A communication method and base station

Abstract

The embodiment of the application discloses a communication method, which comprises: determining historical frequency offset data; wherein the historical frequency offset data represents a frequency offset condition existing when data is transmitted between a terminal; predicting a current frequency offset condition according to the historical frequency offset data to obtain target frequency offset data; and performing frequency offset compensation on a signal transmitted with the terminal based on the target frequency offset data. The embodiment of the application also discloses a base station.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of communication, in particular to a communication method and a base station. BACKGROUND

[0002] If the terminal is in a moving state in the process of communicating with the base station, at this time, in order to ensure the reliability of data transmission, the transmitted data is compensated for frequency offset. In the prior art, after the base station receives the uplink data sent by the terminal through the uplink, the base station predicts the frequency offset estimation value of the downlink data sent to the terminal according to the uplink data, and compensates for the frequency offset of the downlink data based on the frequency offset estimation value; however, in the case of no uplink data for a long time, the frequency offset estimation value is not updated, at this time, the current downlink data is compensated for frequency offset based on the previous frequency offset estimation value, the compensation effect is poor and the downlink demodulation performance of the terminal is reduced. SUMMARY

[0003] To solve the above technical problems, the embodiments of the present application expect to provide a communication method and a base station, which solve the problems of poor compensation effect and reduced downlink demodulation performance of the terminal in the related art.

[0004] The technical scheme of the present application is implemented as follows:

[0005] A communication method, wherein the method comprises:

[0006] determining historical frequency offset data; wherein the historical frequency offset data represents the frequency offset condition existing when data is transmitted between the terminal and the base station;

[0007] predicting the current frequency offset condition according to the historical frequency offset data to obtain target frequency offset data;

[0008] compensating for the frequency offset of the signal transmitted between the terminal and the base station based on the target frequency offset data.

[0009] In the above scheme, the determination of the historical frequency offset data comprises:

[0010] cyclically storing a frequency offset estimation data sequence of a target length, the frequency offset estimation data sequence including a target number of frequency offset estimation data;

[0011] determining the historical frequency offset data based on the frequency offset estimation data sequence of the target length.

[0012] In the above scheme, the determination of the historical frequency offset data comprises:

[0013] in the case where data transmitted by the terminal is not received within a target time period but data needs to be sent to the terminal, determining the historical frequency offset data; the target time period is determined based on the period of the time slot duration of the carrier scheduled when the terminal transmits data and the scheduling number threshold.

[0014] The method further comprises:

[0015] updating the target frequency offset data into the frequency offset estimation data sequence.

[0016] In the above solution, the method further comprises:

[0017] determining a sequence accuracy of the frequency offset estimation data sequence.

[0018] In a case where the sequence accuracy meets a correct rate threshold requirement, taking the frequency offset estimation data sequence as the historical frequency offset data.

[0019] In the above solution, the frequency offset estimation data comprises a frequency offset value and a corresponding data verification identifier; the data verification identifier is used to represent whether the data is decoded correctly under the corresponding frequency offset value.

[0020] The method further comprises:

[0021] determining the sequence accuracy based on the data verification identifier and a number of frequency offset values in the frequency offset estimation sequence.

[0022] In the above solution, the method further comprises:

[0023] In a case where the sequence accuracy does not meet the correct rate threshold requirement, taking frequency offset data used for frequency offset compensation of a signal transmitted between the terminal in the last time as target frequency offset data, and performing frequency offset compensation on the signal transmitted between the terminal based on the target frequency offset data.

[0024] In the above solution, the method further comprises:

[0025] processing the historical frequency offset data based on a target frequency offset prediction model to predict the current frequency offset condition.

[0026] The training method of the target frequency offset prediction model comprises:

[0027] determining to-be-trained frequency offset data in different scenarios and a confidence degree corresponding to the to-be-trained frequency offset data.

[0028] performing model training on an initial frequency offset prediction model based on the to-be-trained frequency offset data and the confidence degree to obtain the target frequency offset prediction model.

[0029] In the scheme, the determining of the to-be-trained frequency offset data in different scenarios comprises:

[0030] Obtaining initial historical record data in different scenarios;

[0031] Filtering the initial historical record data based on the running parameters corresponding to the terminal to obtain target historical record data;

[0032] Determining the to-be-trained frequency offset data based on the frequency offset values in the target historical record data.

[0033] In the scheme, the filtering of the initial historical record data based on the running parameters corresponding to the terminal to obtain target historical record data comprises:

[0034] Filtering intermediate historical record data generated when the terminal transmits data at different running speeds from the initial historical record data;

[0035] Filtering target historical record data in which data decoding of consecutive N data frames is correct from the intermediate historical record data.

[0036] A base station, comprising a processor and a memory;

[0037] The communication bus is used to realize the communication connection between the processor and the memory;

[0038] The processor is used to execute the communication program stored in the memory to realize the following steps:

[0039] Determining historical frequency offset data; wherein the historical frequency offset data represents the frequency offset condition existing when transmitting signals between the terminal;

[0040] According to the historical frequency offset data, the current frequency offset condition is predicted to obtain target frequency offset data;

[0041] Based on the target frequency offset data, the frequency offset compensation is performed on the signals transmitted between the terminal. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 A flowchart of a communication method according to an embodiment of the present application;

[0043] Figure 2 A flowchart of a model method according to an embodiment of the present application;

[0044] Figure 3 A network architecture diagram of an initial frequency domain prediction model in a communication method according to an embodiment of the present application;

[0045] Figure 4A schematic diagram of the network architecture of the initial frequency domain prediction model in another communication method provided in this application embodiment;

[0046] Figure 5 This is a flowchart illustrating the process of obtaining target historical data in a communication method provided in an embodiment of this application.

