Service bearing capacity evaluation method

Abstract

The invention belongs to the technical field of communication, and particularly relates to a service bearing capacity evaluation method, which comprises the following steps: acquiring a data set formed by each grid at each time point; performing grid processing on RSRP, RSRQ, SINR and MOS values in a first original data set based on the VORR voice environment according to a preset time period to obtain a first data set, performing OCNS scrambling on base stations in a target area, and obtaining a first utilization rate index based on timestamp information; performing fitting through a first preset fitting model based on the first data set and the first utilization rate index to obtain a corresponding expected MOS value; similarly, respectively obtaining expected downlink average throughput rate and uplink average throughput rate of the RLC layer in the video downloading environment and the video uploading environment based on similar data processing processes; based on the expected MOS value, the expected downlink average throughput rate and the uplink average throughput rate of the RLC layer, whether the satisfaction degree in the target area meets the requirement or not is judged; and if the requirements are not met, checking by taking the grid as a unit to calculate the service satisfaction of the terminal in different environments.

Description

Technical Field

[0001] The present invention relates to the field of communications, and in particular to a method for evaluating service carrying capacity. Background Art

[0002] In existing technologies, VoNR voice service is one of the important services of operators. Providing users with stable and high-quality voice services is an important part of improving 5G network perception and service experience. To improve user experience, shorten call setup latency, and enhance the clarity of real-time voice communications, VoNR (Voice over New Radio), based on 5G and IMS (IP Multimedia Subsystem), has been introduced in the 5G era as a target voice solution for operators. For operators, deploying VoNR marks the beginning of the evolution towards mobile broadband voice, improving wireless spectrum utilization and reducing network costs. VoNR leverages the advantages of 5G's large bandwidth, high spectrum utilization, and robust fading resistance of new air interface / antenna technologies. By using the ultra-high-definition EVS (Enhanced Voice Services) coding scheme, it provides users with shorter voice call access latency and an ultra-high-definition voice experience. However, VoNR is carried entirely by 5G networks, and voice quality is strongly dependent on network coverage and antenna transceiver performance. Commercial terminals do not report voice service perception metrics, making accurate evaluation of VoNR service quality and user perception critical to ensuring IMS voice service perception on 4G / 5G networks.

[0003] Since 4G VoLTE, Mean Opinion Score (MOS) has become the mainstream method for operations and maintenance personnel to evaluate voice quality.

[0004] End-to-end perception of voice services usually uses the voice MOS (Mean Opinion Score) scoring method, which is divided into subjective and objective evaluations.

[0005] Subjective evaluation is accomplished by manually answering and perceiving voice quality. This method is labor-intensive and time-consuming, and the scoring is somewhat arbitrary. Therefore, this specification recommends an objective evaluation method.

[0006] Objective evaluation can be performed in two ways: active and passive. The active method typically uses a terminal and test instrument to initiate a test service on the terminal side, and performs evaluation based on a comparison of the original voice signal and the distorted voice signal. This type of objective evaluation often quantifies voice quality using numerical distances or auditory models that describe how the auditory system perceives quality. Key scoring algorithms include PSQM / PSQM+, PAMS, PESQ, and POLQA. The passive method directly evaluates voice quality by converting factors such as the delay, jitter, and packet loss of the actual voice service output signal into a MOS score. The primary scoring algorithm is the E-Model. This method is primarily accomplished by collecting metrics from real-time voice services occurring on the network.

[0007] In most cases, the method for obtaining voice MOS values ​​still relies on DT / CQT test acquisition on the existing network and calculation and filling of core network DPI data.

[0008] DT / CQT testing, for example, requires significant time and manpower to collect data and is difficult to meet the requirements of real-time network evaluation. This approach cannot be implemented around the clock, at all times, and in all regions. It requires significant manpower and resources for long-term data collection, which is not only costly but also susceptible to human error and incidental events that can affect the accuracy of VoLTE MOS, making the evaluation results less comprehensive, accurate, and objective.

[0009] MOS values ​​are calculated and filled using core network DPI data: the packet loss, jitter, and delay of user-side RTP packets are fitted and filled. This requires deploying probes at the core network interface to capture real-time reporting data. The data volume is large and inaccurate (there is a risk of data loss during periods of high load), and there is a lack of accurate cell, frequency band, and terminal information.

[0010] Furthermore, the MOS values ​​of voice services vary under different network loads. Similar problems also exist for video uplink and video download. Summary of the Invention

[0011] This application is proposed based on the above-mentioned needs of the existing technology. The technical problem to be solved by this application is to provide a method for evaluating business carrying capacity to facilitate the evaluation of business carrying capacity, accurately discover and explore network problems, and guide the resolution of network and terminal problems.

