moment;

[0170] The dynamic threshold at the current moment is calculated based on the preset basic threshold constant, the proportional coefficient, and the average value. The basic threshold constant and the proportional coefficient are set and adjusted periodically based on the statistical analysis results of user historical complaint data.

[0171] Optionally, determining whether to trigger an early warning based on the perceived saliency index at the current moment and the dynamic threshold at the current moment includes:

[0172] Determine whether the perceived saliency index at the current moment is greater than or equal to the dynamic threshold at the current moment. If the perceived saliency index at the current moment is greater than or equal to the dynamic threshold at the current moment and the direction of change of the dimension index with the maximum weighted relative deviation of each dimension is the direction of deterioration, then trigger an early warning.

[0173] The rule for determining the direction of degradation is as follows: when the dimension index with the maximum weighted relative deviation of each dimension is a positive index, if the value of the dimension that contributes the most in the perception significance index at the current moment is less than the value of the dimension corresponding to the perception baseline at the current moment, then it is considered degradation.

[0174] When the dimension index with the maximum weighted relative deviation of each dimension is a negative index, and the value of the dimension that contributes the most in the perception significance index at the current moment is greater than the value of the dimension corresponding to the perception baseline at the current moment, then it is considered deterioration.

[0175] Please refer to Figure 3 The present invention also provides an electronic device 310, including a processor 311, a memory 312, and a computer program stored in the memory 312 and executable on the processor 311. When the computer program is executed by the processor 301, it implements the various processes of the above-described home broadband network quality diagnosis and early warning method embodiments and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0176] This invention also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the various processes of the above-described home broadband network quality diagnosis and early warning method embodiments, achieving the same technical effects. To avoid repetition, it will not be described again here. The computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0177] This application also provides a computer program product, including computer instructions, which, when executed by a processor, implement the above-described... Figure 1 The various processes of the method embodiments shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0178] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0179] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0180] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of the present invention.

Claims

1. A method for diagnosing and warning the quality of home broadband networks, characterized in that, include: Collect network quality indicators, and obtain network quality information based on the network quality indicators. The network quality information includes network quality indicator values ​​of at least one dimension. The current perception baseline is determined based on the network quality information at the current moment and the perception baseline at the previous moment; the perception baseline reflects the level of the network quality information expected by the user; the perception baseline includes a perception baseline value of at least one dimension, and the dimension of the perception baseline is the same as the dimension of the network quality information. Based on the network quality information at the current moment and the perception baseline at the current moment, the perception salience index at the current moment is obtained; the perception salience index reflects the degree to which the user can perceive changes in the network quality information. Based on the current perception baseline, determine the dynamic threshold at the current moment; Based on the perceived saliency index at the current moment and the dynamic threshold at the current moment, determine whether to trigger an early warning.

2. The method according to claim 1, characterized in that, The step of determining the current sensing baseline based on the current network quality information and the previous sensing baseline includes: Based on the network quality information at the current moment and the perception baseline at the previous moment, determine whether to update the perception baseline at the previous moment; When determining to update the perception baseline of the previous moment, the intermediate perception baseline of the current moment is calculated based on the network quality information of the current moment and the perception baseline of the previous moment. Based on the context category of the user-driven business at the current moment, the intermediate perception baseline at the current moment is adjusted to obtain the perception baseline at the current moment.

3. The method according to claim 2, characterized in that, Determining whether to update the perception baseline at the previous moment based on the current network quality information and the previous moment's perception baseline includes: Calculate the absolute value of the deviation between the network quality information at the current moment and the perception baseline at the previous moment in each dimension, and obtain the maximum value of the absolute value of the deviation in each dimension; The maximum value is compared with a preset noise threshold. If the maximum value is less than the noise threshold, the perception baseline of the previous moment is determined to be the perception baseline of the current moment. If the maximum value is greater than or equal to the noise threshold, it is determined that the perception baseline of the previous moment should be updated.

