A method, device, equipment and medium for controlling inter-frequency and inter-system handover

Abstract

The application discloses a kind of frequency and system switching control method, device, equipment and medium, the switching control method includes obtaining the RSRP data, SINR data and RSRQ data of cell;RSRP data, SINR data and RSRQ data of cell are substituted into voice MOS fitting model, and the predicted value of cell user voice MOS is obtained;The voice MOS fitting model is built according to RSRP historical data, RSRQ historical data, SINR historical data and voice MOS historical data;The predicted value of cell user voice MOS is compared with preset threshold value, if the predicted value of cell user voice MOS is less than preset threshold value, then frequency switching or system switching is executed, so that when 5G user voice service perception is poor, frequency system switching process is initiated in advance, and user is guided to switch to the 5G frequency cell of more optimal MOS value, improves 5G user VoNR voice service perception, simultaneously avoids because voice service perception deterioration causes network problems such as call drop, word swallowing, guarantees the continuity of 5G user VoNR voice service.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a method, apparatus, device, and medium for switching control between different frequencies and different systems. Background Technology

[0002] Voice services are a crucial business for operators in the 5G era, and providing users with stable and high-quality voice services is a vital part of enhancing the 5G experience. To improve user experience, shorten call setup latency, and enhance the clarity of real-time voice communication, the 3GPP standard introduced VoNR (Voice over New Radio) based on 5G and IMS (IP Multimedia Subsystem). VoNR is the target voice solution for operators. VoNR fully leverages the advantages of 5G's large bandwidth and the high spectrum utilization and strong anti-fading characteristics of New Radio / antenna technologies. By using ultra-high definition EVS (Enhanced Voice Services) coding, it provides users with shorter voice call access latency and an ultra-high definition voice experience. However, VoNR is entirely carried by the 5G NR network, and voice quality is strongly correlated with network coverage and antenna transceiver performance. Therefore, accurately evaluating the 5G network's ability to carry VoNR is a key aspect of ensuring a positive VoNR service experience.

[0003] Since 4G VoLTE, Mean Opinion Score (MOS) has become the mainstream method for operations and maintenance personnel to evaluate voice quality. However, current technologies still rely on live network DT / CQT testing to obtain voice MOS values. This method cannot achieve all-weather, all-time, and all-area coverage, and requires a lot of manpower and resources for long-term data collection. Not only is it costly, but human error or accidental events can also affect the accuracy of VoLTE MOS, making it difficult to achieve comprehensive, truthful, and objective evaluation results. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to address the above-mentioned shortcomings of the prior art by proposing a method, device, equipment and medium for inter-frequency and inter-system handover control, so that 5G base stations can realize inter-frequency and inter-system handover based on the expected MOS value of voice users, and can carry out long-term data collection in all weather, all time period and all area, with comprehensive, true and objective evaluation results, and low cost, so as to improve the experience of 5G users' VoNR voice services.

[0005] In a first aspect, the present invention provides a method for switching control between different frequencies and different systems, which is applied to the network side of a 5G communication system. The method includes the following steps:

[0006] S1: Obtain the cell's RSRP data, SINR data, and RSRQ data;

[0007] S2: Substitute the RSRP data, SINR data, and RSRQ data of the cell into the voice MOS fitting model to obtain the predicted voice MOS value for the cell user; the voice MOS fitting model is built based on historical RSRP data, historical RSRQ data, historical SINR data, and historical voice MOS data;

[0008] S3: Compare the predicted MOS value of the cell user's voice with a preset threshold. If the predicted MOS value of the cell user's voice is less than the preset threshold, perform inter-frequency handover or inter-system handover to ensure the continuity of 5G user's VoNR voice service.

[0009] Furthermore, the method also includes:

[0010] S0: Build a speech MOS fitting model;

[0011] The construction of the speech MOS fitting model specifically includes:

[0012] A1: Collect a large amount of historical road test data. Each road test data includes a timestamp, RSRP value, RSRQ value, SINR value, and MOS value.

[0013] A2: Based on timestamps, construct a grid for the road test data;

[0014] A3: After grid resampling the MOS value, the RSRP value, RSRQ value, SINR value and MOS value are associated with the grid to obtain the drive test grid data;

[0015] A4: Based on the road test grid data, fit the speech MOS fitting model.

[0016] Further, in step A4, a speech MOS fitting model is constructed based on the drive test grid data, specifically including:

[0017] The first fitting model of MOS was constructed by fitting the road test grid data using multivariate nonlinear regression. The first fitting model of MOS using multivariate nonlinear regression is as follows:

[0018] MOS MNR =f(RSRP) MNR RSRQ MNR SINR MNR )+C MNR

[0019] In the formula: C MNR C is the bias constant for fitting a multivariate nonlinear regression. MNRThe value range is -0.5 to 0.5, MOS MNR RSRP is the MOS value in the multivariate nonlinear regression fitting. MNR RSRP and RSRQ values ​​in multivariate nonlinear regression fitting MNR SINR represents the RSRQ value in the multivariate nonlinear regression fitting. MNR SINR value in multivariate nonlinear regression fitting;