[0012] In order to solve the above problems, the technical solutions provided in this application include:

[0013] A method for evaluating service carrying capacity is provided, comprising: dividing a geographic layer of a target area into grids, obtaining a data set formed by each grid at each time point; obtaining a first original data set based on a VoNR voice environment in the data set, and performing grid processing on RSRP, RSRQ, SINR, and MOS values ​​therein according to a preset time period to obtain a first data set, performing OCNS scrambling on base stations in the target area, and obtaining a first utilization index by association based on timestamp information; fitting the first data set and the first utilization index through a first preset fitting model to obtain a corresponding expected MOS value; obtaining a second original data set based on a video download environment in the data set, performing OCNS scrambling on each base station in the target area, and obtaining a second utilization index by association based on timestamp information; and obtaining a second utilization index by association based on RSRP, RSRQ, SINR, and RLC layer downlink throughput THP in the second original data set. DL The second utilization index is fitted with the second preset fitting model to obtain the corresponding expected RLC layer downlink average throughput; a third original data set based on the video upload environment in the data set is obtained; OCNS scrambles each base station in the target area, and based on the timestamp information, associates and obtains the third utilization index; based on the RSRP, RSRQ, SINR, RLC layer uplink throughput THP in the third original data set UL The third utilization index is fitted through a third preset fitting model to obtain the corresponding expected RLC layer uplink average throughput; based on the expected MOS value, the expected RLC layer downlink average throughput and the expected RLC layer uplink average throughput, it is judged whether the satisfaction in the target area meets the demand; if the satisfaction in the target area does not meet the demand, it is checked in grid units, the number of sampling points of each grid at different times is selected according to different environments, and the service satisfaction of the terminal in different environments is calculated.

[0014] Preferably, the method further comprises obtaining the service satisfaction probability in each grid at a certain time point based on the service satisfaction of different terminals in combination with the terminal penetration rate.

[0015] Preferably, the RSRP, RSRQ, SINR and MOS values ​​in the first original data set are grid-processed according to a preset time period to obtain the first data set, including processing all sampling points in the data set according to a time grid granularity of 1s, and aggregating each RSRP, RSRQ, SINR and MOS value to 1 second; backfilling the MOS value upward until the previous MOS value, thereby obtaining the first data set containing RSRP, RSRQ, SINR and MOS values ​​arranged in order of seconds.

[0016] Preferably, the first preset fitting model is used to fit the first data set and the first utilization index to obtain the expected MOS value under different wireless environments, which is expressed as: MOS = f(RSRP, RSRQ, SINR, PRB%) DL ,PRB% UL )+c, where c is a constant; the preset fitting methods include but are not limited to: multivariate nonlinear regression fitting, LSTM-based time series prediction or artificial neural network.

[0017] Preferably, the accuracy of the prediction model is calculated by at least one of MSE, RMSE, MAE, MAPE and R2.

[0018] Preferably, the method further comprises calculating the RLC layer downlink throughput rate according to different video types, cache durations, video compression ratios, and video codec standards to obtain a minimum downlink guaranteed rate required for continuous video playback.

[0019] Preferably, the RSRP, RSRQ, SINR, RLC layer downlink throughput THP in the second original data set DL The second utilization index is fitted by the second preset fitting model to obtain the expected average downlink throughput of the RLC layer under different wireless environments, which is expressed as: THP DL =f(RSRP,RSRQ,SINR,PRB% DL )+c, where c is a constant; the preset fitting methods include but are not limited to: multivariate nonlinear regression fitting, LSTM-based time series prediction or artificial neural network.

[0020] Preferably, the RSRP, RSRQ, SINR, RLC layer uplink throughput THP in the third original data set DL The third utilization index is fitted by the third preset fitting model to obtain the expected average uplink throughput of the RLC layer under different wireless environments, which is expressed as: THP UL =f(RSRP,RSRQ,SINR,PRB% UL )+c, where c is a constant; the preset fitting methods include but are not limited to: multivariate nonlinear regression fitting, LSTM-based time series prediction or artificial neural network.

[0021] Preferably, the method of screening the number of sampling points of each grid at different times according to different environments and calculating the service satisfaction of the terminal in different environments includes: expressing the service satisfaction of the VoNR voice environment as: The service satisfaction of the video download environment is expressed as: The business satisfaction of the video upload environment is expressed as: Among them, UE A For a terminal, T i For a moment, N Ti，qualified is the number of sampling points screened out in the corresponding environment, N i is the total number of sampling points in the grid. The screening conditions include: for the VoNR voice environment, its expected MOS value is greater than the MOS threshold; for the video download environment, its expected RLC layer downlink average throughput is greater than the RLC layer downlink average throughput threshold and cannot be lower than the minimum downlink guaranteed rate; for the video upload environment, its expected RLC layer uplink average throughput is greater than the RLC layer uplink average throughput threshold and cannot be lower than the minimum uplink guaranteed rate.

