A communication performance testing method and system for base station network side transmission path

Abstract

The present application provides a communication performance test method and system for the transmission path on the network side of a base station. Among them, the target monitoring node is determined from multiple monitoring nodes deployed on the transmission path on the network side of the base station to be tested, and the transmission delay data, packet loss rate data and bandwidth utilization data of the test data stream on the specific section of the transmission path are monitored in real time; the periodic and non-periodic changes of the transmission delay are detected to obtain analysis results; the impact of the packet loss rate data on the transmission performance is evaluated to obtain evaluation results; the change trend of the bandwidth utilization data is predicted to obtain prediction results; the three results are combined to calculate the communication performance index of the specific section; according to the comparison result of whether the communication performance of the specific section meets the predetermined standard, a communication performance test report is generated. The present application comprehensively evaluates the communication performance of the transmission path on the network side of the base station by integrating chaos theory, fuzzy logic system and deep learning algorithm, thereby improving the efficiency of communication performance testing.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of performance testing technology, and in particular, to a method and system for testing the communication performance of a transmission path on a base station network side. Background Art

[0002] In modern communication networks, especially in 5G and later communication systems, the performance of the transmission path directly affects the user experience and service quality. With the emergence of high-bandwidth applications such as high-definition video, virtual reality, and telemedicine, higher requirements are placed on real-time monitoring and performance evaluation of the transmission path.

[0003] At present, the communication performance test of the transmission path on the base station network side usually relies on traditional statistical analysis methods, such as simple time series analysis, basic statistics such as mean and variance.

[0004] Traditional statistical methods can often only provide single-dimensional performance evaluation, and it is difficult to comprehensively reflect the combined impact of multiple factors on communication performance, resulting in poor efficiency in testing the communication performance of the transmission path on the base station network side. Summary of the invention

[0005] The embodiments of the present application provide a method and system for testing the communication performance of a transmission path on a base station network side, so as to solve the problem of poor efficiency in testing the communication performance of a transmission path on a base station network side in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a method for testing communication performance of a transmission path on a base station network side, including:

[0007] Determine a target monitoring node from a plurality of monitoring nodes deployed on the transmission path of the network side of the base station to be tested, and monitor the transmission delay data, packet loss rate data and bandwidth utilization data of the received test data stream on a specific section of the transmission path in real time through the target monitoring node;

[0008] Analyzing the transmission delay data to detect periodic and non-periodic changes of the transmission delay and obtaining analysis results;

[0009] Processing the packet loss rate data to evaluate the impact of the packet loss rate data on transmission performance and obtain an evaluation result;

[0010] Analyze the bandwidth utilization data, predict the change trend of the bandwidth utilization data, and obtain a prediction result;

[0011] Combining the analysis result, the evaluation result and the prediction result, calculating the communication performance index of the specific road section;

[0012] The communication performance index is compared with a preset communication performance standard to evaluate whether the communication performance of the specific road section meets the predetermined standard, and based on the comparison result, a communication performance test report of the transmission path is generated and optimization suggestions are provided.

[0013] In a second aspect, an embodiment of the present application provides a communication performance test system for a base station network side transmission path, including:

[0014] A monitoring module, used to determine a target monitoring node from a plurality of monitoring nodes deployed on the transmission path of the network side of the base station to be tested, and to monitor the transmission delay data, packet loss rate data and bandwidth utilization data of the received test data stream on a specific section of the transmission path in real time through the target monitoring node;

[0015] An analysis module, used for analyzing the transmission delay data to detect periodic and non-periodic changes of the transmission delay and obtain an analysis result;

[0016] An evaluation module, used for processing the packet loss rate data to evaluate the impact of the packet loss rate data on transmission performance and obtain an evaluation result;

[0017] A prediction module, used to analyze the bandwidth utilization data, predict the change trend of the bandwidth utilization data, and obtain a prediction result;

[0018] A calculation module, used to combine the analysis result, the evaluation result and the prediction result to calculate the communication performance index of the specific road section;

[0019] The processing module is used to compare the communication performance index with the preset communication performance standard to evaluate whether the communication performance of the specific road section meets the preset standard, and generate a communication performance test report of the transmission path based on the comparison result, and provide optimization suggestions.

[0020] In a third aspect, an embodiment of the present application provides a computing device, characterized in that it includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a communication performance testing method for a base station network side transmission path as described in the first aspect above.

[0021] In a fourth aspect, an embodiment of the present application provides a computer storage medium, characterized in that a computer program is stored therein, and when the computer program is executed by a computer, a communication performance testing method of a base station network side transmission path as described in the first aspect is implemented.

[0022] In an embodiment of the present application, a target monitoring node is determined from a plurality of monitoring nodes deployed on the transmission path on the network side of the base station to be tested, and the transmission delay data, packet loss rate data and bandwidth utilization data of the test data stream on a specific section of the transmission path are monitored in real time; the periodic and non-periodic changes of the transmission delay are detected to obtain analysis results; the impact of the packet loss rate data on the transmission performance is evaluated to obtain evaluation results; the changing trend of the bandwidth utilization data is predicted to obtain prediction results; the three results are combined to calculate the communication performance index of the specific section; and a communication performance test report is generated based on the comparison results of whether the communication performance of the specific section meets the predetermined standard. The present application comprehensively evaluates the communication performance of the transmission path on the network side of the base station by integrating chaos theory, fuzzy logic system and deep learning algorithm, thereby improving the efficiency of communication performance testing.

[0023] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0025] Figure 1 A flow chart of a method for testing the communication performance of a transmission path on a base station network side provided in an embodiment of the present application;

[0026] Figure 2 A schematic diagram of the structure of a communication performance test system for a base station network side transmission path provided in an embodiment of the present application;

[0027] Figure 3 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0029] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0030] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0031] Figure 1 A flow chart of a method for testing the communication performance of a transmission path on a base station network side is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0032] 101. Determine a target monitoring node from a plurality of monitoring nodes deployed on a transmission path on the network side of the base station to be tested, and monitor the transmission delay data, packet loss rate data and bandwidth utilization data of the received test data stream on a specific section of the transmission path in real time through the target monitoring node;

[0033] In this step, one or more monitoring points (such as routers, switches, etc.) are selected as target monitoring nodes in the base station network to be tested. These nodes will be used to collect data on transmission delay, packet loss rate and bandwidth utilization on a specific section.

[0034] In the embodiment of the present application, a dedicated network probe is deployed or an existing network device (supporting SNMP protocol) is used to obtain data. For example, a network performance monitor is set on the optical fiber link between two major cities, which can send test traffic and transmit information back through the receiving end to measure the actual delay and packet loss.

[0035] 102. Analyze the transmission delay data to detect periodic and non-periodic changes of the transmission delay and obtain analysis results;

[0036] In this step, by analyzing the transmission delay data collected from the target monitoring nodes, it is identified whether there are periodic changes (such as morning and evening peak hours) or non-periodic changes (abnormalities caused by emergencies).

[0037] In the embodiment of the present application, a time series analysis method, such as an ARIMA model, is used to analyze the average delay trend in the same time period every day, and statistical means are used to detect outliers.

