An automated testing method, equipment, and medium for a broadband trunking communication system.

Abstract

This invention discloses an automated testing method, device, and medium for a broadband trunking communication system, relating to the field of communication network testing technology. The method includes: on the access side of the broadband trunking communication network, collecting voice group call service data, emergency video service data, and location information service data through edge computing nodes, and generating a collection window dataset under a unified time reference; jointly analyzing the latency characteristics, packet loss characteristics, and continuity characteristics in the collection window dataset to extract abnormal distribution characteristics and generate abnormal feature fingerprints; after generating the abnormal feature fingerprints, applying controllable perturbations during service operation, constructing a perturbation response curve, identifying abnormal change points, and generating an elastic threshold. This invention, by applying controllable perturbations during service operation and establishing a perturbation response curve, achieves dynamic identification of multi-dimensional communication data boundaries and can form an elastic threshold to adapt to real-time changes in complex scenarios.

Description

Technical Field

[0001] This invention relates to the field of communication network testing technology, and in particular to an automated testing method, equipment and medium for broadband trunking communication systems. Background Technology

[0002] Broadband trunked communication networks are dedicated communication networks for public safety, emergency dispatch, and industrial sites, requiring simultaneous support for voice group calling, emergency video services, and location information services. Because these services typically operate under high concurrency and complex environments, stringent requirements are placed on the real-time performance and stability of the communication process. Existing communication testing methods usually rely on edge node deployment, acquiring metrics such as communication latency, packet loss, and queuing depth through mechanisms like service triggering, data collection, and time alignment. These metrics are then combined with statistical analysis and distribution characteristic comparisons to achieve network performance testing, reflecting network operational quality to a certain extent and providing a foundation for operation, maintenance, and fault diagnosis.

[0003] However, conventional testing methods still have limitations in dynamic scenarios: on the one hand, anomaly identification is usually based on static threshold comparison, which makes it difficult to fully reflect the dynamic changes of communication latency data, access latency data, mouth-to-ear latency data, packet loss data, and queuing depth data in multiple dimensions, resulting in insufficient verification of anomaly authenticity; on the other hand, the reproduction process of anomaly sources lacks unified constraints, making it impossible to accurately locate the correspondence between operating parameters and anomaly distribution in various business environments, resulting in inaccurate anomaly attribution. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an automated testing method for broadband trunking communication systems, which solves the problems of insufficient verification of the authenticity of abnormal features in multi-dimensional dynamic scenarios and inaccurate location of abnormal sources.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides an automated testing method for a broadband trunking communication system, which includes collecting voice group call service data, emergency video service data and location information service data through edge computing nodes on the access side of the broadband trunking communication network, and generating a collection window dataset under a unified time reference.

[0008] Joint analysis is performed on the latency features, packet loss features, and continuity features in the dataset collected from the acquisition window to extract abnormal distribution features and generate abnormal feature fingerprints;

[0009] After generating the abnormal feature fingerprint, a perturbation response curve is constructed and abnormal change points are identified by applying controllable perturbations during business operation, thereby generating an elastic threshold.

[0010] Using the elastic threshold as a reproduction condition, the running parameters are configured within the elastic threshold constraint range to generate candidate reproduction scenarios. The candidate reproduction scenarios are then compared with the abnormal feature fingerprints, and the candidate reproduction scenario with the smallest difference is selected as the minimum reproduction scenario.

[0011] Based on the minimum reproducibility scenario, determine whether the abnormal feature fingerprint corresponds to the real abnormality, record the source of the real abnormality, and generate test conclusions.

[0012] As a preferred embodiment of the automated testing method for the broadband trunking communication system described in this invention, the acquisition of voice group call service data, emergency video service data, and location information service data includes, on the access side of the broadband trunking communication network, triggering the voice group call service, emergency video service, and location information service respectively at the start time of the acquisition window through an edge computing node.

[0013] In the voice group call service, access latency data, mouth-to-ear latency data, and queuing depth data are collected; in the emergency video service, communication latency data and packet loss data are collected; and in the location information service, the results of location information reporting sequence continuity checks are collected.

[0014] As a preferred embodiment of the automated testing method for the broadband trunking communication system described in this invention, the generation of the acquisition window dataset includes aligning and organizing voice group call service data, emergency video service data, and location information service data according to the order from the start time to the end time of the acquisition window under a unified time reference, to form the acquisition window dataset.

[0015] As a preferred embodiment of the automated testing method for the broadband trunking communication system described in this invention, the joint analysis of the latency characteristics, packet loss characteristics, and continuity characteristics in the data collection window includes sorting the communication latency data, access latency data, and mouth-to-ear latency data according to their numerical values ​​and extracting quantiles.

[0016] Linear interpolation was used to connect adjacent quantiles one by one to generate communication delay quantile curves, access delay quantile curves, and mouth-to-ear delay quantile curves, respectively.

[0017] When the communication delay quantile curve, access delay quantile curve, and mouth-to-ear delay quantile curve of adjacent acquisition windows show an overall shift and shape change, it is determined that the delay characteristics of the acquisition window dataset have changed.

[0018] Extract the number of missing packets and the time period of packet loss from the packet loss data, and arrange them in order of the time period of packet loss to form a packet loss cluster length sequence;

[0019] The time period of each packet loss in the packet loss cluster length sequence is compared one by one with the time periods with discontinuous numbers in the location information reporting sequence continuity check results to generate a confirmed packet loss cluster length sequence.