[0022] Preferably, the service satisfaction probability of a corresponding specific environment in a grid at a moment is obtained based on the service satisfaction of different terminals in combination with the terminal penetration rate, wherein the service satisfaction probability for the VoNR voice environment is expressed as:

[0023] Among them, UE A UE B and UE RedCap Indicates different terminals.

[0024] Compared with the existing technology, this application can realize the dynamic evaluation of the network's ability to carry services and whether it meets the planning requirements for different services, different terminals, different network standards and frequency bands, different carrier bandwidths, and different equipment types according to the actual network planning requirements through the data in the OTT data. Furthermore, for areas that do not meet the business requirements, it is possible to further locate problem grids, problem cells, problem terminal groups, etc., accurately discover and explore network problems, and guide the resolution of network and terminal problems. Furthermore, considering that this patent can be evaluated through grids, it also uses multi-frequency network construction scenarios to accurately discover frequency bands and carriers in regions and grids that do not meet the problems, better guide network construction and network optimization work, and maximize the utilization efficiency of the network. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of this specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the embodiments of this specification. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0026] Figure 1 This is a flowchart of the steps of a service carrying capacity evaluation method in an embodiment of the present invention. DETAILED DESCRIPTION

[0027] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0028] In the description of the embodiments of this application, it should be noted that, unless otherwise expressly specified or limited, the term "connected" should be understood in a broad sense. For example, it can mean a fixed connection, a detachable connection, or an integral connection. It can be a mechanical connection, an electrical connection, a direct connection, or an indirect connection through an intermediate medium. Those skilled in the art will understand the specific meanings of the above terms in this application based on specific circumstances.

[0029] The terms "top," "bottom," "above," "below," and "on" used throughout the description refer to relative positions of components of a device, such as the relative positions of top and bottom substrates within a device. It will be understood that devices are multifunctional regardless of their orientation in space.

[0030] To facilitate understanding of the embodiments of the present application, further explanation will be given below with reference to specific embodiments in conjunction with the accompanying drawings. The embodiments do not constitute a limitation on the embodiments of the present application.

[0031] This embodiment provides a method for evaluating service carrying capacity. Figure 1 shown.

[0032] The service carrying capacity evaluation method includes:

[0033] The geographic layer of the target area is divided into grids of preset specifications, and the OTT data in each grid within a preset time period are obtained respectively. The grids without data and the grids with data amounts less than the preset amount are eliminated.

[0034] Furthermore, the grid is divided according to the specification frame required by the network operation. For example, the size of the grid is 20m×20m. The preset time period is a time period. For example, 7 days is a period, and the OTT data within 7 days is divided into the above-mentioned grids, that is, the data is divided into each geographic grid according to the latitude and longitude information. All empty grids or low-sampling grids (such as grids with 0 or less than 100 sampling points for 7 consecutive days) are eliminated from all grids contained in the target area, and the total number of grids is recorded as N.

[0035] Among them, OTT is the abbreviation of Over The Top. With the large-scale construction of 4 / 5G wireless networks and the extensive use of smart phones, it has been extended to Internet application services, that is, providing various online services to users through the Internet, such as WeChat, Alipay, Baidu and other applications. This type of application service is generally provided by a third party outside the operator. These service providers directly face users and achieve the effect of crossing the operator network. The data generated in these application services is OTT data.

[0036] OTT data from smart terminals is collected by the collector through SDKs integrated into various APPs of its partners. Under the premise of complying with the terminal's requirements for user privacy permissions, it collects and returns the time of occurrence, user current location and network measurement information through various preset mechanisms (trigger type, periodic type, task type, etc.).

[0037] OTT data collection includes the following processes: installing an APP that supports OTT data collection and authorizing data collection; during the APP application process, triggering data collection and storing it in the smart terminal; after the collected data meets the predetermined requirements, triggering the data to be transmitted back to the back-end collection server through the smart terminal, such as when the size of the collected data file reaches the predetermined threshold to trigger transmission or when the collection time reaches the predetermined threshold to trigger transmission; the back-end data collection server completes the collection and collation of the data.

[0038] For wireless network analysis, the following fields are important to focus on in Internet OTT:

[0039] a. Network signal information

[0040] 1) 5G network information collection

[0041]

[0042]

[0043]

[0044] 2) 4G network information collection

[0045]

[0046]

[0047] 3) 3G network information collection

[0048]

[0049]

[0050] 4) 2G network information collection

[0051]

[0052]

[0053] b. Terminal and user related information

[0054]

[0055]

[0056] c. Location information

[0057]

[0058] The original data set formed by the remaining grids at each time point is obtained, where the original data set includes RSRP, RSRQ, SINR, primary serving cell information, and downlink PRB utilization and uplink PRB utilization corresponding to the primary serving cell information.