[0038] Optionally, the step 102 of "analyzing the transmission delay data to detect periodic and non-periodic changes of the transmission delay to obtain analysis results" includes: constructing a chaotic time series, converting the transmission delay data into a chaotic time series for analyzing the dynamic characteristics corresponding to the transmission delay data; calculating the embedding dimension and delay time of the chaotic time series to reconstruct the phase space, and applying the phase space reconstruction technology to reconstruct the chaotic time series into a phase space trajectory according to the embedding dimension and the delay time; and analyzing the stability of the phase space trajectory using the Lyapunov exponent in chaos theory. Detect non-periodic changes in transmission delay; analyze the spectrum of the chaotic time series by Fourier transform to detect periodic changes in transmission delay; use wavelet transform to analyze local characteristics of the chaotic time series, detect changes in transmission delay on different time scales, and obtain wavelet transform analysis results; based on the wavelet transform analysis results, correct the Lyapunov exponent analysis results, Fourier transform spectrum analysis results, and wavelet transform analysis results to obtain adjusted periodic and non-periodic changes in the transmission delay, and generate analysis results based on the adjusted periodic and non-periodic changes in the transmission delay.

[0039] Chaotic time series: refers to a time series that appears random but actually follows certain deterministic rules.

[0040] Phase space reconstruction: By selecting appropriate embedding dimensions and delay times, one-dimensional time series are converted into trajectories in high-dimensional space to better reflect the internal structure of the system.

[0041] Lyapunov index: One of the indicators to measure the stability of a dynamic system. A positive value indicates that the system has divergent characteristics (i.e. chaos), while a negative value indicates that the system tends to be stable.

[0042] Fourier transform: A mathematical tool that converts a signal from the time domain to the frequency domain and is used to analyze the frequency content of a signal.

[0043] Wavelet transform: A signal processing technique that has both time and frequency resolution, particularly suitable for analyzing non-stationary signals.

[0044] First, the collected transmission delay data is regarded as a time series, and it is assumed that there is a chaotic mechanism behind it. The original data is preprocessed (such as removing noise) to ensure the effectiveness of subsequent analysis. Secondly, the mutual information method and other methods are used to determine the optimal delay time τ; the pseudo nearest neighbor algorithm is used to estimate the minimum embedding dimension m. This step is to find the reconstruction parameters that best represent the dynamic characteristics of the original system. Further, based on the m and τ obtained in step 2, the one-dimensional time series is mapped to the m-dimensional space using Takens' theorem to form a series of new vectors. Each vector represents a snapshot of the system state. The maximum Lyapunov exponent λmax is then estimated by tracking the growth of the distance between adjacent trajectories in the reconstructed phase space over time. If λmax>0, it means that the time series contains chaotic components, and the larger λmax means stronger instability. Then the fast Fourier transform (FFT) is applied to convert the time series into a spectrum, from which the main periodic fluctuation components and their intensities are identified. Secondly, the continuous wavelet transform (CWT) or discrete wavelet transform (DWT) technology is used to decompose the original time series at different scales to reveal the multi-level structural features hidden in it. Finally, combining the results of the previous steps, especially using the multi-resolution perspective provided by the wavelet transform, the Lyapunov exponent and Fourier spectrum are corrected to obtain a more accurate description of periodic / non-periodic changes.

[0045] In the embodiment of the present application, it is assumed that there is a set of transmission delay data {d_1, d_2, ..., d_n} continuously recorded on a certain link. After preliminary cleaning, it is found that these data may have potential chaotic behavior. So follow the above process:

[0046] First, we select an appropriate τ=3 sampling interval as the delay time; determine the embedding dimension m=5; then use this set of parameters to reconstruct the original sequence into a phase space trajectory; the maximum Lyapunov exponent is calculated to be about λmax≈0.25, suggesting a certain degree of chaos; further FFT results show that there is an obvious daily periodicity (about 24 hours); CWT shows that in addition to daily fluctuations, there are also rapid changes in shorter periods. The final report pointed out that although there is a stable daily pattern overall, there will still be sudden and significant delay increases in certain periods, and it is recommended to take measures to enhance network robustness and optimize resource scheduling strategies.

[0047] The method provided by the present invention can not only effectively distinguish the periodic and non-periodic components in the transmission delay, but also accurately locate those subtle changes that are difficult to detect by traditional means. This helps operation and maintenance personnel to more comprehensively understand the network operation status and take corresponding countermeasures in time, thereby significantly improving service quality and user experience.

[0048] The present application takes into account that in the prior art, the analysis of network transmission delay usually adopts traditional statistical methods or simple trend analysis. These methods are often difficult to capture the complex dynamic characteristics in the data, especially when there are nonlinear, periodic and non-periodic mixed changes. Traditional linear models may not fully reflect all the characteristics under the actual network conditions, which leads to insufficient understanding of transmission delay, thus affecting the effect of communication performance optimization. For example, describing the delay only by the mean and standard deviation may cover up some sudden abnormal behaviors, which are crucial for network operation and maintenance. In order to solve the above problems, the present application proposes a time series analysis method based on chaos theory. This method can not only identify periodic changes, but also effectively detect changes in non-periodic and local characteristics, thereby providing more accurate and reliable transmission delay analysis results. By combining Lyapunov exponents, Fourier transforms and wavelet transforms, the dynamic characteristics of network delay can be more comprehensively understood, and more accurate performance evaluation and optimization suggestions can be made accordingly.

[0049] The options are as follows:

[0050] Optionally, the step of “analyzing the transmission delay data to detect periodic and non-periodic changes of the transmission delay to obtain analysis results” in step 102 includes:

[0051] Construct a chaotic time series:

[0052] ;

[0053] in, represents the time series of transmission delay data, Indicates at time The transmission delay value of is the total number of data points in the time series;

[0054] Calculate embedding dimension and delay time And reconstruct the phase space:

[0055] ;

[0056] in, is a sequence of increasing dimensions calculated from the maximum Lyapunov exponent, is the dimension;

[0057] ;

[0058] in, Respectively at the time and The transmission delay value of is the length of the time series;

[0059] Phase space reconstruction:

[0060] ;

[0061] in, represents the reconstructed phase space trajectory, represents the first dimensional coordinates, Indicates The dimensional coordinates are determined by the original time series The elements in the Calculated.

[0062] Lyapunov exponent analysis:

[0063] ;

[0064] in, is the Lyapunov exponent, is the neighboring trajectory in phase space, is the norm of the vector;

[0065] Fourier transform spectrum analysis:

[0066] ;

[0067] in, is the result of Fourier transform, is the frequency, is the length of the time series, is a time series, is a complex exponential function;

[0068] Wavelet transform analysis:

[0069] ;

[0070] in, are the wavelet coefficients, is the scale parameter, is the displacement parameter, is the wavelet basis function, is the conjugate of the wavelet basis function;

[0071] Correction of non-periodic and periodic changes:

[0072] Corrected non-cyclical changes ;

[0073] Corrected cycle changes ;

[0074] in, and is the adjustment coefficient, which is used to correct the non-periodic and periodic changes according to the wavelet transform results;

[0075] Comprehensive analysis results:

[0076] ;

[0077] in, It indicates the analysis result that comprehensively considers non-periodic and periodic changes.

[0078] Construct chaotic time series: Construct a time series consisting of transmission delay data, where each data point represents the transmission delay value at a certain moment.

[0079] Calculate embedding dimension and delay time And reconstruct the phase space: calculate the embedding dimension and delay time , to determine the parameters required to reconstruct the phase space.