[0047] Figure 6 A flowchart illustrating another communication method provided in an embodiment of this application;

[0048] Figure 7 A flowchart illustrating yet another communication method provided in an embodiment of this application;

[0049] Figure 8 A schematic diagram of a frequency offset estimation data sequence in a communication method provided in an embodiment of this application;

[0050] Figure 9 A flowchart illustrating the determination of a target time period in a communication method provided in an embodiment of this application;

[0051] Figure 10 This is a schematic diagram of the structure of the communication device provided in the embodiments of this application;

[0052] Figure 11 This is a schematic diagram of the structure of a base station provided in an embodiment of this application. Detailed Implementation

[0053] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0054] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0055] Before providing a further detailed description of the embodiments of this application, the nouns and terms involved in the embodiments of this application will be explained first. The nouns and terms involved in the embodiments of this disclosure are subject to the following interpretations.

[0056] (I) Doppler frequency shift

[0057] Orthogonal Frequency Division Multiplexing (OFDM) is a key technology in Long Term Evolution (LTE), 5th Generation Mobile Communication Technology New Radio (5G NR) and other communication systems. OFDM has high spectral efficiency, supports high-speed data transmission, and supports flexible bandwidth configuration. OFDM is very sensitive to phase noise and carrier frequency offset, and the carrier frequency offset will destroy the orthogonality between subcarriers.

[0058] When a mobile terminal is in motion, especially in high-speed communication, the signal frequency of the mobile terminal and the base station receiving end will change, which is called the Doppler effect. The frequency shift caused by the Doppler effect is called the Doppler shift, and its formula is: Where θ is the angle between the moving direction of the mobile terminal and the incident wave direction, v is the moving speed of the mobile terminal, c is the propagation speed of electromagnetic waves, and f is the carrier frequency.

[0059] (II) Uplink and downlink

[0060] A wireless communication system can support communication for multiple wireless communication devices at the same time. A wireless communication device can communicate with one or more base stations (also referred to as access points, or nodes) via transmission on the uplink and the downlink. The uplink (or reverse link) is the communication link from the wireless communication device (terminal) to the base station, and the downlink (or forward link) is the communication link from the base station to the wireless communication device.

[0061] (III) Wireless channel

[0062] During transmission, the wireless signal is reflected, scattered, diffracted, etc., and the signal received at the receiving end is a superposition of multiple time-delayed, phase-differentiated signal components. Changes in the communication environment will cause the signal at the receiving end to change over time, resulting in interference and distortion, and causing the received signal to fade.

[0063] During the movement of the mobile terminal, for example, when the terminal user is on a high-speed train, the terminal user experiences a variety of scenarios, such as viaducts, open areas, mountains, stations, etc. The different geographical structures and scatterer distributions of different scenarios will cause differences in the multipath effect and large-scale fading of the high-speed train wireless channel.

[0064] The related technologies involved in the embodiments of the present application are described below.

[0065] When a terminal is in motion, especially in high-speed scenarios such as high-speed railway scenarios and flight scenarios, a base station receiving system needs to accurately estimate a frequency offset and compensate for the frequency offset of received uplink data to ensure system performance. Similarly, when downlink data is transmitted, frequency offset compensation also needs to be performed to ensure the demodulation performance of the terminal. At present, most of the frequency offset compensation methods are based on uplink service channel (Physical Uplink Shared Channel, PUSCH) data for frequency offset estimation. For scenarios without uplink service channel data in a period of time, a sounding reference signal (Sounding Reference Signal, SRS) is also used for frequency offset estimation, but the periodicity of the SRS determines that it cannot perform frequency offset estimation at every scheduling moment. Therefore, the commonly used solution is to compensate for the frequency offset of downlink data according to historical frequency offset estimation values.

[0066] Solution 1 in the related art: In any cell corresponding to a remote radio unit (RRU) of a base station, a frequency offset estimation value of each user equipment (terminal) is estimated and saved according to a demodulation reference signal of an uplink between the RRU and each user equipment, a frequency offset pre-compensation value is determined based on the frequency offset estimation value of each user equipment, and the downlink data is compensated by using the frequency offset pre-compensation value. The disadvantage of this solution is that it needs to rely on uplink data for frequency offset estimation, and in a scenario without downlink data for a period of time, the frequency offset estimation value is not continuously updated, and the historical value of the frequency offset estimation is used for frequency offset compensation of the downlink data, which reduces the downlink demodulation performance of the terminal.

[0067] Solution 2 in the related art: Reference signals are respectively obtained at continuous sampling points, a cyclic prefix and tail data of each reference signal are extracted, a phase difference of each reference signal and an average phase difference of all reference signals are calculated through the cyclic prefix and the tail data of each reference signal, a frequency offset value of each reference signal is obtained through the average phase difference and the phase difference of each reference signal, and each reference signal is respectively compensated for the frequency offset through the frequency offset value of each reference signal. The disadvantage of this solution is that it still needs to rely on uplink data for frequency offset estimation of the current frame, and cannot handle scenarios without uplink data.

[0068] Therefore, an embodiment of the present application provides a communication method, which can be applied to a base station. As shown in FIG. 1, the method comprises the following steps: Figure 1

[0069] Step 101: Determine historical frequency offset data.