[0020] or,

[0021] The second fitting model of MOS is constructed by fitting the drive test grid data using LSTM time-series prediction. The second fitting model of MOS when using LSTM time-series prediction is as follows:

[0022] MOS LSTM =f(RSRP) LSTM RSRQ LSTM SINR LSTM )+C LSTM

[0023] In the formula: C LSTM C is the bias constant for LSTM time series prediction. LSTM The value range is -0.5 to 0.5, MOS LSTM RSRP represents the MOS value in LSTM timing prediction. LSTM RSRP and RSRQ values ​​in LSTM time series prediction LSTM SINR represents the RSRQ value in LSTM time series prediction. LSTM The SINR value in LSTM time series prediction;

[0024] or,

[0025] The third fitting model of MOS was constructed by fitting the road test grid data using an artificial neural network training algorithm. The third fitting model when using the artificial neural network training algorithm is as follows:

[0026] MOS AI =f(RSRP) AI RSRQ AI SINR AI )+C AI

[0027] In the formula: C AI C is the bias constant for the artificial neural network training algorithm. AI The value range is -0.5 to 0.5, MOS AI RSRP is the MOS value in the artificial neural network training algorithm. AI RSRP and RSRQ values ​​in artificial neural network training algorithms AISINR is the RSRQ value in the artificial neural network training algorithm. AI This refers to the SINR value in the artificial neural network training algorithm.

[0028] Further, in step S3, the preset threshold includes a first preset threshold and a second preset threshold, where the first preset threshold is greater than the second preset threshold. The step of comparing the predicted voice MOS value of the cell user with the preset threshold, and if the predicted voice MOS value of the cell user is less than the preset threshold, then performing inter-frequency handover or inter-system handover, specifically:

[0029] The predicted voice MOS value of the cell user is compared with the first preset threshold. If the predicted voice MOS value of the cell user is less than the first preset threshold, then inter-frequency handover is performed.

[0030] or,

[0031] The predicted voice MOS value of the cell user is compared with the second preset threshold. If the predicted voice MOS value of the cell user is less than the second preset threshold, a handover between systems is performed.

[0032] Secondly, the present invention provides a switching control device for different frequencies and different systems, applied to the network side of a 5G communication system, the device comprising:

[0033] The acquisition unit is used to acquire RSRP data, SINR data, and RSRQ data of the cell, and also to acquire historical RSRP data, historical RSRQ data, historical SINR data, and historical voice MOS data.

[0034] A construction unit, connected to the acquisition unit, is used to build a speech MOS fitting model based on RSRP historical data, RSRQ historical data, SINR historical data, and speech MOS historical data.

[0035] The calculation unit is connected to the acquisition unit and the construction unit respectively, and is used to substitute the RSRP data, SINR data and RSRQ data of the cell into the speech MOS fitting model to obtain the predicted value of the cell user's speech MOS.

[0036] A comparison unit, connected to the calculation unit, is used to compare the predicted MOS value of a cell user's voice with a preset threshold.

[0037] A switching unit, connected to the comparison unit, is used to perform inter-frequency handover or inter-system handover when the predicted MOS value of the cell user's voice is less than a preset threshold according to the comparison result of the comparison unit, so as to ensure the continuity of 5G user's VoNR voice service.

[0038] Furthermore, the building unit includes:

[0039] The collection module, connected to the acquisition unit, is used to collect a large amount of historical road test data. Each piece of road test data includes a timestamp, RSRP value, RSRQ value, SINR value, and MOS value.

[0040] The first construction module is connected to the collection module and is used to construct a grid of road test data based on timestamps.

[0041] The second construction module is connected to the collection module and the first construction module respectively. It is used to perform grid resampling on the MOS value and associate the RSRP value, RSRQ value, SINR value and MOS value with the grid to obtain the drive test grid data.

[0042] The fitting module, connected to the second construction module, is used to fit a speech MOS fitting model based on the road test grid data.

[0043] Furthermore, the fitting module includes:

[0044] The multivariate nonlinear regression fitting submodule, connected to the second construction module, is used to fit the road test grid data using multivariate nonlinear regression to construct the first fitting model of MOS. The first fitting model when using multivariate nonlinear regression is as follows:

[0045] MOS MNR =f(RSRP) MNR RSRQ MNR SINR MNR )+C MNR

[0046] In the formula: C MNR C is the bias constant for fitting a multivariate nonlinear regression. MNR The value range is -0.5 to 0.5, MOS MNR RSRP is the MOS value in the multivariate nonlinear regression fitting. MNR RSRP and RSRQ values ​​in multivariate nonlinear regression fitting MNR SINR represents the RSRQ value in the multivariate nonlinear regression fitting. MNR SINR value in multivariate nonlinear regression fitting;

[0047] The LSTM time-series prediction fitting submodule, connected to the second construction module, is used to fit the drive test grid data using LSTM time-series prediction to construct the second MOS fitting model. The second MOS fitting model when using LSTM time-series prediction for fitting is as follows:

[0048] MOS LSTM =f(RSRP) LSTM RSRQ LSTM SINRLSTM )+C LSTM

[0049] In the formula: C LSTM C is the bias constant for LSTM time series prediction. LSTM The value range is -0.5 to 0.5, MOS LSTM RSRP represents the MOS value in LSTM timing prediction. LSTM RSRP and RSRQ values ​​in LSTM time series prediction LSTM SINR represents the RSRQ value in LSTM time series prediction. LSTM SINR in LSTM time series prediction;

[0050] The artificial neural network training algorithm fitting submodule, connected to the second construction module, is used to fit the road test grid data using the artificial neural network training algorithm to construct the third fitting model of MOS. The third fitting model of MOS when using the artificial neural network training algorithm for fitting is as follows:

[0051] MOS AI =f(RSRP) AI RSRQ AI SINR AI )+C AI

[0052] In the formula: C AI C is the bias constant for the artificial neural network training algorithm. AI The value range is -0.5 to 0.5, MOS AI RSRP is the MOS value in the artificial neural network training algorithm. AI RSRP and RSRQ values ​​in artificial neural network training algorithms AI SINR is the RSRQ value in the artificial neural network training algorithm. AI This refers to the SINR value in the artificial neural network training algorithm. Further, the preset threshold includes a first preset threshold and a second preset threshold, where the first preset threshold is greater than the second preset threshold.

[0053] The switching unit includes a first switching unit and a second switching unit, and the first switching unit and the second switching unit are respectively connected to the comparison unit;

[0054] The first switching unit is used to perform inter-frequency handover when the predicted voice MOS value of the cell user is less than a first preset threshold according to the comparison result of the comparison unit; the second switching unit is used to perform inter-system handover when the predicted voice MOS value of the cell user is less than a second preset threshold according to the comparison result of the comparison unit.

[0055] Thirdly, the present invention provides an electronic device including a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the inter-frequency and inter-system switching method described in the first aspect.

[0056] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the inter-frequency and inter-system switching method described in the first aspect.

[0057] The beneficial effects achieved by this invention are:

[0058] 1. The methods and apparatus in this invention are based on historical RSRP data, historical RSRQ data, historical SINR data, and historical voice MOS data to build a voice MOS fitting model. Based on this model, the acquired RSRP measurement data, SINR measurement data, and RSRQ measurement data of the cell (including measurement reports from users of cells with different frequency points, different bandwidths, and different equipment types) are fitted to obtain the predicted voice MOS value of the cell user. The 5G base station can realize inter-frequency and inter-system handover based on the expected MOS value. When the 5G user voice service perception is poor, the inter-frequency and inter-system handover process is initiated in advance, and the user is guided to switch to a 5G inter-frequency cell with a better MOS value. This improves the 5G user VoNR voice service perception and avoids network problems such as dropped calls and swallowed words caused by the deterioration of voice service perception, thus ensuring the continuity of 5G user VoNR voice service.

[0059] 2. The device in this invention can accurately assess the VoNR voice service perception of 5G users by taking various indicators of the wireless environment as input, even when the terminal cannot measure MOS;

[0060] Meanwhile, considering that 5G cells under different frequencies, bandwidths, and equipment configurations have different capabilities to carry VoNR voice services, the method and apparatus for fitting voice MOS based on RSRP / RSRQ / SINR of the present invention can distinguish different frequencies, bandwidths, and equipment configurations to form models for fitting voice MOS, ensuring the accuracy and differentiation of voice MOS fitting.

[0061] 3. This invention differs from traditional interoperability schemes that use a single network metric (such as RSRP or RSRQ) for inter-frequency and inter-system handover. For VoNR voice service users in a cell, the invention determines the user's voice service perception based on the measurement report of the primary serving cell reported by the user. For users whose voice service perception is lower than a set threshold, inter-frequency and inter-system measurements are sent to ensure that the user can initiate the inter-frequency and inter-system handover process before the voice perception deteriorates, thereby improving the user's VoNR voice service perception.

[0062] 4. This invention is applicable to a cross-frequency handover scenario where there are more than two cells that meet the handover conditions. Based on the neighbor cell measurement report, the expected MOS value of the neighbor cell is further fitted to assist VoNR voice users in selecting cross-frequency neighbor cells with higher expected MOS values. This ensures that VoNR voice users can initiate handover to a better voice cell, thus guaranteeing users' inter-frequency mobility and VoNR voice service awareness. Attached Figure Description

[0063] Figure 1 This is a flowchart of the switching control between different frequencies and different systems in an embodiment of the present invention;

[0064] Figure 2 This is a schematic diagram of MOS testing in an embodiment of the present invention. Detailed Implementation

[0065] To enable those skilled in the art to better understand the technical solution of the present invention, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0066] It is understood that the specific embodiments and accompanying drawings described herein are merely for explaining the invention and are not intended to limit the invention.

[0067] It is understood that, without conflict, the various embodiments and features in the embodiments of the present invention can be combined with each other.

[0068] It is understood that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, while the parts unrelated to the present invention are not shown in the drawings.

[0069] It is understood that each unit or module involved in the embodiments of the present invention may correspond to only one entity structure, or may be composed of multiple entity structures, or multiple units or modules may be integrated into one entity structure.

[0070] It is understood that, without conflict, the functions and steps marked in the flowcharts and block diagrams of this invention may occur in a different order than that marked in the accompanying drawings.