[0059] Among them, RSRP (Reference Signal Received Power) is the reference signal received power, which measures the strength of the base station reference signal received by the terminal device and is used to evaluate the network coverage quality; RSRQ (Reference Signal Received Quality) is the reference signal received quality, which combines signal strength and interference level to comprehensively evaluate channel quality; SINR (Signal-to-Interference-plus-Noise Ratio) is the signal to interference plus noise ratio, which reflects the signal purity and measures the ratio of useful signal to interference and noise.

[0060] A first original data set based on a VoNR voice environment is obtained from the data set according to a time point, where the data set includes RSRP, RSRQ, SINR, and MOS values.

[0061] The first raw data set includes a large amount of road test data including timestamps, RSRP, RSRQ, SINR, and MOS values ​​collected from different terminals in a well-optimized 5G network with continuous coverage of a single type of device in different frequency bands, different bandwidths, and different device types. The data is collected by voice MOS boxes. The data includes: RSRP, RSRQ, SINR, and MOS values. The MOS value (Mean Opinion Score) is the mean opinion score, which subjectively or objectively evaluates voice / video quality, ranging from 1 to 5 points (5 points is the best). RSRP, RSRQ, and SINR are used for network optimization, and the MOS value is used for user experience evaluation.

[0062] For example, different frequency bands include 900M NR / 2.1G NR / 3.5G NR, etc.; different bandwidths include 900M NR5M / 10M, 2.1G NR 20M / 40M, 3.5G NR 100M, etc.; different device types include 900M NR 4T4R, 2.1G 4T4R / 2T2R, 3.5G NR 64TR, etc.

[0063] The RSRP, RSRQ, SINR, and MOS values ​​in the first original data set are subjected to grid processing according to a preset time period to obtain a first data set.

[0064] The MOS value is generated every 8-9 seconds, while the RSRP, RSRQ, and SINR values ​​are generated at different intervals depending on the capabilities of the chip and MOS device, generally 200ms or less. This results in the MOS value not being completely synchronized with the RSRP, RSRQ, and SINR values. Therefore, it is necessary to perform time-based grid processing on the RSRP, RSRQ, SINR, and MOS values ​​in the first original data set. First, all sampling points in the data set are processed according to a 1s time grid granularity, and each RSRP, RSRQ, SINR, and MOS value is aggregated to the nearest full second. Second, since the MOS value is generated every 8-9 seconds, a large number of null values ​​will be generated in the data table based on a 1-second grid. The generation of the MOS value is closely related to the RSRP, RSRQ, and SINR values ​​of the previous 8-9 seconds. In this case, the MOS value can be backfilled upwards to the previous MOS value, thereby obtaining a first data set containing RSRP, RSRQ, SINR, and MOS values ​​arranged in second order. For example, the first data set is shown in the following table.

[0065]

[0066] OCNS scrambling is performed on base stations in the target area, and based on timestamp information, a first utilization index of the calling terminal at each moment under different loads is obtained through association.

[0067] The first utilization index includes the first uplink utilization index PRB% of the 5G cell occupied by the calling terminal at each moment UL And the 5G cell second downlink utilization index PRB% occupied by the called terminal at each moment DL .

[0068] OCNS (Orthogonal Channel Noise Source) scrambling simulates interference in a multi-user environment by generating pseudo-random noise orthogonal to real user signals, for system testing and performance evaluation. For example, OCNS scrambling is performed at 10%, 30%, 50%, and 70%. PRBs (Physical Resource Blocks) are physical resource blocks in wireless communication systems, used to transmit wireless signals and data. PRB utilization refers to the ratio of the number of PRBs actually used to the number of available PRBs within a given timeframe.

[0069] Based on the first data set and the first utilization index, fitting is performed through a first preset fitting model to obtain expected MOS values ​​under different wireless environments.

[0070] The first preset fitting model is a model formed after pre-training using historical data.

[0071] Methods for training using historical data include, but are not limited to, multivariate nonlinear regression fitting, LSTM-based time series prediction, or artificial neural network training algorithms such as RNN and DNN to obtain the following expression.

[0072] MOS=f(RSRP,RSRQ,SINR,PRB% DL ,PRB% UL )+c

[0073] Where c is a constant. Whenever a set of RSRP, RSRQ, and SINR values ​​is given, a unique MOS value can be obtained.

[0074] Furthermore, the accuracy of the prediction model is calculated by regression evaluation indicators such as MSE (Mean Squared Error), RMSE (Root Mean Squared Error), MAE (Mean Absolute Error), MAPE (Mean Absolute Percentage Error), and R2 (Coefficient of Determination). When the accuracy of the model meets the expected requirements, it can be considered that the MOS prediction model can be based on RSRP, RSRQ, SINR, PRB% UL 、PRB% DL The corresponding MOS value is obtained from the measurement value, that is, the expected MOS value in different wireless environments.