[0070] ​The historical frequency offset data represents a frequency offset existing between the base station and the terminal when transmitting data. The historical frequency offset data can be frequency offset data generated by the base station and the terminal when transmitting data at a historical time point. Alternatively, the historical frequency offset data can include at least one frequency offset estimation data, and each frequency offset estimation data can include a frequency offset value and a data check code of the frequency offset value. If the terminal is in a moving state when transmitting a signal with the base station, the vibration frequency of the received signal of the base station will be different from the frequency of the signal transmitted by the terminal, and the resulting frequency deviation is the frequency offset. By processing the signal through the frequency offset value, the frequency offset existing in the transmission process of the signal can be eliminated, thereby improving the reliability of communication.

[0071] In a feasible implementation, during normal operation of the base station and when there is uplink data, a plurality of frequency offset estimation data corresponding to the uplink data can be cyclically stored, and a sequence formed by the plurality of frequency offset estimation data is used as the historical frequency offset data. Alternatively, when there is uplink data and downlink data during data interaction between the base station and the terminal, the frequency offset estimation data corresponding to the uplink data and the frequency offset estimation data corresponding to the downlink data can be cyclically stored as the historical frequency offset data.

[0072] Step 102, predicting the current frequency offset condition according to the historical frequency offset data to obtain target frequency offset data.

[0073] The target frequency offset data is frequency offset data obtained by predicting the current frequency offset condition according to the historical frequency offset data.

[0074] Since the frequency offset condition existing in the uplink data and the frequency offset condition existing in the downlink data have a certain correlation at the same time, the current frequency offset condition can be predicted by including the frequency offset estimation data corresponding to the uplink data and the frequency offset estimation data corresponding to the downlink data to obtain the target frequency offset data.

[0075] In a feasible implementation, the historical frequency offset data can be processed by using a target frequency offset prediction model to obtain the target frequency offset data.

[0076] The target frequency offset prediction model is a pre-trained model used to predict the frequency offset data existing between the base station and the terminal. The target frequency offset data is a frequency offset value obtained by processing the historical frequency offset data by using the target frequency offset prediction model.

[0077] In some embodiments, the target frequency offset prediction model can be trained by using the to-be-trained frequency offset data and the confidence corresponding to the to-be-trained frequency offset data. In this way, in the process of training the target frequency offset prediction model, not only the frequency offset data under different time points, different scenarios, and the like is considered, but also the confidence of the frequency offset data is considered, so that the accuracy of the trained target frequency offset prediction model is higher, and thus the predicted target frequency offset data is more accurate and more suitable for the current data transmission scenario.

[0078] Step 103, performing frequency offset compensation on the signal transmitted between the terminal based on the target frequency offset data.

[0079] In the embodiments of the present application, the signal transmitted by the base station to the terminal can be compensated for frequency offset according to the target frequency offset data, and the signal transmitted by the terminal to the base station can also be compensated for frequency offset according to the target frequency offset data.

[0080] In the embodiments of the present application, whether there is uplink data at the current time point or not, the target frequency offset data can be obtained by predicting the current frequency offset condition according to the historical frequency offset data; and the historical frequency offset data represents the frequency offset condition existing when the terminal transmits data, that is, the historical frequency offset data is the frequency offset condition existing when the base station and the terminal transmit data at the historical time point. Therefore, when predicting the current frequency offset condition, the frequency offset condition generated at the historical time point is comprehensively considered, and the predicted target frequency offset data is more accurate and more suitable for the current data transmission scenario, thereby improving the downlink demodulation performance of the terminal and solving the problem of poor compensation effect and reduced downlink demodulation performance of the terminal in the related art.

[0081] Based on the foregoing embodiments, an embodiment of the present application provides a model method applied to an electronic device, which can refer to a server, a notebook computer, a tablet computer, a desktop computer, a smart television, a set-top box, a mobile device, and the like, which is a device with data processing capability. As shown in Figure 2 The method comprises the following steps:

[0082] Step 201, determining to-be-trained frequency offset data under different scenarios and confidence corresponding to the to-be-trained frequency offset data.

[0083] The to-be-trained frequency offset data is used for training the target frequency offset prediction model; and the to-be-trained frequency offset data can include a plurality of frequency offset data. In a feasible implementation manner, the to-be-trained frequency offset data can include frequency offset data of the terminal under different scenarios; the different scenarios can include but are not limited to a viaduct, an open land, a mountainous area, and a plain, and the like. Further, the to-be-trained frequency offset data can also include frequency offset data corresponding to different running parameters of the terminal under different scenarios.

[0084] In implementation, the to-be-trained frequency offset data can be determined from log data generated by the terminal and the base station when transmitting data at a historical time point.

[0085] At step 202, the initial frequency offset prediction model is trained based on the to-be-trained frequency offset data and the confidence to obtain a target frequency offset prediction model.

[0086] In some embodiments, the initial frequency offset prediction model can be constructed according to a recurrent neural network (RNN). Specifically, the initial frequency offset prediction model can include a neural network composed of an input layer, a hidden layer and an output layer. As shown in FIG. 1, the neural network in FIG. 1 includes an input layer, a hidden layer and an output layer; X1 is the input at the first time point, Y1 is the output at the first time point; X2 is the input at the second time point, Y2 is the output at the second time point; X3 is the input at the third time point, Y3 is the output at the third time point; the state memory (M1) obtained by the hidden layer at the first time point is stored to assist the learning of the neural network at the second time point, and the state memory (M2) obtained by the hidden layer at the second time point is also stored to assist the learning of the neural network at the third time point. In this way, by considering the state memory learned at the previous time point when processing the input data at the current time point, the accuracy of the model can be improved. Figure 3 Figure 3 Figure 3 Figure 3