[0071] It is understood that the flowcharts and block diagrams of this invention illustrate the possible architecture, functions, and operations of systems, apparatuses, devices, and methods according to various embodiments of this invention. Each block in the flowchart or block diagram may represent a unit, module, program segment, or code, containing executable instructions for implementing the specified function. Furthermore, each block or combination of blocks in the block diagram and flowchart can be implemented using a hardware-based system to achieve the specified function, or using a combination of hardware and computer instructions.

[0072] It is understood that the units and modules involved in the embodiments of the present invention can be implemented by software or by hardware. For example, the units and modules can be located in a processor.

[0073] Example 1:

[0074] like Figure 1 and Figure 2 As shown, this embodiment provides a method for inter-frequency and inter-system handover control, which is applied to the network side of a 5G communication system. This embodiment includes the following specific steps:

[0075] Step 1. Build a speech MOS fitting model

[0076] Step 1-1. Raw MOS Data Acquisition: Under well-optimized and continuously covered 5G networks with different frequency bands (e.g., 900M / 2.1G / 3.5GNR), different bandwidths (900MNR 5M / 10M, 2.1GNR 20M / 40M, 3.5GNR 100M), and different device types (900MNR 4T4R, 2.1G 4T4R / 2T2R), a large amount of drive test data containing timestamps, RSRP values, RSRQ values, SINR values, and MOS values ​​is collected through voice MOS boxes.

[0077] Step 1-2. Data Cleaning and Processing: Since the MOS value is generated every 8-9 seconds, while the RSRP, RSRQ, and SINR values ​​are generated at different intervals depending on the chip and the capabilities of the MOS device, typically 200ms or less, the MOS value cannot be perfectly synchronized with the RSRP / RSRQ / SINR values. Therefore, the RSRP / RSRQ / SINR and MOS values ​​in the table need to be gridded according to time.

[0078] In this embodiment, the grid processing can employ the nearest neighbor allocation method, the mode algorithm, bilinear interpolation, or cubic convolution interpolation. Taking the nearest neighbor allocation method as an example: First, all sampling points in the table are processed according to a 1-second time grid granularity, and each RSRP, RSRQ, SINR, and MOS value is grouped to the nearest whole second. Second, since a MOS value is generated every 8-9 seconds, a large number of null values ​​will be generated in the 1-second grid data table. The generation of MOS values ​​is closely related to the RSRP / RSRQ / SINR values ​​of the previous 8-9 seconds. Therefore, the MOS values ​​can be backfilled upwards until the previous MOS value is reached, resulting in a data table containing RSRP / RSRQ / SINR / MOS values ​​arranged in second order, as shown in the table below.

[0079]

[0080]

[0081] Steps 1-3. Constructing a speech MOS fitting model: Based on the data table obtained in Steps 1-2, a speech MOS fitting model can be established. The speech MOS fitting model can be used for MOS prediction and fitting. Specific methods can include multivariate nonlinear regression fitting, LSTM-based time-series prediction, or artificial neural training algorithms (such as RNN, DNN, etc.) to obtain the following formula.

[0082] MOS = f(RSRP, RSRQ, SINR) + c

[0083] Where c is a constant, and a unique MOS value can be obtained whenever a set of RSRP, RSRQ, and SINR values ​​are given.

[0084] Specifically, based on the drive test grid data, a speech MOS fitting model is derived, which includes:

[0085] The first fitting model of MOS was constructed by fitting the road test grid data using multivariate nonlinear regression. The first fitting model of MOS using multivariate nonlinear regression is as follows:

[0086] MOS MNR =f(RSRP) MNR RSRQ MNR SINR MNR )+C MNR

[0087] In the formula: C MNR C is the bias constant for fitting a multivariate nonlinear regression. MNR The value range is -0.5 to 0.5, MOS MNR RSRP is the MOS value in the multivariate nonlinear regression fitting. MNR RSRP and RSRQ values ​​in multivariate nonlinear regression fitting MNR SINR represents the RSRQ value in the multivariate nonlinear regression fitting. MNR SINR value in multivariate nonlinear regression fitting;

[0088] or,

[0089] The second fitting model of MOS is constructed by fitting the drive test grid data using LSTM time-series prediction. The second fitting model of MOS when using LSTM time-series prediction is as follows:

[0090] MOS LSTM =f(RSRP) LSTM RSRQ LSTM SINR LSTM )+CLSTM

[0091] In the formula: C LSTM C is the bias constant for LSTM time series prediction. LSTM The value range is -0.5 to 0.5, MOS LSTM RSRP represents the MOS value in LSTM timing prediction. LSTM RSRP and RSRQ values ​​in LSTM time series prediction LSTM SINR represents the RSRQ value in LSTM time series prediction. LSTM The SINR value in LSTM time series prediction;

[0092] or,

[0093] An artificial neural network training algorithm was used to fit the drive test grid data to construct the third fitting model of MOS. The third fitting model of MOS when using the artificial neural network training algorithm is as follows:

[0094] MOS AI =f(RSRP) AI RSRQ AI SINR AI )+C AI

[0095] In the formula: C AI C is the bias constant for the artificial neural network training algorithm. AI The value range is -0.5 to 0.5, MOS AI RSRP is the MOS value in the artificial neural network training algorithm. AI RSRP and RSRQ values ​​in artificial neural network training algorithms AI SINR is the RSRQ value in the artificial neural network training algorithm. AI This refers to the SINR value in the artificial neural network training algorithm.