[0080] Phase space reconstruction: according to the embedding dimension and delay time Reconstruct the phase space to form a multi-dimensional vector.

[0081] Lyapunov exponent analysis: Calculates Lyapunov exponents, which are used to detect non-periodic changes (chaotic behavior) in a system.

[0082] Fourier transform spectrum analysis: Analyze the spectrum characteristics of transmission delay data through Fourier transform to detect periodic changes in the data.

[0083] Wavelet Transform Analysis: Use wavelet transform to analyze the local characteristics of transmission delay data and detect non-periodic and periodic changes in the data.

[0084] Correct non-periodic changes and periodic changes: Correct non-periodic changes and periodic changes based on the wavelet transform results to make the results more accurate and reliable.

[0085] Comprehensive analysis results: The results of non-periodic changes and periodic changes are combined to obtain the comprehensive analysis results of the transmission delay data.

[0086] In the embodiment of the present application, it is assumed that there is a set of transmission delay data {X(t)={x_1,x_2,...,x_1000}} continuously recorded on a certain link, and the sampling interval is 1 second.

[0087] Assume that the first 10 data points are {x_1=2ms,x_2=3ms,...,x_10=4ms}. After calculation, assume that the optimal embedding dimension De=5 and the delay time τ=10 seconds. The reconstructed phase space trajectory Y(t)={y_1(t),y_2(t),...,y_5(t)}, where y_1(t)=x_t,y_2(t)=x_(t-10),... Assume that the calculated maximum Lyapunov exponent λ≈0.1 indicates that there is a slight chaotic behavior. The Fourier transform shows that there is an obvious daily periodicity (about 86400 seconds). The CWT results show that in addition to daily fluctuations, there are also rapid changes in short periods. Assuming α=0.5,β=0.3, after adjustment according to the wavelet transform results, the corrected non-periodic change is about 0.15, and the corrected periodic change is about 0.09. A_"chaos"=√(0.15^2+0.09^2)≈0.167, indicating that the transmission delay of the link exhibits a certain degree of chaotic characteristics.

[0088] By using this chaos theory-based approach, the complex dynamic behavior in network transmission delays can be captured more accurately. Specifically, compared with traditional methods, this method can better identify periodic and non-periodic change patterns, which helps to discover potential problems. Secondly, through detailed analysis of non-periodic changes, sudden events such as hardware failures or network attacks can be located more quickly. Based on the detailed delay analysis results, network administrators can take more targeted measures to optimize network performance and improve service quality. And by timely adjusting network configuration and resource allocation, service interruptions caused by delays are reduced, ensuring a smooth user experience.

[0089] 103. Process the packet loss rate data to evaluate the impact of the packet loss rate data on transmission performance and obtain an evaluation result;

[0090] In this step, based on the collected packet loss rate data, its impact on the overall network transmission quality is evaluated.

[0091] In the embodiment of the present application, the average packet loss rate per hour / day is calculated and compared with the industry standard; if it is found that the packet loss rate exceeds the preset threshold for several consecutive time periods, it is considered that this may seriously affect the service quality.

[0092] Optionally, the "processing the packet loss rate data to evaluate the impact of the packet loss rate data on the transmission performance and obtain an evaluation result" in step 103 includes: defining fuzzy sets defined in different ranges according to the packet loss rate data, the fuzzy sets including low packet loss rate, medium packet loss rate and high packet loss rate; constructing a fuzzy logic system, setting the input variable as the packet loss rate data, the output variable as the transmission performance impact level, and setting the fuzzy logic rules corresponding to the fuzzy logic system according to the fuzzy set corresponding to the packet loss rate data, wherein the fuzzy logic rules include: if the packet loss rate data belongs to a low packet loss rate, the transmission performance impact level is a slight impact, if the packet loss rate data belongs to a medium packet loss rate, the transmission performance impact level is a medium impact, and if the packet loss rate data belongs to a high packet loss rate, the transmission performance impact level is a severe impact; inputting the packet loss rate data into the fuzzy logic system, performing reasoning according to the set fuzzy logic rules, and obtaining the transmission performance impact level; using defuzzification processing to convert the transmission performance impact level into a transmission performance impact level value, and determining the evaluation result of the impact of the packet loss rate on the transmission performance according to the transmission performance impact level value.

[0093] Fuzzy set: Different from the binary division of classical sets where elements belong to or do not belong to the set, in fuzzy sets, each element can belong to the set to some extent.

[0094] Fuzzy logic system: A method that simulates the human reasoning process, which allows for uncertain relationships between variables and is suitable for processing imprecise and ambiguous data.

[0095] Defuzzification: The process of converting fuzzy logic output into clear numerical values ​​so that the final result can be directly used for decision making.

[0096] First, the different levels of packet loss rate (low, medium, high) are determined based on experience or experimental data, and the corresponding membership functions are given to describe the extent to which each packet loss rate value belongs to these levels. Secondly, the input variable is set to the specific packet loss rate value, and the output variable is the transmission performance impact level. Then, based on the fuzzy set defined in the first step, a series of fuzzy logic rules are formulated. For example, if the packet loss rate is low, it is considered that the impact on transmission performance is small; vice versa. Then, the actual measured packet loss rate is input into the established fuzzy logic system, and the calculation is performed according to the preset rules to obtain the transmission performance impact level under the current packet loss rate. The last step is to convert the results obtained from the above reasoning from fuzzy form into specific numbers for further analysis and use.

[0097] In an embodiment of the present application, it is assumed that a video streaming service provider wants to monitor the quality of its network service. They collected data on the average packet loss rate every day in the past month and decided to use this fuzzy logic method for evaluation. For example, a packet loss rate of <1% is defined as low, 1%-5% is medium, and >5% is high; at the same time, it is stipulated that if the packet loss rate is low, the impact is slight (0.1), medium has a large impact (0.5), and high has a serious impact (0.9). When it is detected that the average packet loss rate on a certain day is 3.2%, the fuzzy logic system calculates that the service quality of that day has been affected to a certain extent, but has not reached a serious level. The quantified result may be around 0.6. This shows that although there is a problem, it is still within an acceptable range.

[0098] By adopting such a fuzzy logic evaluation method, it can not only more accurately reflect the actual impact of packet loss rate changes on network transmission performance, but also help managers to promptly discover potential problems and take corresponding measures to improve them, thereby ensuring that users get a better experience. In addition, this method is also very adaptable and scalable, and parameter settings can be adjusted according to specific circumstances to make it suitable for more diverse scenario requirements.

[0099] The present application takes into account that in the current network transmission performance evaluation, packet loss rate is an important indicator for measuring service quality. However, traditional evaluation methods usually regard packet loss rate as an absolute value and make judgments based on it. This approach ignores the gradual characteristics of the impact of packet loss rate on service and the differences in packet loss tolerance of different applications. For example, in video streaming transmission, slight packet loss may not significantly affect the user experience, while the same packet loss rate may cause a significant degradation in quality in real-time voice communication. Therefore, a more flexible method that can reflect actual conditions is needed to evaluate the impact of packet loss rate on transmission performance. In order to solve the above problems, the present application proposes a packet loss rate evaluation scheme based on a fuzzy logic system. By introducing fuzzy sets and fuzzy logic rules, this method can more accurately describe the relationship between packet loss rate and transmission performance, and provide more detailed and practical evaluation results.