[0020] When the confirmed packet loss cluster length sequence shows multiple instances where the number of missing packets exceeds the size of a single normal packet loss, it is determined that the packet loss characteristics in the dataset collected in the acquisition window have changed.

[0021] Calculate the local maxima point by point for the queuing depth data, mark all local maxima as peaks, and record the time difference between adjacent peaks in sequence to form a sequence of adjacent peak intervals.

[0022] When the peak interval sequence shows a continuous shortening, it is determined that the continuity characteristic of the dataset in the acquisition window has changed;

[0023] When delay features, packet loss features, and continuity features change simultaneously between consecutive acquisition windows, it is determined that the dataset in the acquisition window has an abnormal distribution in terms of multidimensional features.

[0024] As a preferred embodiment of the automated testing method for the broadband trunking communication system described in this invention, the step of extracting abnormal distribution features and generating abnormal feature fingerprints includes arranging the numerical differences between the communication delay quantile curves, access delay quantile curves, and mouth-to-ear delay quantile curves of adjacent acquisition windows in the order of the acquisition windows to form communication delay distribution change vectors, access delay distribution change vectors, and mouth-to-ear delay change vectors.

[0025] The number of missing packets in each of the confirmed packet loss cluster length sequences is arranged in chronological order of packet loss occurrence to form a packet loss ratio vector.

[0026] The peak intervals of the queuing depth data are arranged sequentially according to the detection order to form a queuing rhythm vector;

[0027] The combined feature vector is formed by concatenating the communication delay distribution change vector, access delay distribution change vector, mouth-to-ear delay distribution change vector, packet loss ratio vector, and queuing rhythm vector in that order. The combined feature vector is then encoded using a one-way hash algorithm to generate anomaly feature fingerprints.

[0028] As a preferred embodiment of the automated testing method for the broadband trunking communication system described in this invention, the generation of elastic thresholds includes, during the operation of voice group call service and emergency video service, recording the voice group call service data and emergency video service data under various disturbance intensities of the data collection window dataset by changing the number of call right requests and the continuity of video data transmission, thereby forming a corresponding disturbance response pairing sequence.

[0029] Curve fitting is performed on each perturbation response pair sequence to generate perturbation response curves, and inflection points are identified based on the slope changes of the perturbation response curves to generate a preliminary set of identification points;

[0030] The initial set of identification points is compared with the abnormal feature fingerprint to obtain the final identification points;

[0031] In the final identification points, the disturbance intensity values ​​are sorted and selected. The disturbance intensity with the smallest value that is consistent with the time period of the abnormal feature fingerprint is selected and determined as the communication delay elastic threshold, access delay elastic threshold, mouth-to-ear delay elastic threshold, packet loss elastic threshold, and queuing depth elastic threshold, respectively.

[0032] As a preferred embodiment of the automated testing method for the broadband trunking communication system described in this invention, the minimum reproducibility scenario includes, in the voice group call service, generating a first type of candidate reproducibility scenario by configuring the number of call right request attempts and constraining them within the access delay elasticity threshold, the mouth-to-ear delay elasticity threshold, and the queuing depth elasticity threshold.

[0033] In emergency video services, by configuring the stability of video data transmission and constraining it within the elastic thresholds of communication latency and packet loss, a second type of candidate reproducible scenario is generated.

[0034] All candidate reproducible scenarios of the first type and candidate reproducible scenarios are gathered into a candidate reproducible scenario set, and the candidate reproducible scenario set is compared with the abnormal feature fingerprint one by one;

[0035] The Euclidean distance method was used to calculate the difference between access latency data, mouth-to-ear latency data, and queuing depth data and abnormal feature fingerprints in the first type of candidate reproducibility scenario, and the difference between communication latency data and packet loss data and abnormal feature fingerprints in the second type of candidate reproducibility scenario. The candidate reproducibility scenario with the smallest difference was selected as the minimum reproducibility scenario.

[0036] As a preferred embodiment of the automated testing method for the broadband trunking communication system described in this invention, the step of determining whether the abnormal feature fingerprint corresponds to a real anomaly includes comparing the communication latency data, access latency data, mouth-to-ear latency data, packet loss data, and queuing depth data in the minimum reproducibility scenario with the corresponding elastic threshold constraints. When all the comparison results are within the constraint range, the abnormal feature fingerprint is determined to correspond to a real anomaly.

[0037] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, and the computer program, when executed by the processor, implements any step of the automated testing method for a broadband trunking communication system as described in the first aspect of the present invention.

[0038] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the automated testing method for a broadband trunking communication system as described in the first aspect of the present invention.

[0039] The beneficial effects of this invention are as follows: by applying controllable disturbances and establishing disturbance response curves during business operations, dynamic identification of multidimensional communication data boundaries is achieved, thereby forming elastic thresholds to adapt to real-time changes in complex scenarios; by applying the elastic thresholds to the configuration of operating parameters and generating reproducible scenarios, the precise correspondence between abnormal distributions and operating conditions is achieved, enabling the source of anomalies to be clearly located, and ultimately achieving accurate diagnosis and reliable conclusion output of the operating status of broadband trunking communication networks. Attached Figure Description

[0040] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 A flowchart for an automated testing method for a broadband trunking communication system.

[0042] Figure 2 A flowchart for generating the dataset for the acquisition window.

[0043] Figure 3 The flowchart for generating anomalous feature fingerprints.