[0096] The speech MOS fitting model is obtained based on the constructed MOS first fitting model, MOS second fitting model, or MOS third fitting model.

[0097] After obtaining the speech MOS fitting model, the accuracy of the prediction model can be calculated using regression evaluation indicators such as MSE, RMSE, MAE, MAPE, and R2. When the accuracy of the model meets the expected requirements, it can be considered that the MOS prediction model can obtain the corresponding MOS value based on the measured values ​​of RSRP, RSRQ, and SINR, that is, the expected MOS value in different wireless environments. It should be noted that, as mentioned earlier, the MOS prediction model is different for different frequency bands, different bandwidths, and different device types. Therefore, steps 1-1 to 1-3 need to be repeated separately to establish the MOS prediction and fitting model. Finally, MOS prediction and fitting models for different frequency bands, different bandwidths, and different device types can be obtained. The specific steps for calculating the accuracy of the prediction model using regression evaluation indicators such as MSE, RMSE, MAE, MAPE, and R2 are as follows:

[0098] When using multivariate nonlinear regression to fit and determine the first fitting model of MOS, the specific steps include the following:

[0099] B1.1: The first fitting model of MOS is obtained by using multivariate nonlinear regression;

[0100] B1.2: The accuracy of the first fitting model of MOS is obtained by performing a first regression evaluation on the first regression evaluation index;

[0101] B1.3: Compare the accuracy of the first fitting model of MOS with the first accuracy: If the accuracy of the first fitting model of MOS is greater than or equal to the first accuracy, then the first fitting model of MOS is determined; if the accuracy of the first fitting model of MOS is less than the first accuracy, then repeat steps B1.1 to B1.3 until the first fitting model of MOS is determined.

[0102] When using LSTM time-series prediction to fit and determine the second fitting model for MOS, the specific steps include the following:

[0103] B2.1: The second fitting model of MOS is obtained by fitting the time series prediction using LSTM;

[0104] B2.2: The accuracy of the MOS second-fit model is obtained by performing a second regression evaluation on the second regression evaluation index.

[0105] B2.3: Compare the accuracy of the second fitting model of MOS with the second accuracy: If the accuracy of the second fitting model of MOS is greater than or equal to the second accuracy, then the second fitting model of MOS is determined; if the accuracy of the second fitting model of MOS is less than the second accuracy, then repeat steps B2.1 to B2.3 until the second fitting model of MOS is determined.

[0106] When using an artificial neural network training algorithm to fit and determine the third fitting model of MOS, the specific steps include the following:

[0107] B3.1: The third fitting model of MOS is obtained by fitting using an artificial neural network training algorithm;

[0108] B3.2: The accuracy of the MOS third-fit model is obtained by performing a third-regression evaluation on the third-regression evaluation index.

[0109] B3.3: Compare the accuracy of the MOS third fitting model with the third accuracy: If the accuracy of the MOS third fitting model is greater than or equal to the third accuracy, then the MOS third fitting model is determined; if the accuracy of the MOS third fitting model is less than the third accuracy, then repeat steps B3.1 to B3.3 until the MOS third fitting model is determined.

[0110] Wherein: the first precision, the second precision, and the third precision are all preset values; the first regression evaluation index, the second regression evaluation index, and the third regression evaluation index all include mean square error (MSE), root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and coefficient of determination (R-Squared).

[0111] Step 2. Users conducting VoNR voice services in this 5G cell report RSRP data, SINR data, and RSRQ data containing the data of this cell. After receiving the RSRP data, SINR data, and RSRQ data, the base station can obtain the user's voice MOS value corresponding to the measurement report based on the MOS prediction and fitting model obtained and put in Step 1.

[0112] Step 3. For inter-frequency handover, when the MOS value fitted from the measurement report reported by the VoNR voice user is less than the set threshold 1 (this threshold can be set to 3.5), an inter-frequency cell measurement command is issued to the user through the RRCConnection Reconfiguration message, instructing the user to perform inter-frequency cell measurement, and inter-frequency handover is performed based on the measurement results; if there are more than two neighboring cells that meet the handover conditions in the target neighboring cell, the traditional handover is to initiate a handover request based on the best neighboring cell in the measurement results. In this scheme, it is necessary to further perform MOS fitting on the measurement results of the neighboring cells through MOS prediction and fitting models, select the neighboring cell with the largest fitted MOS value as the target neighboring cell, and issue a handover command to instruct the terminal to perform the handover;

[0113] Step 4. For inter-system handover, if the MOS value fitted by the measurement report reported by the user before the inter-frequency handover is completed is less than the set threshold 2 (this threshold can be set to 3.0), then an inter-system measurement command will be issued to the user through the RRCConnection Reconfiguration message. The user will then perform inter-system cell measurement and perform inter-system handover based on the measurement results.