[0087] In other embodiments, to further improve the performance of the model, a long short-term memory neural network can be used to construct the initial frequency offset prediction model. Further, an improved long short-term memory neural network can also be used to construct the initial frequency offset prediction model. As shown in FIG. 2, in addition to the basic input gate, forget gate and output gate, a confidence module is added; by the confidence module, different weights are given to the frequency offset values in X n Figure 4 n

[0088] Figure 4 ​​​​​The operation of the initial frequency offset prediction model in the method for processing any to-be-trained frequency offset data can be: determining, by the confidence module, a confidence of the to-be-trained frequency offset data; inputting the to-be-trained frequency offset data and the confidence of the to-be-trained frequency offset data to the input gate for processing to obtain processed data; after the memory unit stores the processed data, the forgetting gate determines whether to discard some data according to newly stored data and previously stored data; and then the output gate outputs the processed data to obtain a prediction result.

[0089] In implementation, the frequency offset estimation sequence used initially when training the target frequency offset prediction model can be obtained by using a traditional algorithm, for example, can be obtained by using the method for estimating frequency offset based on uplink demodulation reference signals in the foregoing related technologies, or can be obtained by using the method for estimating frequency offset based on phase difference of reference signals in the foregoing related technologies.

[0090] It should be noted that when the frequency offset estimation sequence obtained by using a traditional algorithm is used for model training, the frequency offset value output during the training process can also be verified. If the output frequency offset value can correctly decode the data, it is added to the frequency offset estimation sequence as a new training sample for continuing model training. In this way, the training sample is continuously expanded, and the model performance is improved based on the expanded training sample.

[0091] In a feasible implementation manner, the length of the frequency offset value sequence in the to-be-trained frequency offset data is N+1, and the frequency offset value sequence composed of the first N frequency offset values can be used as the input of the model, and the (N+1)th frequency offset value can be used as the prediction value of the model.

[0092] In a feasible implementation manner, the confidence of the to-be-trained frequency offset data can be set according to the decoding condition of the frequency offset value in the to-be-trained frequency offset data. Specifically, the confidence can be set according to the number of frequency offset values in the to-be-trained frequency offset data and the decoding condition of each frequency offset value. For example, if the total number of frequency offset values in the to-be-trained frequency offset data is a, and the number of frequency offset values in the to-be-trained frequency offset data whose data result is decoding success is b, then the confidence can be b / a.

[0093] In another feasible implementation manner, the confidence of the to-be-trained frequency offset data can also be set according to the source of the frequency offset value in the to-be-trained frequency offset data. Specifically, since the frequency offset value obtained by using a traditional algorithm is relatively accurate, the frequency offset value obtained by using a traditional algorithm can be set to be relatively high. Since the frequency offset value predicted by the model is not completely accurate, in the case where the frequency offset value predicted by the model cannot be verified, the confidence of the frequency offset value predicted by the model can be set to be relatively low, and in the case where the frequency offset value predicted by the model can be verified, the confidence of the frequency offset value predicted by the model can be set to be relatively high.

[0094] In some embodiments, step 201 can be implemented through step 201a to step 201b:

[0095] Step 201a, obtaining initial historical record data under different scenarios.

[0096] The initial historical record data can be initial log data of the terminal under different scenarios.

[0097] In a feasible implementation, log data generated when the terminal and the base station transmit data at a historical time point can be obtained from multiple network management systems, and the obtained log data is determined as the initial historical record data.

[0098] Exemplarily, as shown in Figure 5 , the network management system in Figure 5 may include multiple network management systems 1 to m, and each network management system can manage multiple base stations 1 to n. Specifically, multiple base stations under typical high-speed scenarios such as mountainous area scenarios, viaduct scenarios, open land scenarios, and plain scenarios can be selected from the network management system through a screening module 1, and base station logs of the multiple base stations are obtained as initial historical record data. The screening module 1 mainly performs base station selection; in a feasible implementation, the screening module 1 can be a script corresponding to screening logic for screening historical log data under different scenarios.

[0099] Step 201b, screening the initial historical record data based on the running parameters corresponding to the terminal to obtain target historical record data.

[0100] The target historical record data is historical record data selected from the initial historical record data; in other words, the target historical record data can be target log data selected from the initial log data.

[0101] In a feasible implementation, as shown in Figure 3 , the initial historical record data can be screened based on the running parameters of the terminal through a screening module 2 to obtain target historical record data. Further, the initial historical record data can be screened based on the running speed of the terminal and the decoding situation of the frequency offset value in the initial historical record data through the screening module 2 to obtain the target historical record data; in this way, it is convenient to construct a data set for training a target frequency offset prediction model based on the selected target historical record data. The screening module 2 mainly performs time period selection, which can include but is not limited to time periods of running speed changes, time periods between different time points, etc.; in a feasible implementation, the screening module 2 can be a script corresponding to screening logic for screening log data corresponding to different running parameters.

[0102] Step 201c, determining the to-be-trained frequency offset data based on the frequency offset values in the target historical record data.

[0103] In a feasible implementation, after obtaining the target historical record data in which the downlink decoding is correct for consecutive N+1 frames, the frequency offset value sequence in the target historical record data can be saved to obtain the to-be-trained frequency offset data.

[0104] In some embodiments, step 201b can be implemented through steps A1 to A2.

[0105] Step A1, filtering, from the initial historical record data, intermediate historical record data generated by the terminal when transmitting data at different running speeds.