[0100] The options are as follows:

[0101] Optionally, the step of “processing the packet loss rate data to evaluate the impact of the packet loss rate data on transmission performance to obtain an evaluation result” in step 103 includes:

[0102] Define fuzzy sets:

[0103] ;

[0104] Fuzzy logic reasoning:

[0105] ;

[0106] Defuzzification:

[0107] ;

[0108] Comprehensive evaluation results:

[0109] ;

[0110] Define fuzzy sets:

[0111] Low packet loss rate : Define the fuzzy membership function of the packet loss rate in the low range.

[0112] Moderate packet loss rate : Define the fuzzy membership function of the packet loss rate in the medium range.

[0113] High packet loss rate : Define the fuzzy membership function of the packet loss rate in the high range.

[0114] Imaginary logical reasoning:

[0115] According to the packet loss rate data The fuzzy membership of the selected object is the highest membership as the result of fuzzy logic reasoning. .

[0116] Defuzzification:

[0117] Use the Centroid Defuzzification method or other defuzzification methods to convert the result of fuzzy logic reasoning into a specific value. .

[0118] Comprehensive evaluation results:

[0119] The final fuzzy logic evaluation result Represents the degree of impact of packet loss rate data on transmission performance.

[0120] In the present embodiment, it is assumed that an online education platform wants to evaluate the network transmission performance of its live courses. To distinguish low, medium and high levels of packet loss. If the average packet loss rate during a live broadcast is :

[0121] Calculate membership:

[0122] ;

[0123] Fuzzy logic reasoning: because , at this time, you can decide which one to use or a combination of the two according to the specific application scenario; here, assuming that a higher impact level is selected, as the main influencing factor.

[0124] Defuzzification assumes that the simplified processing only considers the value corresponding to the highest membership degree, so directly take .

[0125] Comprehensive evaluation results , which means that the service performance was affected to a certain extent due to packet loss during the live broadcast, but it was not completely unacceptable.

[0126] By adopting the above-mentioned fuzzy logic-based evaluation method, we can not only better capture the real situation of the impact of packet loss rate changes on service performance, but also provide more customized solutions for different application scenarios. In addition, this method also enhances the system's ability to cope with uncertainty and volatility, which helps to improve the overall service quality and user experience.

[0127] 104. Analyze the bandwidth utilization data, predict a change trend of the bandwidth utilization data, and obtain a prediction result;

[0128] In this step, historical bandwidth usage records are used to predict future bandwidth requirements and development trends.

[0129] In an embodiment of the present application, a machine learning algorithm (such as linear regression, neural network, etc.) is applied to perform short- to medium-term predictions based on bandwidth usage patterns over the past few months or even years.

[0130] Optionally, the "analyzing the bandwidth utilization data, predicting the changing trend of the bandwidth utilization data, and obtaining a prediction result" in step 104 includes: constructing a training data set, dividing the bandwidth utilization data into a training set and a test set, and selecting and initializing a deep learning model; training the deep learning model using the training set, optimizing the model parameters through a back propagation algorithm, and verifying the trained deep learning model using the test set to evaluate the prediction performance of the deep learning model; predicting the bandwidth utilization data monitored in real time using the trained deep learning model to obtain the changing trend of the bandwidth utilization data.

[0131] Training dataset: A collection of data used to train a machine learning model.

[0132] Training set: A part of the total data set, specifically used to train the model.

[0133] Test set: another part of the data that is not involved in training, used to evaluate model performance.

[0134] Deep learning model: An artificial neural network with a multi-layer structure that can automatically extract high-level abstract features of input data.

[0135] Back propagation algorithm: The process of adjusting the weights of each layer by calculating the output error is one of the main means of training neural networks.

[0136] Predictive performance: The ability of the model to make correct predictions about future data.

[0137] First, select data within a certain time range from the historical bandwidth utilization records as samples. Randomly divide these data into training sets and test sets, with a general ratio of 70% to 80% for training and the rest for testing. Secondly, select a suitable deep learning architecture based on the characteristics of the problem, such as LSTM (Long Short-Term Memory Network) for sequence data prediction. Initialize model parameters and set appropriate hyperparameter values. Further use the training set data to iteratively train the model, during which the back propagation algorithm is used to continuously adjust the model weights to minimize the prediction error. After each round of training, use the validation set to check for overfitting and stop the training process in a timely manner. Then use the test set to evaluate the performance of the final model, including but not limited to indicators such as mean square error (MSE) and mean absolute error (MAE). Finally, apply the trained model to the newly collected bandwidth utilization data to generate prediction results for a period of time in the future.

[0138] In the embodiment of the present application, it is assumed that a large data center needs to monitor its internal network traffic and predict future bandwidth requirements. The center has collected daily bandwidth utilization data over the past year. In order to implement the above solution:

[0139] First, we randomly selected the first 9 months of data from this set of data as the training set, and the last 3 months of data were kept as the test set. Secondly, we decided to use a neural network architecture with two layers of LSTM units, each layer containing 64 nodes, and set the initial learning rate to 0.001. After about 200 rounds of training, we observed that the training loss gradually decreased and the performance on the validation set was stable and good. Further test results showed that the MAE reached 0.05, indicating that the model has good generalization ability and high prediction accuracy. Recently, this model was deployed to the production environment, and the latest bandwidth utilization information is received regularly every day, and the traffic forecast for the next day is made based on this, helping the operation team to adjust the resource allocation strategy in a timely manner.

[0140] By introducing a deep learning-based method to predict bandwidth utilization, not only the accuracy of the prediction results is improved, but also the system's ability to cope with dynamic changes is enhanced. This method can better identify potential trends and patterns, thereby providing decision makers with more reliable information support and helping to achieve more effective resource management and optimization. In addition, the deep learning model has good scalability and can easily process larger data sets to meet the growing data processing needs.

[0141] This application takes into account that in modern network management and operation and maintenance, accurate prediction of bandwidth utilization is crucial to ensuring service quality, optimizing resource allocation, and planning future network expansion. However, traditional prediction methods such as time series analysis (such as ARIMA models) and simple linear regression models often have difficulty capturing complex nonlinear relationships in the data, especially when faced with large amounts of data and highly dynamically changing network environments. These methods may not provide sufficiently accurate prediction results, resulting in resource waste or service interruption. Therefore, this application proposes a bandwidth utilization prediction scheme based on deep learning, which uses the powerful feature learning ability of deep neural networks to improve prediction accuracy, and introduces adjustment coefficients to further refine estimates of future trends, in order to solve the problems of insufficient prediction accuracy and technical limitations in the prior art.

[0142] The options are as follows:

[0143] Optionally, the step of “analyzing the bandwidth utilization data, predicting a change trend of the bandwidth utilization data, and obtaining a prediction result” in step 104 includes:

[0144] Construct a training dataset:

[0145] ;

[0146] in, is the training dataset, is the first data points, is the input feature vector, is the actual bandwidth utilization value, is the number of data points in the training dataset;

[0147] Training a deep learning model:

[0148] ;

[0149] in, is the loss function, are model parameters, is the number of samples in the training dataset, The model is The predicted value of samples, It is The actual value (or target value) of samples, is the squared difference between the predicted value and the actual value, It is the average of the squared errors of all m samples.