[0114] Frequency switching and system switching are separate processes. Generally, frequency switching is performed first, followed by system switching.

[0115] Example 2:

[0116] This embodiment discloses a switching control device for different frequencies and different systems. This device is applied to the network side of a 5G communication system and includes:

[0117] The acquisition unit is used to acquire RSRP data, SINR data, and RSRQ data of the cell, and also to acquire historical RSRP data, historical RSRQ data, historical SINR data, and historical voice MOS data.

[0118] The construction unit, connected to the acquisition unit, is used to build a speech MOS fitting model based on historical RSRP data, historical RSRQ data, historical SINR data, and historical speech MOS data.

[0119] The calculation unit, connected to the acquisition unit and the construction unit respectively, is used to substitute the RSRP data, SINR data and RSRQ data of the cell into the speech MOS fitting model to obtain the predicted speech MOS value of the cell user.

[0120] The comparison unit, connected to the calculation unit, is used to compare the predicted MOS value of the cell user's voice with a preset threshold.

[0121] The switching unit, connected to the comparison unit, is used to perform inter-frequency handover or inter-system handover when the predicted MOS value of the cell user's voice is less than a preset threshold, based on the comparison result of the comparison unit, so as to ensure the continuity of 5G user's VoNR voice service.

[0122] Specifically, the building blocks include:

[0123] The collection module, connected to the acquisition unit, is used to collect a large amount of historical road test data. Each piece of road test data includes a timestamp, RSRP value, RSRQ value, SINR value, and MOS value.

[0124] The first construction module, connected to the collection module, is used to build a grid of road test data based on timestamps.

[0125] The second construction module is connected to the collection module and the first construction module respectively. It is used to perform grid resampling on the MOS value and associate the RSRP value, RSRQ value, SINR value and MOS value with the grid to obtain the road test grid data.

[0126] The fitting module, connected to the second construction module, is used to fit a speech MOS fitting model based on the drive test grid data. The fitting module includes:

[0127] The multivariate nonlinear regression fitting submodule, connected to the second construction module, is used to fit the road test grid data using multivariate nonlinear regression to construct the first fitting model for MOS. The first fitting model when using multivariate nonlinear regression is as follows:

[0128] MOS MNR =f(RSRP) MNR RSRQ MNR SINR MNR )+C MNR

[0129] In the formula: C MNR C is the bias constant for fitting a multivariate nonlinear regression. MNR The value range is -0.5 to 0.5, MOS MNR RSRP is the MOS value in the multivariate nonlinear regression fitting. MNR RSRP and RSRQ values ​​in multivariate nonlinear regression fitting MNR SINR represents the RSRQ value in the multivariate nonlinear regression fitting. MNR SINR value in multivariate nonlinear regression fitting;

[0130] The LSTM time-series prediction fitting submodule, connected to the second construction module, is used to fit the drive test grid data using LSTM time-series prediction to construct the second MOS fitting model. The second MOS fitting model when using LSTM time-series prediction is as follows:

[0131] MOS LSTM =f(RSRP) LSTM RSRQ LSTM SINR LSTM )+C LSTM

[0132] In the formula: C LSTM C is the bias constant for LSTM time series prediction. LSTM The value range is -0.5 to 0.5, MOS LSTM RSRP represents the MOS value in LSTM timing prediction. LSTM RSRP and RSRQ values ​​in LSTM time series prediction LSTMSINR represents the RSRQ value in LSTM time series prediction. LSTM SINR in LSTM time series prediction;

[0133] The artificial neural network training algorithm fitting submodule, connected to the second construction module, is used to fit the road test grid data using the artificial neural network training algorithm to construct the third MOS fitting model. The third MOS fitting model when using the artificial neural network training algorithm is as follows:

[0134] MOS AI =f(RSRP) AI RSRQ AI SINR AI )+C AI

[0135] In the formula: C AI C is the bias constant for the artificial neural network training algorithm. AI The value range is -0.5 to 0.5, MOS AI RSRP is the MOS value in the artificial neural network training algorithm. AI RSRP and RSRQ values ​​in artificial neural network training algorithms AI SINR is the RSRQ value in the artificial neural network training algorithm. AI This refers to the SINR value in the artificial neural network training algorithm.

[0136] Specifically, the preset threshold includes a first preset threshold (which can be set to 3.5) and a second preset threshold (which can be set to 3.0), wherein the first preset threshold is greater than the second preset threshold;

[0137] The switching unit includes a first switching unit and a second switching unit, which are respectively connected to the comparison unit;

[0138] The first handover unit is used to perform inter-frequency handover when the predicted voice MOS value of the cell user is less than a first preset threshold according to the comparison result of the comparison unit; the second handover unit is used to perform inter-system handover when the predicted voice MOS value of the cell user is less than a second preset threshold according to the comparison result of the comparison unit.

[0139] The switching control process between different frequencies and different systems implemented in Example 2 is the same as that implemented in Example 1.