[0075] The VoNR voice MOS prediction models are different for different frequency bands, different bandwidths, different device types, different typical terminals, and different cell power configurations. Therefore, it is necessary to establish different fitting models based on different situations and input the acquired data into the matching fitting model for prediction according to the actual situation.

[0076] A second original data set based on a video download environment is obtained from the data set according to a time point, wherein the data set includes RSRP, RSRQ, SINR and RLC layer downlink throughput THP DL .

[0077] Specifically, the second original data set includes a large amount of timestamps, RSRP, RSRQ, SINR and RLC layer downlink throughput THP of different terminals collected by FTP download in a 5G network with continuous coverage of a well-optimized single type of device in different frequency bands (such as 900M NR / 2.1G NR / 3.5G NR, etc.), different bandwidths (exemplarily, 900M NR 5M / 10M, 2.1G NR 20M / 40M, 3.5G NR 100M, etc.), and different device types (for example, 900M NR 4T4R, 2.1G 4T4R / 2T2R, 3.5G NR 64TR, etc.). DL Drive test data.

[0078] The RLC layer downlink throughput is calculated based on different video types (such as 480p, 720p, 1080p, 2K, and 4K sources), cache duration (typically 4 seconds, which can be modified), video compression ratio (typically 1 / 80, which can be modified), and video codec standard (typically H.264 / H.265, which can be modified; H.265 is used here) to obtain the minimum downlink guaranteed rate required for continuous video playback.

[0079] OCNS scrambling is performed on each base station in the target area (scrambling 10%, 30%, 50%, 70%), and a second utilization index is obtained by association based on the timestamp information.

[0080] Specifically, the second utilization index is the utilization index PRB% of the 5G cell occupied by the terminal at each moment under different loads DL .

[0081] Based on the RSRP, RSRQ, SINR and the second utilization index in the second original data set, fitting is performed through a preset fitting method to obtain the expected RLC layer downlink average throughput under different wireless environments.

[0082] The preset fitting methods include but are not limited to: a multivariate nonlinear regression fitting method, a time series prediction method based on LSTM, or a method based on an artificial neural network training algorithm, such as RNN, DNN, etc., to obtain the following expression.

[0083] THP DL =f(RSRP,RSRQ,SINR,PRB% DL )+c

[0084] Where c is a constant. Whenever a set of RSRP, RSRQ, and SINR values ​​is given, a unique THP can be obtained. D value.

[0085] Furthermore, the accuracy of the prediction model is calculated by regression evaluation indicators such as MSE (Mean Squared Error), RMSE (Root Mean Squared Error), MAE (Mean Absolute Error), MAPE (Mean Absolute Percentage Error), and R2 (Coefficient of Determination). When the accuracy of the model meets the expected requirements, it can be considered that the MOS prediction model can be based on RSRP, RSRQ, SINR, PRB% UL 、PRB% DL The measurement value of the corresponding RLC layer downlink average throughput is obtained, that is, the expected THP in different wireless environments DL value.

[0086] A third original data set based on a video upload environment is obtained from the data set according to a time point, wherein the data set includes RSRP, RSRQ, SINR and RLC layer uplink throughput THP UL .

[0087] Specifically, the second original data set includes a large amount of timestamps, RSRP, RSRQ, SINR and RLC layer downlink throughput THP of different terminals collected by FTP download in a 5G network with continuous coverage of a well-optimized single type of device in different frequency bands (such as 900M NR / 2.1G NR / 3.5G NR, etc.), different bandwidths (exemplarily, 900M NR 5M / 10M, 2.1G NR 20M / 40M, 3.5G NR 100M, etc.), and different device types (for example, 900M NR 4T4R, 2.1G 4T4R / 2T2R, 3.5G NR 64TR, etc.). DL Drive test data.

[0088] OCNS scrambling is performed on each base station in the target area (scrambling 10%, 30%, 50%, 70%), and the third utilization index is obtained by association based on the timestamp information.

[0089] Specifically, the third utilization index is the utilization index PRB% of the 5G cell occupied by the terminal at each moment under different loads DL .

[0090] Based on the RSRP, RSRQ, SINR and the third utilization index in the third original data set, fitting is performed through a third preset fitting model to obtain the expected RLC layer uplink average throughput under different wireless environments.

[0091] The third preset fitting model is a model formed after pre-training using historical data.

[0092] Methods for training using historical data include, but are not limited to, multivariate nonlinear regression fitting, LSTM-based time series prediction, or artificial neural network training algorithms such as RNN and DNN to obtain the following expression.