[0106] The intermediate historical record data is historical record data generated by the terminal at different running speeds and filtered from the initial historical record data; in other words, the intermediate historical record data is base station logs generated by the terminal at different running speeds and filtered from the initial log data.

[0107] In a feasible implementation, log data generated by the terminal when the train speed is between 250 km / h and 350 km / h can be selected from the initial historical record data as the intermediate historical record data; further, log data generated by the terminal when the train is in the acceleration phase, deceleration phase, constant speed phase, etc. can also be selected as the intermediate historical record data; the specific setting can be made according to actual business requirements, which is not limited in the embodiments of the present application.

[0108] Step A2, filtering, from the intermediate historical record data, target historical record data in which the data decoding is correct for consecutive N data frames.

[0109] The base station logs can include the scheduler logs; the scheduler logs can include the results of uplink decoding and downlink decoding.

[0110] In a feasible implementation, the scheduler logs can be selected from the base station logs (intermediate historical record data), and log data in which the downlink decoding is correct for consecutive N+1 frames can be filtered from the scheduler logs to obtain the target historical record data.

[0111] In the embodiments of the present application, the frequency offset values in the historical record data of different running parameters in different scenarios are used to train the target frequency offset prediction model, which can enable the target frequency offset prediction model to learn the frequency offset conditions of different running parameters in different scenarios to more accurately predict the frequency offset conditions at the current time point. Moreover, the decoding conditions of each frequency offset value are also added when training the target frequency offset prediction model, which further improves the model performance and model accuracy of the target frequency offset prediction model.

[0112] Based on the foregoing embodiments, after the target frequency offset prediction model is trained, the target frequency offset prediction model can be deployed at the base station for use when the base station communicates with the terminal. The specific use manner can be referred to Figure 6 The communication method is shown as in Figure 6 The method includes the following steps:

[0113] Step 601, in a case where data transmitted by the terminal is not received in a target time period but data needs to be sent to the terminal, historical frequency offset data is determined.

[0114] The target time period is determined based on a period of a time slot duration of a carrier scheduled when the terminal transmits data and a scheduling number threshold.

[0115] In the embodiments of the present application, the target time period is a trigger condition for predicting the target frequency offset data by using the target frequency offset prediction model. In a feasible implementation manner, the period of the time slot duration of the carrier scheduled when the terminal transmits data and the scheduling number threshold can be multiplied to obtain the target time period. The calculation formula of the target time period can be: target time period = scheduling period * scheduling number threshold; wherein the scheduling period refers to the period of the time slot duration. For example, in the 5G NR, the time slot duration in the central processor (CP) parameter set in the conventional wireless communication device is as follows: 1 ms, 0.5 ms, 0.25 ms, 0.125 ms, 0.0625 ms. Correspondingly, the scheduling period is equal to the period of the time slot duration, and the scheduling number threshold can be adjusted, and the default value is 64. For example, when the period of the time slot duration is 0.5 ms, the target time period is 32 ms.

[0116] In a feasible implementation manner, in a case where data transmitted by the terminal is not received in the target time period but data needs to be sent to the terminal, the historical frequency offset data is obtained to process the historical frequency offset data by using the target frequency offset prediction model to obtain the predicted target frequency offset data.

[0117] Step 602, the historical frequency offset data is processed by using the target frequency offset prediction model to obtain the target frequency offset data.

[0118] In a feasible implementation manner, the historical frequency offset data can be input into the target frequency offset prediction model for prediction to obtain the target frequency offset data.

[0119] Step 603, the signal transmitted between the terminal and the base station is frequency offset compensated based on the target frequency offset data.

[0120] Step 604, the target frequency offset data is updated into the frequency offset estimation data sequence.

[0121] In the embodiment of the present application, after the target frequency offset data is obtained, the target frequency offset data can be updated into the frequency offset estimation data sequence, so that the frequency offset estimation data in the frequency offset estimation data sequence is always in the latest state, and then the historical frequency offset data determined based on the frequency offset estimation data sequence is more consistent with the current scene, and then the target frequency offset data obtained by processing the historical frequency offset data by using the target frequency offset prediction model is more accurate, thereby improving the downlink demodulation performance of the terminal.

[0122] In some embodiments, the "determining historical frequency offset data" in step 601 can be implemented by the following steps 601a to 601b:

[0123] Step 601a, circularly storing a frequency offset estimation data sequence with a target length, and the frequency offset estimation data sequence includes a target number of frequency offset estimation data.

[0124] The target length is the total length of the circularly stored frequency offset estimation data sequence; the target length can be pre-set, and the target length can be set according to actual business requirements, which is not limited in the embodiment of the present application. The target number is the number of frequency offset estimation data included in each frequency offset estimation data sequence; the target number can be pre-set, and the target number can be set according to actual business requirements, which is not limited in the embodiment of the present application.

[0125] In a feasible implementation manner, the corresponding frequency offset data can be circularly stored according to the time when the base station receives the uplink data, to obtain a frequency offset estimation data sequence with a target length; wherein the frequency offset estimation data sequence includes frequency offset data obtained by a traditional algorithm and model-predicted frequency offset data.

[0126] Step 601b, determining historical frequency offset data based on the frequency offset estimation data sequence with the target length.

[0127] The historical frequency offset data represents the frequency offset existing when data is transmitted between the terminal and the base station.

[0128] In the embodiment of the present application, the frequency offset estimation data sequence with the target length can be divided according to the target number, to obtain at least one frequency offset estimation data sequence, and then the historical frequency offset data is selected from the at least one frequency offset estimation data sequence; wherein the number of frequency offset estimation data included in each frequency offset estimation data sequence is the target number.