[0150] Predict the changing trend of bandwidth utilization data:

[0151] ;

[0152] in, Indicates at time The predicted value of the bandwidth utilization data by the deep learning model is represents the predicted value of the deep learning model at time t+1, represents the parameters of the model, is the rate of change of the predicted value, is the adjustment factor.

[0153] First, we collect historical bandwidth utilization data, including input features (such as timestamp, traffic type, etc.) and the corresponding actual bandwidth utilization values. We organize these data into a structured form, that is, each record contains an input feature vector and a corresponding output value . Secondly, the loss function is defined as mean square error (MSE), the purpose is to minimize the difference between the predicted value and the actual value. Use gradient descent or other optimization algorithms to iteratively update the model parameters , until the loss function converges to a smaller value. The trained model is further used to predict the bandwidth utilization at a certain time t+1 in the future. Combined with the current prediction value and its rate of change , by adjusting the coefficient To fine-tune the forecast results, so as to obtain a trend forecast that is more in line with the actual situation.

[0154] In the embodiment of the present application, it is assumed that a telecommunications operator wants to make a long-term prediction of the bandwidth utilization of its backbone network in order to formulate a reasonable expansion plan. The company has collected bandwidth utilization data for each day in the past year and decided to use the above solution for prediction.

[0155] First, extract the timestamp, day type (weekday / weekend), holiday information, etc. of each day from the collected data as input features , and the actual bandwidth utilization of the day is output as . The entire dataset is divided into a training set (first 9 months) and a test set (last 3 months). Secondly, LSTM (Long Short-Term Memory Network) was selected as the deep learning architecture because it is good at processing sequence data and can remember information for a long time. The number of hidden layer units is set to 50, the batch size is 64, and the learning rate is 0.001. The LSTM model is then trained using the training set data, and the weights are continuously adjusted through the back propagation algorithm to minimize the mean square error. After about 100 epochs, the loss on the validation set stabilized at a low level. Further predictions were made on the test set and compared with the actual values. The calculated mean square error was 0.02, indicating that the model has a high prediction accuracy. Finally, the trained model was applied to the data of the latest day to obtain preliminary predictions for the next day. At the same time, the rate of change is calculated based on the data trends in recent weeks. . Set the adjustment factor , then the final prediction value is .

[0156] By adopting a bandwidth utilization prediction solution based on deep learning, operators can obtain more accurate and reliable predictions of future trends. This not only helps to identify potential bottlenecks in a timely manner, but also supports more effective resource planning and management decisions, reducing cost waste and service quality issues caused by over-configuration or insufficient resources. In addition, this method also has good generalization and adaptability, and can maintain high prediction performance in different scenarios.

[0157] 105. Calculate the communication performance index of the specific road section by combining the analysis result, the evaluation result and the prediction result;

[0158] In this step, the results of the above steps are combined with comprehensive consideration of factors such as delay, packet loss rate and bandwidth utilization to obtain an overall communication performance score for a specific road section.

[0159] In the embodiment of the present application, a weight system is defined to assign corresponding scores to different performance parameters, and then the weighted sum is taken according to the actual situation to obtain the final performance index.

[0160] Optionally, the "combining the analysis results, the evaluation results and the prediction results to calculate the communication performance index of the specific road section" in step 105 includes: constructing a comprehensive evaluation model based on the analysis results, the evaluation results and the prediction results; determining the weight coefficients of the analysis results, the evaluation results and the prediction results based on the importance of each indicator; and performing weighted fusion of the analysis results, the evaluation results and the prediction results to obtain the communication performance index of the specific road section.

[0161] Analysis results: refers to the specific observations and analysis conclusions based on the data collected in the current and past period of time.

[0162] Evaluation results: It is an overall evaluation of the communication status at a certain point in time or stage, which is based on but not limited to the analysis results.

[0163] Prediction results: Use statistical models or other algorithms to estimate the communication situation in a certain period of time in the future.

[0164] Comprehensive evaluation model: A mathematical framework that integrates information from different sources and outputs a comprehensive score based on preset rules.

[0165] Weight coefficient: A numerical parameter that represents the relative importance of each input variable.

[0166] Weighted fusion: The process of combining data from multiple independent sources in proportion to their respective importance.

[0167] First, define which factors are key indicators for measuring communication performance (such as signal stability, network latency, etc.), and select appropriate mathematical tools (such as linear regression, neural network, etc.) to express the relationship between these factors. Secondly, use expert scoring method, entropy value method, etc. to quantify the importance of each indicator, so as to obtain the corresponding weight value. Finally, multiply the raw data obtained in step 1 by its corresponding weight and sum them up to obtain the final communication performance indicator.

[0168] In the embodiment of the present application, it is assumed that the 4G network coverage quality on a main road in a certain city is being evaluated. After preliminary research, it is decided to consider three main factors: average download speed X, packet loss rate Y, and user satisfaction Z. Through investigation, it is found that these three factors account for 40%, 30%, and 30% of the total score respectively. If X=50Mbps, Y=0.02%, and Z=8 / 10 are measured in a certain test, the communication performance index CPI of the road section can be calculated by the following formula:

[0169] ;

[0170] This method can not only comprehensively consider various factors that affect communication performance, but also flexibly adjust the weight of each factor according to the specific situation, making the final result more targeted and accurate. In addition, by introducing a prediction mechanism, it can better cope with possible future changes and improve the foresight and practicality of the entire evaluation system.

[0171] This application takes into account that in existing communication performance evaluation methods, traditional evaluation methods often use linear models to process data, which may not accurately capture nonlinear relationships in complex environments. In addition, in most cases, the weights of different indicators are pre-set and fixed. This static method is difficult to adapt to real-time changes. And relying on a single type of prediction model (such as statistics or machine learning) may lead to large prediction deviations. In order to solve these problems, this application proposes a comprehensive communication performance evaluation scheme based on nonlinear transformation, dynamic weight adjustment, and adaptive fusion mechanism. The scheme aims to improve the accuracy and robustness of the evaluation results and better serve decision support in application scenarios such as intelligent transportation systems.

[0172] The options are as follows:

[0173] Optionally, in step 105, “combining the analysis result, the evaluation result and the prediction result to calculate the communication performance index of the specific road section” includes:

[0174] Nonlinear transformation function:

[0175] ;

[0176] Dynamic weight adjustment mechanism:

[0177] ;

[0178] Adaptive adjustment mechanism:

[0179] ;

[0180] in, and are the standard deviations of the predictions from the deep learning model and the ARIMA model, is the adjustment factor;

[0181] Fusion prediction results:

[0182] ;

[0183] New adaptive adjustment coefficients:

[0184] ;

[0185] in, and are the standard deviations of the chaos analysis results and the fuzzy logic evaluation results, is the standard deviation of the fusion prediction results, is the adjustment factor;

[0186] Calculation of comprehensive communication performance indicators:

[0187] ;

[0188] Final comprehensive communication performance indicators:

[0189] ;

[0190] Non-linear transformation function:

[0191] Chaos analysis results: using Sigmoid function A nonlinear transformation is performed to reflect the importance of periodic and non-periodic variations in transmission delay.

[0192] Fuzzy logic evaluation results: using square root function A nonlinear transformation is performed to reflect the impact of packet loss rate on transmission performance.