4. The method according to claim 2, characterized in that, When determining to update the perception baseline of the previous time step, calculating the intermediate perception baseline of the current time based on the network quality information of the current time and the perception baseline of the previous time step includes: The intermediate perception baseline at the current moment is obtained based on the network quality information at the current moment, the perception baseline at the previous moment, and the personalized adaptation rate parameter. The personalized adaptation rate parameter is determined offline based on the user's historical complaint records and the network quality information for the corresponding time period, and is used to control the rate at which the perception baseline at the current moment converges to the network quality information at the current moment.

5. The method according to claim 2, characterized in that, The step of adjusting the intermediate perception baseline at the current moment based on the context category of the user-dominated service at the current moment to obtain the perception baseline at the current moment includes: Identify the user's dominant business type and classify the dominant business into a preset scenario category to obtain the user's dominant business scenario category; Select the corresponding context bias index according to the dominant business context category, and perform low-pass filtering on the context bias index to obtain a smoothed context bias index. The context bias index represents the adjustment of the intermediate perception baseline under the preset context category. The dimension of the context bias index is the same as the dimension of the network quality information. The smoothed context bias index is superimposed on the intermediate perception baseline at the current moment to obtain the perception baseline at the current moment.

6. The method according to claim 5, characterized in that, Based on the network quality information and the perception baseline at the current moment, obtain the perception saliency index at the current moment, including: Obtain the relative deviation of each dimension between the network quality information at the current moment and the perception baseline at the current moment; The weighting coefficients of the relative deviation of each dimension are determined based on the dominant business scenario category at the current moment; The relative deviations of each dimension are weighted by the weighting coefficients of the relative deviations of each dimension to obtain the weighted relative deviations of each dimension between the network quality information at the current time and the perception baseline at the current time. The maximum value of the weighted relative deviation of each dimension is taken as the perceptual saliency index at the current moment.

7. The method according to any one of claims 1-6, characterized in that, Based on the current perception baseline, determine the dynamic threshold for the current moment, including: Obtain the average value of each dimension of the perception baseline at the current moment; The dynamic threshold at the current moment is calculated based on the preset basic threshold constant, the proportional coefficient, and the average value. The basic threshold constant and the proportional coefficient are set and adjusted periodically based on the statistical analysis results of user historical complaint data.

8. The method according to claim 6, characterized in that, The step of determining whether to trigger an early warning based on the perceived saliency index at the current moment and the dynamic threshold at the current moment includes: Determine whether the perceived saliency index at the current moment is greater than or equal to the dynamic threshold at the current moment. If the perceived saliency index at the current moment is greater than or equal to the dynamic threshold at the current moment and the direction of change of the dimension index with the maximum weighted relative deviation of each dimension is the direction of deterioration, then trigger an early warning. The rule for determining the direction of degradation is as follows: when the dimension index with the maximum weighted relative deviation of each dimension is a positive index, if the value of the dimension that contributes the most in the perception significance index at the current moment is less than the value of the dimension corresponding to the perception baseline at the current moment, then it is considered degradation. When the dimension index with the maximum weighted relative deviation of each dimension is a negative index, and the value of the dimension that contributes the most in the perception significance index at the current moment is greater than the value of the dimension corresponding to the perception baseline at the current moment, then it is considered deterioration.

9. A home broadband network quality diagnosis and early warning device, characterized in that, include: The data acquisition module is used to collect home broadband network quality indicators and obtain network quality information at the current moment based on the home broadband network quality indicators. The network quality information includes network quality indicators of at least one dimension. The sensing baseline calculation module is used to calculate the sensing baseline at the current moment based on the network quality information at the current moment and the sensing baseline at the previous moment. The perception saliency index calculation module is used to obtain the perception saliency index at the current moment based on the network quality information at the current moment and the perception baseline at the current moment. The dynamic threshold calculation module is used to determine the dynamic threshold at the current moment based on the current sensing baseline. The early warning decision module is used to execute early warning decisions and output the results.

10. An electronic device, characterized in that, include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the home broadband network quality diagnosis and early warning method as described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the home broadband network quality diagnosis and early warning method as described in any one of claims 1 to 8.

12. A computer program product, characterized in that, It includes computer instructions, which, when executed by a processor, implement the steps of the home broadband network quality diagnosis and early warning method as described in any one of claims 1 to 8.

Images