[0044] Figure 4 A flowchart for generating the elastic threshold. Detailed Implementation

[0045] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0046] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0047] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0048] Reference Figures 1-4 As an embodiment of the present invention, this embodiment provides an automated testing method for a broadband trunking communication system, comprising the following steps:

[0049] S1. On the access side of the broadband trunking communication network, voice group call service data, emergency video service data and location information service data are collected through edge computing nodes, and a collection window dataset is generated under a unified time reference.

[0050] Furthermore, broadband trunking communication networks are communication networks oriented towards public safety, emergency dispatch, and industrial sites. They typically include voice group calling services, emergency video services, and location information services. Edge computing nodes are deployed on the access side of the broadband trunking communication network, with the edge computing nodes located between the terminal equipment and the core network. The edge computing nodes have service triggering functions, data acquisition functions, and time synchronization functions.

[0051] The service triggering function is used to start voice group call service, emergency video service and location information service within the collection window; the data collection function is used to record service data during service operation, including voice group call service data, emergency video service data and location information service data; the time synchronization function is used to establish a unified time base at the beginning of the collection window and align all service data on the same time axis.

[0052] Furthermore, voice group call service data includes access latency data, mouth-to-ear latency data, and queuing depth data; emergency video service data includes communication latency data and packet loss data; and location information service data refers to the results of location information reporting sequence continuity checks.

[0053] To ensure complete recording of service data, the collection window duration is set on the edge computing nodes of the broadband trunking communication network based on the typical duration of voice group call services, emergency video services, and location information services. Specifically, the collection window duration for voice group call services must cover a complete call process, for emergency video services it must cover a continuous video transmission process, and for location information services it must include at least two periodic reports. Insufficient collection window duration will cause the call process of voice group call services to be truncated, resulting in incomplete recording of access latency data and mouth-to-ear latency data. Excessive collection window duration will cause the video transmission process of emergency video services to be repeatedly superimposed, leading to deviations in the statistical results of communication latency data and packet loss data, and increasing data processing overhead. Therefore, the collection window duration is set to cover the duration of a complete voice group call, a continuous emergency video, and at least two location information reports simultaneously.

[0054] It should be noted that on the edge computing nodes of the broadband trunking communication network, voice group call service, emergency video service and location information service are triggered at the start time of the acquisition window.

[0055] When the voice group call service is triggered at the start time of the collection window, the edge computing node records the call right request time and the permission issuance time during the voice group call service. By calculating the difference between the call right request time and the permission issuance time, the access latency data is obtained. During voice transmission, the voice transmission time and the voice playback time are recorded. By calculating the difference between the voice transmission time and the voice playback time, the mouth-to-ear latency data is obtained. During the call right control process, the call right queue length is continuously recorded, and the call right queue lengths are arranged into a call right queue length sequence according to the time order. The call right queue length sequence is the queue depth data.

[0056] When the emergency video service is triggered at the start time of the acquisition window, the edge computing node records the video data transmission time and video data reception time during the emergency video service. By calculating the difference between the video data transmission time and video data reception time, the communication latency data is obtained. At the video data receiving end, each received video data packet is arranged according to its sequential number, and the arrangement result is compared with the transmission sequence number one by one. When a jump or missing number is detected in the reception sequence number, the start and end range of the missing number and the number of missing packets are calculated, and the time period of packet loss is recorded. The records containing the start and end range of the missing number, the number of missing packets, and the time period of packet loss are arranged in chronological order to generate packet loss data.

[0057] When the location information service is triggered at the start time of the acquisition window, the edge computing node extracts the sequence number of the location information reported during the location information service process and compares whether adjacent sequence numbers are continuous. When a discontinuous sequence number is detected, the start time and end time of the discontinuity are recorded, and a location information reporting sequence continuity check result is generated accordingly.

[0058] It should also be noted that at the end of the collection window, the edge computing nodes, through the time synchronization function, establish a unified time base and align and organize the voice group call service data, emergency video service data, and location information service data in the order from the start time to the end time of the collection window to form the collection window dataset.

[0059] S2. Perform joint analysis on the latency features, packet loss features, and continuity features in the dataset collected in the acquisition window to extract abnormal distribution features and generate abnormal feature fingerprints.

[0060] Furthermore, the communication delay data, access delay data, and mouth-to-ear delay data in the data set of the acquisition window are sorted in ascending order of numerical value to obtain ordered sequences of communication delay, access delay, and mouth-to-ear delay.

[0061] Data points located at a certain proportion of low, middle and high positions are extracted from each ordered sequence as quantiles; the quantiles in the ordered sequence of communication delay, access delay and mouth-to-ear delay are arranged in proportion, and adjacent quantiles are connected one by one by linear interpolation to generate communication delay quantile curves, access delay quantile curves and mouth-to-ear delay quantile curves respectively.

[0062] The three quantile curves obtained from the current acquisition window are compared one by one with the three quantile curves obtained from the adjacent acquisition windows. When the comparison results show that there is an overall translation or morphological change between consecutive acquisition windows, it is determined that the time delay characteristics in the dataset of the acquisition window have changed, thus indicating that the distribution characteristics of communication time delay data, access time delay data and mouth-to-ear time delay data have become abnormal.

[0063] Furthermore, in the dataset collected from the packet loss data, the number of missing packets and the time period in which the packet loss occurred are extracted, and the number of missing packets is used as the packet loss cluster length; all packet loss cluster lengths are arranged in order of the time period in which the packet loss occurred to form a packet loss cluster length sequence.