[0193] Dynamic weight adjustment mechanism:

[0194] Through the exponential function Dynamically adjust the weights of different indicators so that the importance of different indicators can be flexibly adjusted at different time points.

[0195] Adaptive adjustment mechanism:

[0196] use Fusion of the forecast results of the ARIMA model and the deep learning model.

[0197] Fusion prediction results:

[0198] use To integrate the prediction results of the ARIMA model and the deep learning model.

[0199] New adaptive adjustment coefficient:

[0200] Introducing a new adaptive adjustment coefficient , according to the chaos analysis results, model evaluation results and the difference in standard deviation of the fusion prediction results, so as to dynamically adjust the fusion coefficient according to the performance of the model.

[0201] Calculation of comprehensive communication performance indicators:

[0202] In the calculation of comprehensive communication performance indicators, use To balance the weight distribution between chaos analysis and fuzzy logic evaluation results and fusion prediction results.

[0203] when When it is close to 1, it relies more on chaos analysis and fuzzy logic evaluation results; when When it is close to 0, it relies more on the fusion prediction results.

[0204] Final comprehensive communication performance indicators:

[0205] Calculate the final comprehensive communication performance index , to reflect the communication performance status of a specific road section.

[0206] In the embodiment of the present application, it is assumed that the 5G network performance on a busy street in a city is being studied. At a certain moment, .

[0207] First, calculate the value after nonlinear transformation according to the given parameters:

[0208] ;

[0209] Next, calculate the dynamic weight:

[0210] ;

[0211] Then, determine the adaptive adjustment coefficient :

[0212] ;

[0213] And calculate the fusion prediction result based on this:

[0214] ;

[0215] Then, the new adaptive adjustment coefficient is obtained :

[0216] ;

[0217] Finally, the communication performance indicators are calculated based on the above information:

[0218] ;

[0219] Complete normalization:

[0220] ;

[0221] By introducing nonlinear transformation, dynamic weight adjustment and multi-model fusion strategy, this solution can not only capture the actual status of communication performance more accurately, but also has good adaptability and scalability. Especially when facing complex and changing application scenarios, this method can effectively reduce the errors caused by the inherent defects of traditional methods, thereby providing more reliable support for relevant decisions.

[0222] 106. Compare the communication performance index with a preset communication performance standard to evaluate whether the communication performance of the specific road section meets the preset standard, and generate a communication performance test report of the transmission path based on the comparison result, and provide optimization suggestions.

[0223] In this step, the calculated communication performance indicators are compared with the pre-set benchmark levels to determine whether they meet the requirements. A detailed test report is compiled based on this, and improvement suggestions are made.

[0224] In the embodiment of the present application, if the actual performance of a link is lower than expected, the report will point out the problem and suggest measures to improve the situation, such as adding redundant paths, upgrading hardware facilities, etc. In addition, the report should also include charts showing the historical trend of key performance indicators to help managers better understand the current situation and make decisions.

[0225] Optionally, after predicting the real-time monitored bandwidth utilization data using the trained deep learning model to obtain the changing trend of the bandwidth utilization data, it also includes: using the ARIMA model to perform time series analysis on the bandwidth utilization data to predict the changing trend of the bandwidth utilization data in the future; fusing the changing trend of the bandwidth utilization data predicted by the ARIMA model with the changing trend of the bandwidth utilization data predicted by the deep learning model to obtain the changing trend of the adjusted bandwidth utilization data.

[0226] ARIMA (AutoRegressiveIntegratedMovingAverage): is a classic statistical model for time series forecasting, which makes predictions based on past data points and the differences between these data points.

[0227] Deep learning model: refers to a type of machine learning algorithm based on neural networks, which is characterized by a multi-layer structure and can automatically learn high-level abstract representations from raw data.

[0228] Fusion: Here, it refers to the process of combining information or predictions from different sources in order to achieve results that are superior to any single source.

[0229] First, a properly trained deep learning model is used to make a preliminary estimate of bandwidth utilization over a period of time in the future. Next, an ARIMA model is built and trained based on the same historical data set to generate another set of forecasts for future bandwidth utilization. The last step is to combine the results obtained from the two models according to certain rules (such as weighted average) to form the final forecast output.

[0230] In an embodiment of the present application, it is assumed that a data center wants to estimate the total bandwidth demand provided by its server cluster one week in advance. According to the average daily bandwidth usage records of the past year, a deep learning model based on the LSTM (Long Short-Term Memory) network and an ARIMA model with parameters set to (5, 1, 0) were established. The LSTM model predicts that the average daily bandwidth demand in the next 7 days will be approximately 150 Gbps, while the result given by the ARIMA model is slightly lower, at 140 Gbps. Taking into account that LSTM is better at capturing long-term trends, but may ignore some regular changes in the short term, it was decided to give the two equal weights for simple averaging, that is, to predict the bandwidth. As the final forecast of average daily bandwidth demand for the next week.

[0231] Through the above fusion strategy, we can not only make full use of the powerful representation ability of deep learning to understand complex patterns, but also use the ARIMA model to better cope with regular changes in the short term, so that the overall prediction performance is significantly improved and closer to the actual situation. This helps enterprises to rationally plan resource allocation, optimize cost expenditure, and effectively avoid problems such as waste caused by over-configuration or deterioration of service quality caused by insufficient configuration.

[0232] This application takes into account that when predicting bandwidth utilization, although deep learning models can capture long-term trends and complex patterns well, they may not be sensitive enough to short-term fluctuations, especially for those sudden outliers or rapidly changing situations. On the other hand, the ARIMA model is good at processing data with obvious time series characteristics, such as seasonal and trend components, but is weak in the face of nonlinear relationships. In addition, a single model often finds it difficult to balance accuracy and stability at the same time. Therefore, in order to improve the prediction accuracy and enhance the model's adaptability to different situations, this application proposes a fusion method that combines an improved ARIMA model and a dynamic weight adjustment mechanism.

[0233] The options are as follows:

[0234] Optionally, the using of the ARIMA model to perform time series analysis on the bandwidth utilization data to predict a change trend of the bandwidth utilization data in a future period of time includes:

[0235] Forecasting using the ARIMA model:

[0236] ;

[0237] in, It's in time At , the predicted value of bandwidth utilization data by ARIMA model is, is the ARIMA model at time The predicted value of is the rate of change of the ARIMA model forecast, is the adjustment factor;

[0238] The step of fusing the change trend of the bandwidth utilization data predicted by the ARIMA model with the change trend of the bandwidth utilization data predicted by the deep learning model to obtain the change trend of the adjusted bandwidth utilization data includes:

[0239] Fusion prediction results:

[0240] ;

[0241] in, It's in time The comprehensive prediction value obtained by integrating the deep learning model and the ARIMA model is is the dynamically adjusted fusion coefficient, defined as:

[0242] ;

[0243] in, and are the standard deviations of the predictions from the deep learning model and the ARIMA model, is the adjustment factor.

[0244] First, the traditional ARIMA model is used to make a preliminary estimate of future bandwidth utilization, and an additional term is introduced To consider the changing trend of the predicted value, in order to more accurately reflect the actual changes. Then, based on the uncertainty (i.e. standard deviation) of the prediction results of the two models, dynamically determine the proportion of the two in the final prediction. When the deep learning model shows lower uncertainty, it is given a higher weight; vice versa. The last step is to combine the results obtained in the above two steps according to a specific formula to form a unified and optimized prediction plan.