[0129] In some embodiments, the operation of determining whether each frequency offset estimation data sequence is historical frequency offset data can be implemented by the following steps B1 to B2:

[0130] Step B1, determining the sequence accuracy rate of the frequency offset estimation data sequence.

[0131] The frequency offset estimation data includes a frequency offset value and a corresponding data check identifier. The data check identifier is used to represent whether the data is correctly decoded under the frequency offset value in the corresponding frequency offset estimation data.

[0132] In the embodiments of the present application, the sequence correctness rate represents a case that the data is correctly decoded under the frequency offset value in the sequence of frequency offset estimation data.

[0133] The implementation manner of step B1 can be: determining the sequence correctness rate based on the data check identifier and the number of frequency offset values in the sequence of frequency offset estimation.

[0134] In a feasible implementation manner, the number of frequency offset values that are correctly decoded in the sequence of frequency offset estimation data can be determined based on the data check identifier in the sequence of frequency offset estimation data; and the sequence correctness rate can be determined according to the number of frequency offset values that are correctly decoded and the total number of frequency offset values included in the sequence of frequency offset estimation data. Specifically, the sequence correctness rate can be obtained by performing a division operation on the number of frequency offset values that are correctly decoded and the total number of frequency offset values included in the sequence of frequency offset estimation data.

[0135] Step B2: in a case that the sequence correctness rate meets the correctness rate threshold requirement, the sequence of frequency offset estimation data is taken as historical frequency offset data.

[0136] In the embodiments of the present application, the correctness rate threshold requirement is a basis for judging whether the sequence of frequency offset estimation data is historical frequency offset data; and the correctness rate threshold requirement can be set in advance, and can be set according to actual business requirements, for example, the correctness rate threshold requirement can be set to 0.8. In a feasible implementation manner, in a case that the sequence correctness rate is greater than or equal to the correctness rate threshold requirement, the sequence of frequency offset estimation data is taken as historical frequency offset data.

[0137] Based on the foregoing embodiments, in other embodiments of the present application, the communication method can further include the following steps:

[0138] Step 605: in a case that the sequence correctness rate does not meet the correctness rate threshold requirement, frequency offset data used in the latest frequency offset compensation for the signal transmitted between the terminal is taken as target frequency offset data, and the signal transmitted between the terminal is compensated for frequency offset based on the target frequency offset data.

[0139] In the embodiments of the present application, in a case that the sequence correctness rate of the sequence of frequency offset estimation data is less than the correctness rate threshold requirement, the frequency offset data used in the latest frequency offset compensation for the signal transmitted between the terminal can be directly taken as target frequency offset data, and the signal transmitted between the terminal is compensated for frequency offset based on the target frequency offset data.

[0140] It should be noted that the communication method provided by the embodiment of the present application can predict the current frequency offset situation by using the target frequency offset pre-stored model to obtain target frequency offset data, regardless of whether there is uplink data and downlink data in the current time period. In the case of uplink data, the traditional algorithm can be used to calculate the frequency offset value, which is more efficient. In the case of no uplink data but downlink data, the target frequency offset pre-stored model can be used to predict the current frequency offset situation to obtain target frequency offset data.

[0141] The communication method provided by the embodiment of the present application can be applied to multiple different high-speed scenes such as high-speed railway scenes and flight scenes. However, the environment and frequency offset situation involved in different high-speed scenes are different, so the construction method of the data set needs to be adaptively adjusted according to the actual application scene to make the constructed data set more consistent with the actual application scene.

[0142] It should be noted that the description of the same steps and the same content in the embodiments can refer to the description in other embodiments, and will not be repeated here.

[0143] The communication method provided by the embodiment of the present application can process the historical frequency offset data by using the target frequency offset prediction model to predict the frequency offset data (target frequency offset data) at the current time point, regardless of whether there is uplink data at the current time point. Moreover, the historical frequency offset data represents the frequency offset situation existing between the terminal and the terminal when transmitting data, that is, the historical frequency offset data is the frequency offset situation existing between the base station and the terminal when transmitting data at the historical time point. Therefore, by using the target frequency offset prediction model to process the historical frequency offset data to predict the frequency offset data at the current time point, the predicted target frequency offset data is more accurate and more consistent with the current data transmission scene, thereby improving the downlink demodulation performance of the terminal and solving the problem of poor compensation effect and reduced downlink demodulation performance of the terminal in the related art.

[0144] The application of the communication method provided by the embodiment of the present application in the actual scene will be described below, taking multiple terminals in a train scene as an example.

[0145] Figure 7 A flowchart of a communication method provided by the embodiment of the present application is shown in FIG. 7. Figure 7 As shown in FIG. 7, the method includes the following steps 701 to 705:

[0146] Step 701, storing the frequency offset estimation data sequence of the target length in a loop according to the terminal.

[0147] In one feasible implementation, a frequency offset estimation data sequence of the target length can be stored cyclically according to the terminal identifier. Storing the frequency offset estimation data sequence of the target length cyclically according to the terminal allows for efficient and unified compensation for multiple terminals in the same environment, thereby improving processing efficiency.