[0140] Both Examples 1 and 2 establish a voice MOS fitting model based on historical RSRP, RSRQ, SINR, and voice MOS data. This model is then used to fit the acquired RSRP, SINR, and RSRQ measurement data (including measurement reports from users covering different frequency points, bandwidths, and equipment types) of the cell to obtain the predicted voice MOS value for the cell user. The 5G base station can then perform inter-frequency and inter-system handover based on the expected MOS value. When the 5G user's voice service experience is poor, the inter-frequency and inter-system handover process can be initiated in advance, guiding the user to switch to a 5G inter-frequency cell with a better MOS value. This improves the 5G user's VoNR voice service experience and avoids network problems such as dropped calls and swallowed words caused by deteriorating voice service experience, ensuring the continuity of the 5G user's VoNR voice service.

[0141] The device in Example 2 can accurately assess the VoNR voice service perception of 5G users by taking various indicators of the wireless environment as input when the terminal cannot measure MOS. At the same time, considering that the ability of 5G cells to carry VoNR voice services varies under different frequency points, different bandwidths, and different equipment forms, the method and device for fitting voice MOS based on RSRP / RSRQ / SINR in Examples 1 and 2 can distinguish different frequency points, bandwidths, and equipment forms to form models for fitting voice MOS, ensuring the accuracy and differentiation of voice MOS fitting.

[0142] Both Implementation Examples 1 and Implementation Examples 2 differ from traditional interoperability schemes that use a single network metric (such as RSRP or RSRQ) for inter-frequency and inter-system handover. For VoNR voice service users in a cell, the user's voice service perception is determined based on the measurement report of the primary serving cell reported by the user. For users whose voice service perception is lower than a set threshold, inter-frequency and inter-system measurements are sent to ensure that the user can initiate the inter-frequency and inter-system handover process before the voice perception deteriorates, thereby improving the user's VoNR voice service perception.

[0143] Both Example 1 and Example 2 are applicable to a single inter-frequency handover scenario where there are more than two cells that meet the handover conditions. Based on the neighbor cell measurement report, the expected MOS value of the neighbor cell is further fitted to assist VoNR voice users in selecting inter-frequency neighbor cells with higher expected MOS values. This ensures that VoNR voice users can initiate handover to a better voice cell, thus guaranteeing users' inter-frequency mobility and VoNR voice service awareness.

[0144] Example 3:

[0145] Based on the same technical concept as Embodiment 1, the electronic device provided in this embodiment includes a memory and a processor. The memory stores a computer program. When the processor runs the computer program stored in the memory, the processor executes the inter-frequency and inter-system switching method described in Embodiment 1.

[0146] Example 4:

[0147] Based on the same technical concept as Embodiment 1, this embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the inter-frequency and inter-system switching method described in Embodiment 1.

[0148] It is understood that the above embodiments are merely exemplary implementations used to illustrate the principles of the present invention, and the present invention is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A method for switching control between different frequencies and different systems, applied to the network side of a 5G communication system, characterized in that, The method includes the following steps: Obtain the RSRP data, SINR data, and RSRQ data of the cell; Substitute the RSRP data, SINR data, and RSRQ data of the cell into the speech MOS fitting model to obtain the predicted speech MOS value for the cell user. The predicted MOS value of the cell user's voice is compared with a preset threshold. If the predicted MOS value of the cell user's voice is less than the preset threshold, inter-frequency handover or inter-system handover is performed to ensure the continuity of 5G user's VoNR voice service. The speech MOS fitting model is constructed based on historical RSRP, RSRQ, SINR, and speech MOS data. The construction of the speech MOS fitting model specifically includes: collecting historical road test data containing timestamps, RSRP values, RSRQ values, SINR values, and MOS values; constructing a grid of road test data based on the timestamps; resampling the MOS values ​​using a grid, and then associating the RSRP, RSRQ, SINR, and MOS values ​​with the grid to obtain road test grid data; and fitting the speech MOS fitting model based on the road test grid data.

2. The method for switching control between different frequencies and different systems according to claim 1, characterized in that, The process of fitting a speech MOS fitting model based on drive test grid data specifically includes: The first fitting model of MOS was constructed by fitting the road test grid data using multivariate nonlinear regression. The first fitting model of MOS using multivariate nonlinear regression is as follows: In the formula: is the bias constant for fitting the multivariate nonlinear regression. The value range is -0.5 to 0.

5. The MOS value is used in the fitting of a multivariate nonlinear regression. RSRP value in multivariate nonlinear regression fitting The RSRQ value is used in the fitting of a multivariate nonlinear regression. SINR value in multivariate nonlinear regression fitting; or, The second fitting model of MOS is constructed by fitting the drive test grid data using LSTM time-series prediction. The second fitting model of MOS when using LSTM time-series prediction is as follows: In the formula: This is the bias constant for LSTM time series prediction. The value range is -0.5 to 0.

5. The MOS value in LSTM timing prediction. RSRP value in LSTM time series prediction This refers to the RSRQ value in LSTM time series prediction. The SINR value in LSTM time series prediction; or, An artificial neural network training algorithm was used to fit the drive test grid data to construct the third fitting model of MOS. The third fitting model of MOS when using the artificial neural network training algorithm is as follows: In the formula: This is the bias constant for the artificial neural network training algorithm. The value range is -0.5 to 0.