[0245] In an embodiment of the present application, it is assumed that a telecom operator wants to estimate the average daily bandwidth demand of its backbone network in the next week. Based on the historical data of the past year, an LSTM network was constructed as a deep learning model, and an ARIMA (5,1,0) model was set. After training, it was found that the average daily bandwidth demand given by the LSTM model on the test set was about 160Gbps, while the ARIMA model predicted 145Gbps. Further analysis shows that the standard deviation of the prediction results of the LSTM model is smaller (indicating that its prediction is more stable), while that of the ARIMA model is larger.

[0246] set up . According to the given conditions, we can calculate:

[0247] For the ARIMA model, Assume that the simplified processing is not given in detail here The specific value of ARIMA is mainly based on the original prediction value.

[0248] calculate , assuming that at this time ,but

[0249] Final fusion prediction results ;

[0250] In this way, not only can the advantages of each model be fully utilized, but also the limitations of a single model can be alleviated to a certain extent. In particular, the dynamic adjustment coefficient method can automatically adjust the contribution of each model according to the actual situation, making the overall prediction closer to the actual situation, helping enterprises to better plan resources, avoid waste, and ensure service quality.

[0251] Optionally, determining the target monitoring node from multiple monitoring nodes deployed on the transmission path on the network side of the base station to be tested includes: receiving a test data stream on the transmission path on the network side of the base station to be tested, and determining the transmission direction of the test data stream, wherein multiple monitoring nodes are deployed on the transmission path, and the monitoring nodes are distributed at key positions on the transmission path; according to the transmission direction of the test data stream, selecting a target monitoring node located at a specific section on the corresponding transmission path from the multiple monitoring nodes.

[0252] Base station under test: refers to the mobile communication base station that needs to be tested for performance or troubleshooting.

[0253] Network-side transmission path: The communication path consisting of all connection lines from the user device to the core network.

[0254] Monitoring node: A hardware or software entity deployed in the network specifically for collecting and analyzing traffic information.

[0255] Test data flow: A collection of data packets that flows in a network for testing purposes.

[0256] Transmission direction: the direction in which data flows in the network, such as from the base station to the core network or vice versa.

[0257] Target monitoring node: A monitoring point selected according to test requirements that can provide the most valuable information.

[0258] First, the system receives the test data stream sent along a specific transmission path and identifies the direction from which the data stream enters the network. Based on the data stream direction determined above, the system selects the monitoring points on the corresponding transmission path that best reflect the problem characteristics or help diagnose the problem location as the target monitoring nodes from a series of preset monitoring nodes.

[0259] In an embodiment of the present application, it is assumed that an operator wants to evaluate the network delay between 4G base station A and the core network in a certain area. To this end, they plan to use data streams generated by simulating user access behavior for testing. Four monitoring nodes M1, M2, M3, and M4 have been pre-set in this area, located in different positions, covering the entire transmission path from base station A to the core network. Assume that when the test starts, the data stream starts from base station A and is transmitted in the direction of the core network. The system first receives these data streams and determines that they are sent from base station A to the core network. Next, based on this transmission direction and the positional relationship of each monitoring node relative to base station A and the core network, it is decided to use M2 and M3 as the main monitoring objects for this test. The reason for selecting these two nodes is that they are in the important intermediate links through which the data stream passes, and can effectively reflect the link status.

[0260] Through this method, operators can not only locate potential problem sources more accurately, but also avoid resource consumption caused by extensive collection of useless data. This can not only improve operation and maintenance efficiency, but also ensure that limited monitoring resources are used where they are needed, thereby better supporting the quality assurance of network services. For example, in the case mentioned above, through data analysis of the two monitoring nodes M2 and M3, technicians quickly discovered the specific cause of the high latency - a section of optical fiber between M2 and M3 was slightly damaged, and then took timely measures to repair it, avoiding a more serious decline in network service quality.

[0261] Figure 2 A schematic diagram of the structure of a communication performance test system for a base station network side transmission path is provided for an embodiment of the present application, such as Figure 2 As shown, the device comprises:

[0262] The monitoring module 21 is used to determine a target monitoring node from a plurality of monitoring nodes deployed on the transmission path of the network side of the base station to be tested, and to monitor the transmission delay data, packet loss rate data and bandwidth utilization data of the received test data stream on a specific section of the transmission path in real time through the target monitoring node;

[0263] An analysis module 22, used for analyzing the transmission delay data to detect periodic and non-periodic changes of the transmission delay and obtain an analysis result;

[0264] An evaluation module 23, used for processing the packet loss rate data to evaluate the impact of the packet loss rate data on transmission performance and obtain an evaluation result;

[0265] A prediction module 24, configured to analyze the bandwidth utilization data, predict a change trend of the bandwidth utilization data, and obtain a prediction result;

[0266] A calculation module 25, used to combine the analysis result, the evaluation result and the prediction result to calculate the communication performance index of the specific road section;

[0267] The processing module 26 is used to compare the communication performance index with the preset communication performance standard to evaluate whether the communication performance of the specific road section meets the preset standard, and generate a communication performance test report of the transmission path according to the comparison result, and provide optimization suggestions.

[0268] Figure 2 The communication performance test system of the base station network side transmission path can be performed Figure 1 The implementation principle and technical effect of the communication performance test method of the transmission path on the network side of a base station described in the embodiment shown are not repeated here. The specific way in which each module and unit performs operations in the communication performance test system of the transmission path on the network side of a base station in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated here.

[0269] In one possible design, Figure 2 A communication performance test system for a base station network side transmission path in the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0270] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0271] The processing component 32 is used to: determine the target monitoring node from multiple monitoring nodes deployed on the transmission path of the network side of the base station to be tested, and monitor the transmission delay data, packet loss rate data and bandwidth utilization data of the test data stream on the specific section of the transmission path in real time; detect the periodic and non-periodic changes of the transmission delay to obtain analysis results; evaluate the impact of the packet loss rate data on the transmission performance to obtain evaluation results; predict the change trend of the bandwidth utilization data to obtain prediction results; combine the three results to calculate the communication performance index of the specific section; generate a communication performance test report based on the comparison result of whether the communication performance of the specific section meets the predetermined standard.

[0272] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0273] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0274] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0275] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0276] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0277] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0278] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 A method for testing the communication performance of a transmission path on a base station network side is provided in the illustrated embodiment.