[0093] THP UL =f(RSRP,RSRQ,SINR,PRB% UL )+c

[0094] Where c is a constant. Whenever a set of RSRP, RSRQ, and SINR values ​​is given, a unique THP can be obtained. UL value.

[0095] Furthermore, the accuracy of the prediction model is calculated by regression evaluation indicators such as MSE (Mean Squared Error), RMSE (Root Mean Squared Error), MAE (Mean Absolute Error), MAPE (Mean Absolute Percentage Error), and R2 (Coefficient of Determination). When the accuracy of the model meets the expected requirements, it can be considered that the MOS prediction model can be based on RSRP, RSRQ, SINR, PRB% UL 、PRB% DL The measurement value of the corresponding RLC layer uplink average throughput is obtained, that is, the expected THP in different wireless environments UL value.

[0096] Based on the expected MOS value, the expected average downlink throughput of the RLC layer, and the expected average uplink throughput of the RLC layer, it is determined whether the satisfaction in the target area meets the requirements.

[0097] Specifically,

[0098] When the target area is in an urban area:

[0099] For the sampling points in the VoNR voice environment, if the ratio of the number of sampling points with expected MOS values ​​greater than the MOS threshold to the total number of sampling points in the VoNR voice environment is greater than or equal to 92%, then the satisfaction of the area is considered to meet the requirements.

[0100] For the sampling points in the video download environment, if the ratio of the number of sampling points where the expected RLC layer downlink average throughput is greater than the RLC layer downlink average throughput threshold to the total number of sampling points in the video download environment is greater than or equal to 92%, then the satisfaction of the area is considered to meet the requirements.

[0101] For the sampling points in the video upload environment, if the ratio of the number of sampling points where the expected RLC layer uplink average throughput is greater than the RLC layer uplink average throughput threshold to the total number of sampling points in the video upload environment is greater than or equal to 92%, then the satisfaction of the area is considered to meet the requirements.

[0102] When the target area is in an urban area:

[0103] For the sampling points in the VoNR voice environment, if the ratio of the number of sampling points with expected MOS values ​​greater than the MOS threshold to the total number of sampling points in the VoNR voice environment is greater than or equal to 90%, the satisfaction of the area is considered to meet the requirements.

[0104] For the sampling points in the video download environment, if the ratio of the number of sampling points where the expected RLC layer downlink average throughput is greater than the RLC layer downlink average throughput threshold to the total number of sampling points in the video download environment is greater than or equal to 90%, then the satisfaction of the area is considered to meet the requirements.

[0105] For the sampling points in the video upload environment, if the ratio of the number of sampling points where the expected RLC layer uplink average throughput is greater than the RLC layer uplink average throughput threshold to the total number of sampling points in the video upload environment is greater than or equal to 90%, then the satisfaction of the area is considered to meet the requirements.

[0106] If the satisfaction of the target area does not meet the requirements, it is checked in grid units, and the number of sampling points in each grid at different times for different terminals is calculated to calculate the service satisfaction of the terminals in different environments.

[0107] Select the mainstream test terminals of the network as the typical terminals for network carrying service capability evaluation. Here, select several commercial version terminals with high market share and supporting the network frequency band of the operator as the typical terminals. A UE B At the same time, considering the widespread promotion of 5G RedCap terminals, mainstream 5G RedCap terminals can also be considered as one of the typical terminals when evaluating the network's business carrying capacity. The corresponding evaluation results can be applied to the network planning and post-network evaluation work in the target area where the RedCap function is enabled. RedCap .

[0108] Calculate the target area for UE A Business satisfaction, including, from T i N of each grid at each moment i The number of sampling points selected is N Ti，qualified The sampling point, in addition to satisfying that each indicator is greater than the corresponding threshold, also includes the RLC layer downlink throughput THP obtained in the video download environment DL It cannot be lower than the calculated minimum guaranteed downlink rate. Correspondingly, the RLC layer uplink throughput THP obtained in the video upload environment UL The rate cannot be lower than the calculated minimum guaranteed uplink rate.

[0109] In addition, specifically, the screening conditions include: for the VoNR voice environment, its expected MOS value is greater than the MOS threshold; for the video download environment, its expected RLC layer downlink average throughput is greater than the RLC layer downlink average throughput threshold and cannot be lower than the minimum downlink guaranteed rate; for the video upload environment, its expected RLC layer uplink average throughput is greater than the RLC layer uplink average throughput threshold and cannot be lower than the minimum uplink guaranteed rate.

[0110] Specifically, the service satisfaction of the VoNR voice environment is expressed as:

[0111]

[0112] The service satisfaction of the video download environment is expressed as:

[0113]

[0114] The business satisfaction of the video upload environment is expressed as:

[0115]

[0116] UE B and UE RedCap The business satisfaction can be calculated in the same way.