[0064] The time period of each packet loss in the packet loss cluster length sequence is compared one by one with the time periods with discontinuous numbers in the location information reporting sequence continuity check results. When the time period of each packet loss in the packet loss cluster length sequence matches the time period with discontinuous numbers in the location information reporting sequence continuity check results, the packet loss data is deemed valid, and a confirmed packet loss cluster length sequence is generated.

[0065] When the confirmed packet loss cluster length sequence shows multiple instances within a collection window where the number of missing packets exceeds the normal size of a single packet loss, such as when the number of missing packets in a single instance exceeds five packets, it is determined that the packet loss characteristics in the dataset within the collection window have changed, indicating that the size or frequency of missing packets in the packet loss data within the collection window has become abnormal.

[0066] Furthermore, in the dataset collected from the collection window, local maxima are calculated point by point in the queuing depth data according to the time order, and all local maxima are marked as peaks; the time difference between adjacent peaks is recorded in sequence to form a sequence of adjacent peak intervals.

[0067] When the interval between adjacent spikes gradually shortens, it indicates that the continuity characteristic of the dataset in the acquisition window has changed, thus indicating that the rhythm of the queuing depth data within the acquisition window has become abnormal.

[0068] Furthermore, when the distribution characteristics of the communication delay quantile curve, access delay quantile curve, and mouth-to-ear delay quantile curve change between consecutive acquisition windows, and the confirmed packet loss cluster length sequence indicates that the packet loss data shows an abnormal scale or frequency of packet loss within the acquisition window, and the peak interval sequence of queuing depth data indicates that the change rhythm of queuing depth data within the acquisition window is abnormal, it is determined that the acquisition window dataset has changed in multidimensional features, thus indicating that there is an abnormal distribution in the acquisition window dataset.

[0069] It should be noted that when the dataset in the acquisition window has an abnormal distribution in multidimensional features, the abnormal distribution features are extracted from the dataset in the acquisition window and an abnormal feature fingerprint is generated.

[0070] Between two adjacent acquisition windows, the numerical difference of the communication delay quantile curve at the same quantile point is taken as the offset of the communication delay quantile curve, the numerical difference of the access delay quantile curve at the same quantile point is taken as the offset of the access delay quantile curve, and the numerical difference of the mouth-to-ear delay quantile curve at the same quantile point is taken as the offset of the mouth-to-ear delay quantile curve.

[0071] The offsets of the communication delay quantile curves between consecutive acquisition windows are arranged sequentially according to the time order of the acquisition windows. The arrangement result forms a communication delay distribution change vector, which is used to describe the distribution change of communication delay between adjacent acquisition windows. Similarly, the offsets of the access delay quantile curves between consecutive acquisition windows are arranged sequentially according to the time order of the acquisition windows. The arrangement result forms an access delay distribution change vector, which is used to describe the distribution change of access delay between adjacent acquisition windows. Finally, the offsets of the mouth-to-ear delay quantile curves between consecutive acquisition windows are arranged sequentially according to the time order of the acquisition windows. The arrangement result forms a mouth-to-ear delay distribution change vector, which is used to describe the distribution change of mouth-to-ear delay between adjacent acquisition windows.

[0072] The number of missing packets in each of the confirmed packet loss cluster length sequences is arranged in chronological order of packet loss occurrence. The arrangement results form a packet loss ratio vector, which is used to describe the scale and frequency of packet loss within the acquisition window.

[0073] The peak intervals of the queuing depth data are arranged sequentially according to the detection order. The arrangement results form a queuing rhythm vector, which is used to describe the rhythmic changes of queuing depth within the acquisition window.

[0074] The combined feature vector is obtained by concatenating the communication delay distribution change vector, access delay distribution change vector, mouth-to-ear delay distribution change vector, packet loss ratio vector, and queuing rhythm vector in that order. The combined feature vector is then encoded using a public one-way hash algorithm (e.g., SHA-256). The encoding result is an identifier sequence, the length of which can be adjusted according to the size of the dataset in the acquisition window and the required representation accuracy. The identifier sequence is the anomaly fingerprint. The anomaly fingerprint can simultaneously reflect multi-dimensional anomaly features of communication delay distribution, access delay distribution, mouth-to-ear delay distribution, packet loss, and queuing rhythm.

[0075] S3. After the abnormal feature fingerprint is generated, a perturbation response curve is constructed and abnormal change points are identified by applying controllable perturbations during business operation, and an elastic threshold is generated.

[0076] After the abnormal feature fingerprint is generated, in order to further confirm the source of the abnormal features, a controllable disturbance is applied to the service traffic of the broadband trunking communication network based on the communication latency data, access latency data, mouth-to-ear latency data, packet loss data and queuing depth data in the acquisition window dataset.

[0077] During the operation of voice group call services, the frequency of call right requests is changed, setting the call right requests to once, twice, and three or more times within a continuous service cycle, so that the call right requests manifest as low-frequency, medium-frequency, and high-frequency requests during service operation, thereby creating different levels of operational load. During the operation of emergency video services, the continuity of video data transmission is changed, setting the video transmission behavior to stable transmission, limited fluctuation transmission, and large fluctuation transmission, respectively. Stable transmission uses a fixed bit rate and uniform transmission, limited fluctuation transmission uses a constrained variable bit rate and small transmission jitter, and large fluctuation transmission uses an adaptive bit rate, and alternates between burst transmission and pause transmission, thereby creating different levels of data transmission fluctuation.