[0148] During normal base station operation, when uplink data is available, the frequency offset estimation data sequence of the target length is stored cyclically by the terminal. For example... Figure 8 As shown, Figure 8 The frequency offset estimation data sequence 1 includes a frequency offset value sequence consisting of n frequency offset values, such as frequency offset value 1, frequency offset value 2, ..., frequency offset value n, and a data check indicator bit sequence (data check identifier) ​​consisting of the corresponding data check indicator bits; the sequence 2 includes a frequency offset value sequence consisting of n frequency offset values, such as frequency offset value 2, frequency offset value 3, ..., frequency offset value n+1, and a data check indicator bit sequence consisting of the corresponding data check indicator bits; the data check code identifier is used to characterize whether the data decoding is correct under the corresponding frequency offset value; if the downlink data decoding is correct (i.e., ACK), the corresponding bit is 1; if the downlink data decoding is incorrect (i.e., NACK), the corresponding bit is 0. In implementation, n frequency offset values ​​are first stored to obtain frequency offset estimation data sequence 1. When a new frequency offset value exists, the new frequency offset value is added to the frequency offset estimation data sequence as the (n+1)th frequency offset value. Based on the n frequency offset values, such as frequency offset value 2, frequency offset value 3, ..., frequency offset value n+1, a new frequency offset estimation data sequence (i.e., frequency offset estimation data sequence 2) is formed. In this way, the frequency offset estimation data sequence is continuously updated based on the new frequency offset value. The frequency offset estimation data sequence of length n is stored cyclically to ensure that the n frequency offset values ​​in the frequency offset estimation data sequence are up-to-date.

[0149] It should be noted that during the circular storage process, if the frequency offset value is obtained using a traditional algorithm when there is upstream data and a model prediction is used when there is no upstream data, then the frequency offset value in the frequency offset estimation data sequence may fall into one of the following three categories: 1. When upstream data is consistently present over a long period, the frequency offset values ​​in the frequency offset estimation data sequence are all obtained using the traditional algorithm; 2. When upstream data is present for a period and absent for a period, the frequency offset values ​​in the frequency offset estimation data sequence are partly obtained from the traditional algorithm and partly from the model prediction; 3. When there is no upstream data for a long period, the frequency offset values ​​in the frequency offset estimation data sequence are all obtained using the model prediction.

[0150] Step 702, judge whether the time when the terminal exists downlink scheduling and does not exist uplink scheduling exceeds the target time period, and whether the sequence accuracy rate of the frequency offset estimation data sequence meets the accuracy threshold requirement; if the time when the terminal exists downlink scheduling and does not exist uplink scheduling exceeds the target time period, and the sequence accuracy rate of the frequency offset estimation data sequence meets the accuracy threshold requirement, execute step 703.

[0151] In the embodiment of the application, as shown in Figure 9 The implementation manner of determining the target time period can be: step 901, obtaining the period of the time slot duration configured by the base station; step 902, obtaining the scheduling times threshold; step 903, performing product operation on the period of the time slot duration and the scheduling times threshold to obtain the target time period.

[0152] The determination module judges whether the terminal meets condition 1 (i.e. there is downlink scheduling while there is no uplink service channel scheduling in the target time period) and condition 2 (the data check accuracy rate of the terminal's frequency offset estimation value sequence meets the accuracy threshold requirement) in each scheduling period; if both condition 1 and condition 2 are met, execute step 703; if both condition 1 and condition 2 are not met, execute step 706.

[0153] Step 703, processing the frequency offset estimation data sequence of each terminal by using the target frequency offset prediction model to obtain target frequency offset data.

[0154] In a feasible implementation manner, the network architecture of the target frequency offset prediction model can be the network architecture as shown in Figure 5 to ensure the model performance of the target frequency offset prediction model, and further ensure the accuracy rate of the target frequency offset data obtained by prediction.

[0155] Step 704, update the target frequency offset data into the frequency offset estimation data sequence according to the terminal.

[0156] In a feasible implementation manner, the target frequency offset data of each terminal can be updated into the corresponding frequency offset estimation data sequence according to the terminal identifier.

[0157] Step 705, select a suitable frequency offset value for frequency offset compensation of downlink data by comprehensively considering the target frequency offset data of all terminals.

[0158] In a feasible implementation manner, the mean value of the target frequency offset data of all terminals can be determined, and the downlink data is compensated for frequency offset according to the mean value.

[0159] In other embodiments, after receiving the decoding feedback (ACK / NACK) of the interrupted downlink data, the data check indication bit of the frequency offset value in the frequency offset estimation data sequence can also be updated to ensure that the data check indication bit of the frequency offset estimation data sequence is consistent with the actual situation.

[0160] Step 706, using the frequency offset data used for frequency offset compensation of the signal transmitted between the terminal and the base station last time to perform frequency offset compensation on the signal transmitted between the terminal and the base station this time.

[0161] It should be noted that the base station and the terminal can both perform frequency offset compensation on the transmitted data, and generally the base station side performs large-scale frequency offset compensation; for example, if the frequency offset between the base station and the terminal when transmitting data is 1010, the base station side compensates 1000 and the terminal side compensates 10. The communication method provided in the embodiments of the present application is described by taking the application in the base station side as an example, but the communication method provided in the embodiments of the present application can also be extended to the terminal side; specifically, the historical frequency offset condition when transmitting data with the base station is obtained, the historical frequency offset condition is processed by using a target frequency offset prediction model to obtain predicted frequency offset data, and the pre-stored frequency offset data is used to perform frequency offset compensation on the signal transmitted with the base station.

[0162] It is worth mentioning that the advantages of the communication method provided in the embodiments of the present application at least include: 1. using an improved long short-term memory neural network to replace the traditional algorithm to predict the frequency offset value, which can improve the accuracy of prediction; 2. the frequency offset prediction does not need to rely on real-time uplink data; 3. using the predicted frequency offset value to perform relatively accurate compensation on the downlink, which improves the downlink demodulation accuracy.