[0117] According to the service satisfaction of different terminals and the terminal penetration rate, the service satisfaction probability in each grid at a certain time point is obtained.

[0118] The service satisfaction probability within a grid at a certain moment is calculated using the terminal penetration rate that can be directly obtained and is expressed as:

[0119]

[0120] The terminal penetration rate can be calculated based on the capabilities included in the UE Capability reported by the terminal and the number of terminal connections.

[0121] Based on the service satisfaction probability calculated above, the service satisfaction of the grid in each period can be further calculated based on the system busy hour (the period with the highest downlink PRB utilization in the target network) or customized busy hour (for example, for areas with dense office buildings, it can be based on 9:00-11:00 a.m. and 3:00-5:00 p.m. on weekdays; for residential areas, it can be based on 7:00-9:00 p.m. on weekdays and 10:00-12:00 a.m. and 4:00-8:00 p.m. on weekends). The service satisfaction at each time can be calculated as the arithmetic mean.

[0122] Furthermore, according to the service satisfaction at each moment, the threshold Threshold1 can be customized based on the actual needs of network operation (for example, it can be set to 90%). If the grid meets this requirement during the period, the grid is recorded as the service carrying standard grid, and the number of all grids that meet this requirement in the area is recorded as N. 业务承载达标 The grid compliance rate of this area is N 业务承载达标 / N 总栅格 *100%.

[0123] This facilitates further statistics on substandard cells within the target area, allowing for further optimization and problem location for the top poor-quality cells based on service satisfaction, and guiding network optimization personnel to conduct detailed optimization for these cells. If network optimization still fails to meet the service satisfaction requirements within the cell, further network construction measures (such as adding sites, expanding carrier capacity, and carrier aggregation) can be used to improve network coverage and ensure good service perception within the target area.

[0124] This paper invented a method that combines the air interface data, terminal information, location information in OTT data with CM data such as the frequency band, bandwidth, and standard of the main service cell, and cell load KPI data. It can achieve different services (such as voice, uplink and downlink data services, uplink and downlink UDP services, etc.), different terminals (HiSilicon chip terminals, Qualcomm chip terminals, RedCap terminals, etc.), different network standards and frequency bands (2.6GNR / 3.5G NR TDD / 4.9G NR TDD / 2.1G NR FDD / 900M NR FDD / 800M NRFDD / millimeter wave, etc.), different carrier bandwidths (such as 3.5G NR 100M / 3.5G NR 200M / 2.1G NR 40M / 2.1G NR20M / 900M NR 10M / 900M NR 5M / 800M NR 10M / 800M NR 5M, etc.), different equipment types (64TR / 32TR / 8TR / 4T4R / 2T4R / 2T2R, etc.), dynamically evaluate the network's ability to carry services and whether it meets the planning requirements based on the actual network planning requirements. Furthermore, for areas that do not meet business requirements, we can further locate problem grids, problem cells, problem terminal groups, etc., accurately discover and explore network problems, and guide the resolution of network and terminal problems. Furthermore, considering that applications can be evaluated by grid, the multi-frequency network construction scenario can also be used to accurately discover frequency bands and carriers that do not meet the problems in the region and grid, better guide network construction and network optimization work, and maximize the utilization efficiency of the network.

[0125] This application is based on the OTT data fitting method, which can effectively evaluate the minimum coverage requirements that need to be provided by the network to ensure the basic perception / good perception of various services of mainstream terminals under different networks and different carriers, and then guide the network planning, equipment selection and other work in the area; after the multi-frequency network is built, based on this invention, it can be evaluated whether the network coverage in the target area can meet the basic perception / good perception network requirements of mainstream terminals for various services in each grid.

[0126] The specific implementation methods described above further illustrate the purpose, technical solutions and beneficial effects of this application. It should be understood that the above description is only the specific implementation methods of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of this application should be included in the scope of protection of this application.

Claims

1. A method for evaluating service carrying capacity, characterized in that: include: Divide the geographic layer of the target area into grids and obtain the dataset formed by each grid at each time point; Obtain a first original data set based on a VoNR voice environment from the data set, and perform grid processing on the RSRP, RSRQ, SINR, and MOS values ​​therein according to a preset time period to obtain a first data set, perform OCNS scrambling on the base stations in the target area, and obtain a first utilization indicator by association based on timestamp information; Based on the first data set and the first utilization index, a first preset fitting model is used to fit the first data set and the first utilization index to obtain a corresponding expected MOS value; Obtain the second original data set based on the video download environment in the data set, perform OCNS scrambling on each base station in the target area, and obtain the second utilization index based on the timestamp information; based on the RSRP, RSRQ, SINR, RLC layer downlink throughput THP in the second original data set DL The second utilization index is fitted with a second preset fitting model to obtain a corresponding expected RLC layer downlink average throughput; Obtain the third original data set based on the video upload environment in the data set; perform OCNS scrambling on each base station in the target area, and obtain the third utilization index based on the timestamp information; calculate the RSRP, RSRQ, SINR, and RLC layer uplink throughput THP based on the third original data set UL The third utilization index is fitted using a third preset fitting model to obtain a corresponding expected RLC layer uplink average throughput; Based on the expected MOS value, the expected average downlink throughput of the RLC layer, and the expected average uplink throughput of the RLC layer, determine whether the satisfaction level in the target area meets the requirements; If the satisfaction of the target area does not meet the requirements, it will be checked in grid units. The number of sampling points in each grid at different times will be selected according to different environments to calculate the service satisfaction of the terminal in different environments.