[0078] Each change in the number of call right requests or the continuity of video data transmission constitutes a disturbance operation. After the disturbance operation is completed, the results of communication delay data, access delay data, mouth-to-ear delay data, packet loss data, and queuing depth data under the corresponding disturbance intensity are immediately recorded in the dataset of the acquisition window. The correspondence between the disturbance intensity and the target variable is obtained, namely the correspondence between communication delay disturbance intensity and communication delay data, access delay disturbance intensity and access delay data, mouth-to-ear delay disturbance intensity and mouth-to-ear delay data, packet loss disturbance intensity and packet loss data, and queuing depth disturbance intensity and queuing depth data.

[0079] As the disturbance intensity increases step by step, multiple sets of corresponding results between disturbance intensity and target variable are obtained continuously, eventually forming a pairing sequence of communication delay disturbance response, access delay disturbance response, mouth-to-ear delay disturbance response, packet loss disturbance response, and queuing depth disturbance response.

[0080] Furthermore, curve fitting was performed on the paired sequences of communication delay disturbance response, access delay disturbance response, mouth-to-ear delay disturbance response, packet loss disturbance response, and queuing depth disturbance response to obtain the continuous relationship between disturbance intensity and target variable.

[0081] Each disturbance intensity and its corresponding delay data in the communication delay disturbance response pairing sequence, access delay disturbance response pairing sequence, and mouth-to-ear delay disturbance response pairing sequence are used as a set of fitting sample points. The least squares method is used to establish a continuous functional relationship between the disturbance intensity and the communication delay data. The packet loss disturbance response pairing sequence and the queuing depth disturbance response pairing sequence are subjected to the same fitting process to obtain the continuous functional relationships between the disturbance intensity and the packet loss data, and between the disturbance intensity and the queuing depth data, expressed as follows:

[0082] ;

[0083] ;

[0084] ;

[0085] in, Represents the time delay function, when Time represents the communication delay function, when The time indicates the access delay function. Time represents the mouth-to-ear time delay function. Function to indicate packet loss. The queuing depth function is represented by the following: Indicates the first The disturbance intensity set in this disturbance operation has a value range determined by the method of disturbance application. For example, the number of call right requests for a voice group call service is between 1 and 3. This represents the square of the disturbance intensity, used for weighting the disturbance intensity in least squares calculations. Indicates the first The delay data obtained under the second perturbation operation, when Time represents communication delay data, when The time indicates the access latency data. The time represents the mouth-to-ear delay data. Indicates the first Packet loss data obtained under this perturbation operation. Indicates the first Queue depth data obtained under the perturbation operation. Indicates the sequence number of the perturbation operation. This indicates the total number of disturbance operations.

[0086] After obtaining the communication delay function, access delay function, mouth-to-ear delay function, packet loss function, and queuing depth function, plot the communication delay disturbance response curve, access delay disturbance response curve, mouth-to-ear delay disturbance response curve, packet loss disturbance response curve, and queuing depth disturbance response curve respectively, with the disturbance intensity as the x-axis and the function values ​​of the communication delay function, access delay function, mouth-to-ear delay function, packet loss disturbance response curve, and queuing depth disturbance response curve.

[0087] Furthermore, inflection point detection is performed on the communication delay disturbance response curve, access delay disturbance response curve, mouth-to-ear delay disturbance response curve, packet loss disturbance response curve, and queuing depth disturbance response curve to determine the preliminary identification point of the disturbance intensity. Specifically, each disturbance response curve is discretized according to the sampling points of the disturbance intensity, and the slope between adjacent sampling points is calculated for each pair of adjacent sampling points, expressed as:

[0088] ;

[0089] in, Indicates the intensity of the disturbance With disturbance intensity The slope of adjacent sampling points between them Indicates the first The perturbation intensity set in the next perturbation operation. Indicates the intensity of the disturbance The target variable function value, This could be a communication delay function, access delay function, mouth-to-ear delay function, packet loss function, or queuing depth function. The function value at that point, Indicates the intensity of the disturbance The target variable function value.

[0090] After obtaining the slope of each pair of adjacent sampling points, the slopes of all adjacent sampling points are arranged in order of disturbance intensity to form a slope sequence, including the slope sequence of the communication delay disturbance response curve, the slope sequence of the access delay disturbance response curve, the slope sequence of the mouth-to-ear delay disturbance response curve, the slope sequence of the packet loss disturbance response curve, and the slope sequence of the queuing depth disturbance response curve.

[0091] To identify the intensity of slope changes between adjacent sampling points, a difference operation is performed on each slope sequence according to the perturbation intensity. The difference between adjacent slopes is taken as the slope change, expressed as:

[0092] ;

[0093] in, Indicates the intensity of the disturbance The change in slope in the vicinity Indicates the intensity of the disturbance With disturbance intensity The slope of adjacent sampling points between them is used to detect the perturbation response curve at different perturbation intensities. Are there any inflection points nearby?

[0094] When satisfied and At that time, the disturbance intensity The record is the inflection point, among which Indicates the intensity of the disturbance The change in slope in the vicinity Indicates the intensity of the disturbance The slope change in the vicinity; among all the disturbance intensities that meet the conditions, collect the corresponding disturbance intensities in ascending order, and respectively form the preliminary identification points set for communication delay, access delay, mouth-to-ear delay, packet loss, and queuing depth.