5. This refers to the MOS value in the artificial neural network training algorithm. RSRP value in artificial neural network training algorithms This refers to the RSRQ value in the artificial neural network training algorithm. This refers to the SINR value in the artificial neural network training algorithm. The speech MOS fitting model is obtained based on the constructed MOS first fitting model, MOS second fitting model, or MOS third fitting model.

3. The inter-frequency and inter-system switching control method according to claim 1 or 2, characterized in that, The preset thresholds include a first preset threshold and a second preset threshold, where the first preset threshold is greater than the second preset threshold. The process involves comparing the predicted voice MOS value of a cell user with the preset thresholds. If the predicted voice MOS value is less than the preset threshold, then inter-frequency handover or inter-system handover is performed. Specifically: The predicted voice MOS value of the cell user is compared with the first preset threshold. If the predicted voice MOS value of the cell user is less than the first preset threshold, then inter-frequency handover is performed. or, The predicted voice MOS value of the cell user is compared with the second preset threshold. If the predicted voice MOS value of the cell user is less than the second preset threshold, a handover between systems is performed.

4. A switching control device for different frequencies and systems, applied to the network side of a 5G communication system, characterized in that, include: The acquisition unit is used to acquire RSRP data, SINR data, and RSRQ data of the cell, and also to acquire historical RSRP data, historical RSRQ data, historical SINR data, and historical voice MOS data. A construction unit, connected to the acquisition unit, is used to build a speech MOS fitting model based on RSRP historical data, RSRQ historical data, SINR historical data, and speech MOS historical data. The building unit includes: The collection module, connected to the acquisition unit, is used to collect a large amount of historical road test data. Each piece of road test data includes a timestamp, RSRP value, RSRQ value, SINR value, and MOS value. The first construction module is connected to the collection module and is used to build a grid of road test data based on timestamps. The second construction module is connected to the collection module and the first construction module respectively. It is used to perform grid resampling on the MOS value and associate the RSRP value, RSRQ value, SINR value and MOS value with the grid to obtain the drive test grid data. The fitting module, connected to the second construction module, is used to fit a speech MOS fitting model based on the road test grid data. The calculation unit is connected to the acquisition unit and the construction unit respectively, and is used to substitute the RSRP data, SINR data and RSRQ data of the cell into the speech MOS fitting model to obtain the predicted value of the cell user's speech MOS. A comparison unit, connected to the calculation unit, is used to compare the predicted MOS value of a cell user's voice with a preset threshold. A switching unit, connected to the comparison unit, is used to perform inter-frequency handover or inter-system handover when the predicted MOS value of the cell user's voice is less than a preset threshold according to the comparison result of the comparison unit, so as to ensure the continuity of 5G user's VoNR voice service.

5. The inter-frequency and inter-system switching control device according to claim 4, characterized in that, The fitting module includes: The multivariate nonlinear regression fitting submodule, connected to the second construction module, is used to fit the road test grid data using multivariate nonlinear regression to construct the first MOS fitting model. The first MOS fitting model when using multivariate nonlinear regression is as follows: In the formula: is the bias constant for fitting the multivariate nonlinear regression. The value range is -0.5 to 0.

5. The MOS value is used in the fitting of a multivariate nonlinear regression. RSRP value in multivariate nonlinear regression fitting The RSRQ value is used in the fitting of a multivariate nonlinear regression. SINR value in multivariate nonlinear regression fitting; The LSTM time-series prediction fitting submodule, connected to the second construction module, is used to fit the drive test grid data using LSTM time-series prediction to construct the second MOS fitting model. The second MOS fitting model when using LSTM time-series prediction for fitting is as follows: In the formula: This is the bias constant for LSTM time series prediction. The value range is -0.5 to 0.

5. The MOS value in LSTM timing prediction. RSRP value in LSTM time series prediction This refers to the RSRQ value in LSTM time series prediction. SINR in LSTM time series prediction; The artificial neural network training algorithm fitting submodule, connected to the second construction module, is used to fit the road test grid data using the artificial neural network training algorithm to construct the third fitting model of MOS. The third fitting model of MOS when using the artificial neural network training algorithm for fitting is as follows: In the formula: This is the bias constant for the artificial neural network training algorithm. The value range is -0.5 to 0.

5. This refers to the MOS value in the artificial neural network training algorithm. RSRP value in artificial neural network training algorithms This refers to the RSRQ value in the artificial neural network training algorithm. This refers to the SINR value in the artificial neural network training algorithm.

6. The inter-frequency and inter-system switching control device according to any one of claims 4 or 5, characterized in that, The preset threshold includes a first preset threshold and a second preset threshold, wherein the first preset threshold is greater than the second preset threshold; The switching unit includes a first switching unit and a second switching unit, and the first switching unit and the second switching unit are respectively connected to the comparison unit; The first switching unit is used to perform inter-frequency handover when the predicted voice MOS value of the cell user is less than a first preset threshold according to the comparison result of the comparison unit; the second switching unit is used to perform inter-system handover when the predicted voice MOS value of the cell user is less than a second preset threshold according to the comparison result of the comparison unit.

7. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the inter-frequency and inter-system switching method according to any one of claims 1 to 3.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the inter-frequency and inter-system switching method as described in any one of claims 1 to 3.

Images