2. The method for evaluating service carrying capacity according to claim 1, wherein: The method further comprises obtaining the service satisfaction probability in each grid at a certain time point according to the service satisfaction of different terminals in combination with the terminal penetration rate.

3. The method for evaluating service carrying capacity according to claim 1, wherein: The RSRP, RSRQ, SINR and MOS values ​​in the first original data set are subjected to grid processing according to a preset time period to obtain a first data set including: All sampling points in the data set are processed according to the time grid granularity of 1 second, and each RSRP, RSRQ, SINR and MOS value is aggregated to 1 second; The MOS values ​​are backfilled upward until the previous MOS value, thereby obtaining a first data set including RSRP, RSRQ, SINR, and MOS values ​​arranged in order of seconds.

4. The method for evaluating service carrying capacity according to claim 1, wherein: The first preset fitting model is used to fit the first data set and the first utilization index to obtain expected MOS values ​​under different wireless environments, which are expressed as: MOS=f(RSRP,RSRQ,SINR,PRB% DL ,PRB% UL )+c Where c is a constant; Preset fitting methods include but are not limited to: multivariate nonlinear regression fitting, LSTM-based time series prediction, or artificial neural network.

5. The method for evaluating service carrying capacity according to claim 4, wherein: The accuracy of the prediction model is calculated by at least one of MSE, RMSE, MAE, MAPE and R2.

6. The method for evaluating service carrying capacity according to claim 1, wherein: The method also includes calculating the RLC layer downlink throughput rate according to different video types, cache durations, video compression ratios, and video coding and decoding standards to obtain the minimum downlink guaranteed rate required for continuous video playback.

7. The method for evaluating service carrying capacity according to claim 1, wherein: The RSRP, RSRQ, SINR, and RLC layer downlink throughput THP in the second original data set DL The second utilization index is fitted by a second preset fitting model to obtain the expected average downlink throughput of the RLC layer under different wireless environments, which is expressed as: THP DL =f(RSRP,RSRQ,SINR,PRB% DL )+c Where c is a constant; Preset fitting methods include but are not limited to: multivariate nonlinear regression fitting, LSTM-based time series prediction, or artificial neural network.

8. The method for evaluating service carrying capacity according to claim 1, wherein: The RSRP, RSRQ, SINR, and RLC layer uplink throughput THP in the third original data set DL The third utilization index is fitted by a third preset fitting model to obtain the expected average uplink throughput of the RLC layer under different wireless environments, which is expressed as: THP UL =f(RSRP,RSRQ,SINR,PRB% UL )+c Where c is a constant; Preset fitting methods include but are not limited to: multivariate nonlinear regression fitting, LSTM-based time series prediction, or artificial neural network.

9. The method for evaluating service carrying capacity according to claim 1, wherein: The method of screening the number of sampling points of each grid at different times according to different environments and calculating the service satisfaction of the terminal in different environments includes: The service satisfaction of the VoNR voice environment is expressed as: The service satisfaction of the video download environment is expressed as: The business satisfaction of the video upload environment is expressed as: Among them, UE A For a terminal, T i For a moment, N Ti，qualified is the number of sampling points screened out in the corresponding environment, N i The total number of sampling points in the grid. The filtering conditions include: For the VoNR voice environment, the expected MOS value is greater than the MOS threshold; For video downloading, the expected average downlink throughput of the RLC layer must be greater than the threshold of the average downlink throughput of the RLC layer and must not be lower than the minimum guaranteed downlink rate. For a video upload environment, the expected RLC layer uplink average throughput is greater than the RLC layer uplink average throughput threshold and cannot be lower than the minimum uplink guaranteed rate.

10. The method for evaluating service carrying capacity according to claim 2, wherein: The service satisfaction probability of a specific environment in a grid at a given moment is obtained based on the service satisfaction of different terminals and combined with the terminal penetration rate. The service satisfaction probability for the VoNR voice environment is expressed as: Among them, UE A UE B and UE RedCa Indicates different terminals.

Images