[0095] Furthermore, the preliminary identification points for communication delay, access delay, mouth-to-ear delay, packet loss, and queuing depth are compared one by one with the communication delay distribution change vector, access delay distribution change vector, mouth-to-ear delay distribution change vector, packet loss ratio vector, and queuing rhythm vector contained in the abnormal feature fingerprint.

[0096] When the time period corresponding to the disturbance intensity in the preliminary identification point set is consistent with the time period of abnormal change reflected in the abnormal feature fingerprint, the disturbance intensity in the preliminary identification point set is taken as the final identification point; for example, in the preliminary identification point set of communication delay, when the time period corresponding to the disturbance intensity in the preliminary identification point set is consistent with the time period of abnormal change represented by the communication delay distribution change vector in the abnormal feature fingerprint, the disturbance intensity in the preliminary identification point set of communication delay is taken as the final identification point of communication delay.

[0097] Among the final identification points of various disturbance response curves, the disturbance intensity of all final identification points is sorted by numerical value. The disturbance intensity with the smallest value that is consistent with the time period of the abnormal feature fingerprint is selected as the corresponding elastic threshold. For example, among the final identification points of the communication delay disturbance response curve, after sorting the disturbance intensity of all final identification points by numerical value, the disturbance intensity with the smallest value in the communication delay disturbance response curve that is consistent with the time period of the abnormal change corresponding to the communication delay distribution change vector is determined as the communication delay elastic threshold.

[0098] Finally, the following thresholds are generated: communication latency elasticity threshold, access latency elasticity threshold, mouth-to-ear latency elasticity threshold, packet loss elasticity threshold, and queuing depth elasticity threshold.

[0099] S4. Using the elastic threshold as a reproduction condition, configure the running parameters within the elastic threshold constraint range, generate candidate reproduction scenarios, compare the candidate reproduction scenarios with the abnormal feature fingerprint, and select the candidate reproduction scenario with the smallest difference as the minimum reproduction scenario.

[0100] Furthermore, the communication latency elasticity threshold, access latency elasticity threshold, mouth-to-ear latency elasticity threshold, packet loss elasticity threshold, and queuing depth elasticity threshold are used as reproduction conditions.

[0101] A correspondence is established between the communication latency elasticity threshold and the communication latency data to limit the maximum allowable range of variation in the communication latency data; a correspondence is also established between the access latency elasticity threshold and the access latency data to limit the maximum allowable range of variation in the access latency data; a correspondence is established between the mouth-to-ear latency elasticity threshold and the mouth-to-ear latency data to limit the maximum allowable range of variation in the mouth-to-ear latency data; a correspondence is established between the packet loss elasticity threshold and the packet loss data to limit the maximum allowable proportion of missing packets in the packet loss data; and a correspondence is established between the queuing depth elasticity threshold and the queuing depth data to limit the maximum allowable rhythmic variation amplitude of the queuing depth data.

[0102] The correspondences between the communication latency elasticity threshold and communication latency data, the correspondences between the access latency elasticity threshold and access latency data, the correspondences between the mouth-to-ear latency elasticity threshold and mouth-to-ear latency data, the correspondences between the packet loss elasticity threshold and packet loss data, and the correspondences between the queuing depth elasticity threshold and queuing depth data are arranged in order of category to form elasticity threshold constraints.

[0103] Furthermore, the elastic threshold constraint is used as an operational constraint. In the voice group call service, the operational parameters for the number of call right requests are configured based on the constraint range of the access delay elastic threshold, the mouth-to-ear delay elastic threshold, and the queuing depth elastic threshold. When the configuration result of the operational parameters meets the constraint range defined by the access delay elastic threshold, the mouth-to-ear delay elastic threshold, and the queuing depth elastic threshold, the first type of candidate reproducible scenario is generated.

[0104] In emergency video services, operating parameters for video data transmission stability are configured based on the constraints of the communication delay elasticity threshold and the packet loss elasticity threshold. When the configuration result of the operating parameters meets the constraints defined by the communication delay elasticity threshold and the packet loss elasticity threshold, a second type of candidate reconstructed scene is generated.

[0105] All candidate reproducible scenarios of the first category and the candidate reproducible scenarios of the second category are collected in the order of their generation to form a candidate reproducible scenario set.

[0106] Furthermore, the candidate reproducible scenario set is compared one by one with the abnormal feature fingerprint.

[0107] The access latency data, mouth-to-ear latency data, and queuing depth data in each first-type candidate reproducible scenario are arranged in the order of the acquisition window to form access latency candidate sequences, mouth-to-ear latency candidate sequences, and queuing depth candidate sequences, respectively. These sequences are then matched one-to-one with the access latency distribution change vector, mouth-to-ear latency distribution change vector, and queuing rhythm vector in the abnormal feature fingerprint. The Euclidean distance method is used to calculate the difference between the access latency candidate sequence and the access latency distribution change vector, the difference between the mouth-to-ear latency candidate sequence and the mouth-to-ear latency distribution change vector, and the difference between the queuing depth candidate sequence and the queuing rhythm vector, respectively. When the difference value is small, it indicates that the first-type candidate reproducible scenario and the abnormal feature fingerprint have a high degree of agreement in the corresponding dimension.