[0279] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0280] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0281] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0282] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for testing the communication performance of a transmission path on a base station network side, characterized in that: include: Determine a target monitoring node from a plurality of monitoring nodes deployed on the transmission path of the network side of the base station to be tested, and monitor the transmission delay data, packet loss rate data and bandwidth utilization data of the received test data stream on a specific section of the transmission path in real time through the target monitoring node; Analyzing the transmission delay data to detect periodic and non-periodic changes in the transmission delay and obtaining analysis results; Processing the packet loss rate data to evaluate the impact of the packet loss rate data on transmission performance and obtain an evaluation result; Analyze the bandwidth utilization data, predict the change trend of the bandwidth utilization data, and obtain a prediction result; Combining the analysis result, the evaluation result and the prediction result, calculating the communication performance index of the specific road section; Comparing the communication performance index with a preset communication performance standard to evaluate whether the communication performance of the specific road section meets the preset standard, and generating a communication performance test report of the transmission path according to the comparison result, and providing optimization suggestions; The analyzing the transmission delay data to detect periodic and non-periodic changes of the transmission delay and obtaining analysis results includes: Constructing a chaotic time series, converting the transmission delay data into a chaotic time series, for analyzing the dynamic characteristics corresponding to the transmission delay data; Calculating the embedding dimension and delay time of the chaotic time series to reconstruct the phase space, and applying the phase space reconstruction technology to reconstruct the chaotic time series into a phase space trajectory according to the embedding dimension and the delay time; The stability of the phase space trajectory is analyzed by using the Lyapunov exponent in chaos theory to detect the non-periodic changes of the transmission delay; Analyzing the frequency spectrum of the chaotic time series by Fourier transform to detect the periodic change of transmission delay; Using wavelet transform to analyze the local characteristics of the chaotic time series, detecting the changes of transmission delay on different time scales, and obtaining wavelet transform analysis results; The Lyapunov exponent analysis result and the Fourier transform spectrum analysis result are corrected based on the wavelet transform analysis result to obtain the adjusted periodic and non-periodic changes of the transmission delay, and an analysis result is generated according to the adjusted periodic and non-periodic changes of the transmission delay.

2. The method according to claim 1, characterized in that The processing of the packet loss rate data to evaluate the impact of the packet loss rate data on transmission performance and obtain an evaluation result includes: Defining fuzzy sets of different ranges according to the packet loss rate data, wherein the fuzzy sets include a low packet loss rate, a medium packet loss rate, and a high packet loss rate; Construct a fuzzy logic system, set the input variable as packet loss rate data, the output variable as the transmission performance impact level, and set the fuzzy logic rules corresponding to the fuzzy logic system according to the fuzzy set corresponding to the packet loss rate data, wherein the fuzzy logic rules include: if the packet loss rate data belongs to a low packet loss rate, the transmission performance impact level is a slight impact; if the packet loss rate data belongs to a medium packet loss rate, the transmission performance impact level is a medium impact; if the packet loss rate data belongs to a high packet loss rate, the transmission performance impact level is a severe impact; Inputting the packet loss rate data into a fuzzy logic system, performing reasoning according to set fuzzy logic rules, and obtaining a transmission performance impact level; The transmission performance impact level is converted into a transmission performance impact level value by using a defuzzification process, and an impact evaluation result of the packet loss rate on the transmission performance is determined according to the transmission performance impact level value.

3. The method according to claim 1, characterized in that The analyzing the bandwidth utilization data, predicting the change trend of the bandwidth utilization data, and obtaining the prediction result includes: Constructing a training data set, dividing the bandwidth utilization data into a training set and a test set, and selecting and initializing a deep learning model; Using the training set to train the deep learning model, optimizing the model parameters by a back propagation algorithm, and using the test set to verify the trained deep learning model to evaluate the prediction performance of the deep learning model; Use the trained deep learning model to predict the real-time monitored bandwidth utilization data to obtain the change trend of the bandwidth utilization data; The expression for predicting the change trend of bandwidth utilization data is: ； in, Indicates at time The predicted value of the bandwidth utilization data by the deep learning model is represents the predicted value of the deep learning model at time t+1, represents the parameters of the model, is the rate of change of the predicted value, is the adjustment factor.

4. The method according to claim 3, characterized in that After predicting the real-time monitored bandwidth utilization data using the trained deep learning model to obtain the change trend of the bandwidth utilization data, the method further includes: Use the ARIMA model to perform time series analysis on the bandwidth utilization data to predict the change trend of the bandwidth utilization data in the future; The change trend of the bandwidth utilization data predicted by the ARIMA model is integrated with the change trend of the bandwidth utilization data predicted by the deep learning model to obtain an adjusted change trend of the bandwidth utilization data.

5. The method according to claim 1, characterized in that The step of combining the analysis result, the evaluation result and the prediction result to calculate the communication performance index of the specific road section includes: Constructing a comprehensive evaluation model according to the analysis results, the evaluation results and the prediction results; Determine the weight coefficients of the analysis results, the evaluation results, and the prediction results according to the importance of each indicator; The analysis result, the evaluation result and the prediction result are weightedly fused to obtain a communication performance index for the specific road section.

6. The method according to claim 1, characterized in that The determining of the target monitoring node from a plurality of monitoring nodes deployed on the transmission path of the network side of the base station to be tested includes: Receive a test data stream on a transmission path on the network side of the base station to be tested, and determine a transmission direction of the test data stream, wherein a plurality of monitoring nodes are deployed on the transmission path, and the monitoring nodes are distributed at key positions of the transmission path; According to the transmission direction of the test data stream, a target monitoring node located at a specific section on the corresponding transmission path is selected from the multiple monitoring nodes.

7. A communication performance test system for a base station network side transmission path, characterized in that: include: A monitoring module, used to determine a target monitoring node from a plurality of monitoring nodes deployed on the transmission path of the network side of the base station to be tested, and to monitor the transmission delay data, packet loss rate data and bandwidth utilization data of the received test data stream on a specific section of the transmission path in real time through the target monitoring node; An analysis module, used for analyzing the transmission delay data to detect periodic and non-periodic changes of the transmission delay and obtain an analysis result; An evaluation module, used for processing the packet loss rate data to evaluate the impact of the packet loss rate data on transmission performance and obtain an evaluation result; A prediction module, used to analyze the bandwidth utilization data, predict the change trend of the bandwidth utilization data, and obtain a prediction result; A calculation module, used to combine the analysis result, the evaluation result and the prediction result to calculate the communication performance index of the specific road section; A processing module, used to compare the communication performance index with a preset communication performance standard to evaluate whether the communication performance of the specific road section meets the preset standard, and generate a communication performance test report of the transmission path according to the comparison result, and provide optimization suggestions; The analyzing the transmission delay data to detect periodic and non-periodic changes of the transmission delay and obtaining analysis results includes: Constructing a chaotic time series, converting the transmission delay data into a chaotic time series, for analyzing the dynamic characteristics corresponding to the transmission delay data; Calculating the embedding dimension and delay time of the chaotic time series to reconstruct the phase space, and applying the phase space reconstruction technology to reconstruct the chaotic time series into a phase space trajectory according to the embedding dimension and the delay time; The stability of the phase space trajectory is analyzed by using the Lyapunov exponent in chaos theory to detect the non-periodic changes of the transmission delay; Analyzing the frequency spectrum of the chaotic time series by Fourier transform to detect the periodic change of transmission delay; Using wavelet transform to analyze the local characteristics of the chaotic time series, detecting the changes of transmission delay on different time scales, and obtaining wavelet transform analysis results; The Lyapunov exponent analysis result and the Fourier transform spectrum analysis result are corrected based on the wavelet transform analysis result to obtain the adjusted periodic and non-periodic changes of the transmission delay, and an analysis result is generated according to the adjusted periodic and non-periodic changes of the transmission delay.

8. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a communication performance testing method for a base station network side transmission path as described in any one of claims 1 to 6.

9. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, a communication performance testing method for a base station network side transmission path as described in any one of claims 1 to 6 is implemented.

Images