[0108] The communication latency data and packet loss data in each second-type candidate reproducible scenario are arranged according to the acquisition window order to form a communication latency candidate sequence and a packet loss candidate sequence, respectively, and correspond one-to-one with the communication latency distribution change vector and packet loss ratio vector in the abnormal feature fingerprint. The Euclidean distance method is used to calculate the difference between the communication latency candidate sequence and the communication latency distribution change vector, and the difference between the packet loss candidate sequence and the packet loss ratio vector. When the difference value is small, it indicates that the second-type candidate reproducible scenario and the abnormal feature fingerprint have a high degree of consistency in the corresponding dimension.

[0109] Compare the differences between all candidate reproducible scenarios of the first category and candidate reproducible scenarios of the second category, and select the candidate reproducible scenario with the smallest difference as the minimum reproducible scenario.

[0110] S5. Based on the minimum reproducibility scenario, determine whether the abnormal feature fingerprint corresponds to the real abnormality, record the source of the real abnormality, and generate test conclusions.

[0111] Furthermore, a unified analysis is performed on communication latency data, access latency data, mouth-to-ear latency data, packet loss data, and queuing depth data in the minimum reproducibility scenario.

[0112] When the minimum reproducible scenario belongs to the first type of candidate reproducible scenario, the access latency data in the minimum reproducible scenario is compared with the access latency elastic threshold, the mouth-to-ear latency data in the minimum reproducible scenario is compared with the mouth-to-ear latency elastic threshold, and the queuing depth data in the minimum reproducible scenario is compared with the queuing depth elastic threshold. When the access latency data, mouth-to-ear latency data, and queuing depth data are all within the allowable range of the corresponding elastic threshold, it is determined that the abnormal feature fingerprint can be reproduced by the minimum reproducible scenario in the access latency dimension, mouth-to-ear latency dimension, and queuing depth dimension.

[0113] When the minimum reproducible scenario belongs to the second type of candidate reproducible scenario, the correspondence between the communication latency data in the minimum reproducible scenario and the communication latency elastic threshold is compared, and the correspondence between the packet loss data in the minimum reproducible scenario and the packet loss elastic threshold is compared. When both the communication latency data and the packet loss data are within the allowable range of the corresponding elastic threshold, it is determined that the abnormal feature fingerprint can be reproduced by the minimum reproducible scenario in the dimensions of communication latency and packet loss.

[0114] Furthermore, after completing the comparison and judgment of the first type of candidate reproducible scenarios and the abnormal feature fingerprints, and the comparison and judgment of the second type of candidate reproducible scenarios and the abnormal feature fingerprints, the reproducible result corresponding to the minimum reproducible scenario is output as the test conclusion.

[0115] When the minimum reproducible scenario matches the anomaly feature fingerprint across all its dimensions, it is determined that a real anomaly exists in the broadband trunking communication network within the current acquisition window, and the source of the anomaly is recorded as the operating parameter setting method corresponding to the minimum reproducible scenario.

[0116] When the minimum reproducible scenario cannot be consistent with the anomalous feature fingerprint in some dimensions, it is determined that the anomalous feature fingerprint may be affected by random fluctuations and does not constitute a real anomaly.

[0117] This embodiment also provides a computer device applicable to an automated testing method for a broadband trunking communication system, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the automated testing method for the broadband trunking communication system proposed in the above embodiment.

[0118] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0119] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the automated testing method for a broadband trunking communication system as proposed in the above embodiments. The storage medium 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 Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0120] In summary, this invention achieves dynamic identification of multidimensional communication data boundaries by applying controllable disturbances and establishing disturbance response curves during business operations, thereby forming elastic thresholds to adapt to real-time changes in complex scenarios; by applying the elastic thresholds to the configuration of operating parameters and generating reproducible scenarios, it achieves a precise correspondence between abnormal distributions and operating conditions, enabling the source of anomalies to be clearly located, and ultimately achieving accurate diagnosis and reliable conclusion output of the operating status of broadband trunking communication networks.

[0121] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for automated testing of a wideband trunked communication system, the method comprising: Comprising, On the access side of the broadband trunking communication network, collecting voice group call service data, emergency video service data and location information service data through an edge computing node, and generating a collection window data set under a unified time reference; Jointly analyzing time delay features, packet loss features and continuity features in the collection window data set, extracting abnormal distribution features and generating an abnormal feature fingerprint; After the abnormal feature fingerprint is generated, a controllable disturbance is applied in the service running process to construct a disturbance response curve and identify an abnormal change point, and an elasticity threshold is generated; Taking the elasticity threshold as a reproduction condition, configuring running parameters within the elasticity threshold constraint range to generate a candidate reproduction scene, comparing the candidate reproduction scene with the abnormal feature fingerprint, and selecting the candidate reproduction scene with the smallest difference as the minimum reproduction scene; Based on the minimum reproduction scene, it is determined whether the abnormal feature fingerprint corresponds to a real abnormality, and the source of the real abnormality is recorded to generate a test conclusion; The joint analysis of the time delay features, packet loss features and continuity features in the collection window data set includes sorting the communication time delay data, access time delay data and talk-to-ear time delay data according to the numerical value and extracting the quantile points; Linear interpolation is used to connect adjacent quantile points one by one to generate communication time delay quantile number curves, access time delay quantile number curves and talk-to-ear time delay quantile number curves; When the communication time delay quantile number curves, access time delay quantile number curves and talk-to-ear time delay quantile number curves of adjacent collection windows appear overall translation and morphological changes, it is determined that the time delay features of the collection window data set have changed; The missing number of packet loss and the time period of packet loss occurrence are extracted from the packet loss condition data, and are arranged in order according to the time period of packet loss occurrence to form a packet loss cluster length sequence; The time period of each packet loss occurrence in the packet loss cluster length sequence is compared with the time period of number discontinuity in the continuity check result of the location information reporting sequence one by one to generate a confirmed packet loss cluster length sequence; When the confirmed packet loss cluster length sequence appears multiple packet loss missing numbers exceeding the size of a single normal packet loss, it is determined that the packet loss features in the collection window data set have changed; The local maximum values of the queue depth data are calculated point by point, and all local maximum values are marked as spikes, and the time difference between adjacent spikes is recorded to form an adjacent spike interval sequence; When the spike interval sequence continuously appears gradually shortened, it is determined that the continuity features in the collection window data set have changed; When the time delay features, packet loss features and continuity features change simultaneously between consecutive collection windows, it is determined that the collection window data set has abnormal distribution in multiple dimensions.

2. The method of automated testing of a broadband trunking communication system of claim 1, wherein: The collection of voice group call service data, emergency video service data and location information service data includes triggering voice group call service, emergency video service and location information service at the start time of the collection window through the edge computing node on the access side of the broadband trunking communication network; The access delay data, the talk-to-listen delay data and the queue depth data are collected in the voice group call service, the communication delay data and the packet loss condition data are collected in the emergency video service, and the position information reporting sequence continuity checking result is collected in the position information service.

3. The method of automatically testing a broadband trunking communication system of claim 2, wherein: The generating of the collection window data set comprises aligning and arranging the voice group call service data, the emergency video service data and the position information service data in the order of the collection window start time to the collection window end time under the unified time reference to form the collection window data set.

4. The method of automated testing of a broadband trunking communication system of claim 3, wherein: The extracting of the abnormal distribution features and the generating of the abnormal feature fingerprint comprises arranging the numerical differences between the communication delay quantile curves, the access delay quantile curves and the talk-to-listen delay quantile curves of adjacent collection windows in the time sequence of the collection windows to form the communication delay distribution change vector, the access delay distribution change vector and the talk-to-listen delay distribution change vector; The missing number of each packet loss in the confirmed packet loss cluster length sequence is arranged in the time sequence of the packet loss occurrence to form the packet loss rate vector; The peak interval of the queue depth data is arranged in the detection sequence to form the queue rhythm vector; The combined feature vector is spliced in the order of the communication delay distribution change vector, the access delay distribution change vector, the talk-to-listen delay distribution change vector, the packet loss rate vector and the queue rhythm vector, and the combined feature vector is encoded by using the one-way hash algorithm to generate the abnormal feature fingerprint.

5. The method of automatically testing a broadband trunking communication system of claim 4, wherein: The generating of the elastic threshold comprises recording the voice group call service data and the emergency video service data of the collection window data set under each perturbation intensity by changing the talk burst request times and the video data sending continuity during the running of the voice group call service and the emergency video service to form the corresponding perturbation response pairing sequence; The perturbation response curves are generated by curve fitting of each perturbation response pairing sequence, and the preliminary identification point set is generated based on the inflection points identified by the slope changes of the perturbation response curves; The final identification points are obtained by comparing the preliminary identification point set with the abnormal feature fingerprint; The perturbation intensity with the smallest value and the same time period as the abnormal feature fingerprint is selected as the communication delay elastic threshold, the access delay elastic threshold, the talk-to-listen delay elastic threshold, the packet loss condition elastic threshold and the queue depth elastic threshold by sorting the perturbation intensity in the final identification point according to the numerical value.

6. The method of automatically testing a broadband trunking communication system of claim 5, wherein: The minimum reproduction scene comprises generating the first type of candidate reproduction scene by configuring the talk burst request times to constrain within the access delay elastic threshold, the talk-to-listen delay elastic threshold and the queue depth elastic threshold in the voice group call service; The second type of candidate reproduction scene is generated by configuring the video data sending stability to constrain within the communication delay elastic threshold and the packet loss condition elastic threshold in the emergency video service; The candidate reproduction scene set is formed by collecting all the first type of candidate reproduction scene and the second type of candidate reproduction scene, and the candidate reproduction scene set is compared with the abnormal feature fingerprint one by one. The Euclidean distance method is used to calculate the difference degree between the access delay data, the mouth-to-ear delay data and the queue depth data in the first type of candidate complex scene and the abnormal feature fingerprint, and the difference degree between the communication delay data and the packet loss condition data in the second type of candidate complex scene and the abnormal feature fingerprint, and the candidate complex scene with the minimum difference degree is selected as the minimum complex scene.

7. The method of automated testing of a broadband trunking communication system of claim 6, wherein: The method further comprises: comparing the communication delay data, the access delay data, the mouth-to-ear delay data, the packet loss condition data and the queue depth data in the minimum complex scene with corresponding elastic threshold constraint conditions respectively, and determining that the abnormal feature fingerprint corresponds to a real abnormality when all the comparison results are within the constraint range.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that: The processor executes the computer program to realize the steps of the automatic test method of the wideband trunking communication system according to any one of claims 1-7.

9. A computer readable storage medium having stored thereon a computer program, characterized in that: The computer program is executed by the processor to realize the steps of the automatic test method of the wideband trunking communication system according to any one of claims 1-7.

Images