[0163] Based on the foregoing embodiments, the embodiments of the present application provide a communication device, which can be applied to Figure 1 and 6 In the communication method provided in the corresponding embodiments, the reference Figure 10 As shown in the figure, the communication device includes:

[0164] The determining unit 1001 is configured to determine historical frequency offset data; wherein the historical frequency offset data represents the frequency offset condition when transmitting data with the terminal;

[0165] The processing unit 1002 is configured to predict the current frequency offset condition according to the historical frequency offset data to obtain target frequency offset data;

[0166] The processing unit 1002 is further configured to perform frequency offset compensation on the signal transmitted with the terminal based on the target frequency offset data.

[0167] In the embodiments of the present application, the determining unit 1001 is further configured to perform the following steps:

[0168] The frequency offset estimation data sequence of the target length is circularly stored, and the frequency offset estimation data sequence includes the target number of frequency offset estimation data;

[0169] The frequency offset estimation data sequence based on the target length is used to determine the historical frequency offset data.

[0170] In the embodiment of the present application, the determining unit 1001 is further configured to determine the historical frequency offset data in the case that no data transmitted by the terminal is received in a target time period but data needs to be transmitted to the terminal, wherein the target time period is determined based on the time slot duration of the carrier scheduled when the terminal transmits data and a threshold of the scheduling times.

[0171] Correspondingly, the processing unit 1002 is further configured to update the target frequency offset data into the frequency offset estimation data sequence.

[0172] In the embodiment of the present application, the processing unit 1002 is further configured to perform the following steps:

[0173] determine the sequence correctness rate of the frequency offset estimation data sequence;

[0174] in the case that the sequence correctness rate meets the correctness rate threshold requirement, use the frequency offset estimation data sequence as the historical frequency offset data.

[0175] In the embodiment of the present application, the processing unit 1002 is further configured to determine the sequence correctness rate based on the data verification identifier and the number of frequency offset values in the frequency offset estimation sequence.

[0176] In the embodiment of the present application, the processing unit 1002 is further configured to, in the case that the sequence correctness rate does not meet the correctness rate threshold requirement, use the frequency offset data used for frequency offset compensation of the signal transmitted between the terminal in the last time as the target frequency offset data, and perform frequency offset compensation on the signal transmitted between the terminal based on the target frequency offset data.

[0177] In the embodiment of the present application, the processing unit 1002 is further configured to process the historical frequency offset data based on the target frequency offset prediction model to predict the current frequency offset condition.

[0178] The training method of the target frequency offset prediction model comprises:

[0179] determining the to-be-trained frequency offset data in different scenarios and the confidence degrees corresponding to the to-be-trained frequency offset data;

[0180] performing model training on the initial frequency offset prediction model based on the to-be-trained frequency offset data and the confidence degrees to obtain the target frequency offset prediction model.

[0181] In the embodiment of the present application, the processing unit 1002 is further configured to perform the following steps:

[0182] obtain initial historical record data in different scenarios;

[0183] filter the initial historical record data based on the running parameters corresponding to the terminal to obtain target historical record data.

[0184] Based on the frequency offset values ​​in the target historical data, determine the frequency offset data to be trained.

[0185] In this embodiment of the application, the processing unit 1002 is further configured to perform the following steps:

[0186] From the initial historical data, intermediate historical data generated when the terminal transmits data at different operating speeds are filtered out;

[0187] From the intermediate historical data, select the target historical data in which the data decoding of N consecutive data frames is correct.

[0188] It should be noted that the specific implementation process of the steps performed by each unit in this embodiment can be referred to Figure 1 and 6 The implementation process of the communication method provided in the corresponding embodiment will not be described in detail here.

[0189] The communication device provided in the embodiments of this application can predict the frequency offset data (target frequency offset data) at the current time point by processing historical frequency offset data through a target frequency offset prediction model, regardless of whether there is uplink data at the current time point. Furthermore, historical frequency offset data represents the frequency offset situation that exists when transmitting data between the base station and the terminal. That is, historical frequency offset data is the frequency offset situation that exists when the base station and the terminal transmit data at historical time points. Therefore, by using the target frequency offset prediction model to process historical frequency offset data to predict the frequency offset data at the current time point, the frequency offset situation generated at historical time points is comprehensively considered during the prediction. The predicted target frequency offset data is more accurate and more in line with the current data transmission scenario, thereby improving the downlink demodulation performance of the terminal and solving the problem of poor compensation effect and reduced downlink demodulation performance of the terminal in related technologies.

[0190] Based on the foregoing embodiments, embodiments of this application provide a base station that can be applied to... Figure 1 and 6 In the communication method provided in the corresponding embodiment, refer to Figure 11 As shown, the device may include: a processor 111, a memory 112, and a communication bus 113;

[0191] Determine historical frequency offset data; whereby historical frequency offset data characterizes the frequency offset that exists when transmitting signals between the terminal;

[0192] The target frequency offset data is obtained by predicting the current frequency offset based on historical frequency offset data.

[0193] Frequency offset compensation is performed on the signal transmitted between the target frequency offset data and the terminal.

[0194] In other embodiments of the present application, the processor 111 is configured to execute the communication program in the memory 112 to determine the historical frequency offset data to implement the following steps:

[0195] cyclically storing a frequency offset estimation data sequence of a target length, the frequency offset estimation data sequence including a target number of frequency offset estimation data;

[0196] determining the historical frequency offset data based on the frequency offset estimation data sequence of the target length.

[0197] In other embodiments of the present application, the processor 111 is configured to execute the communication program in the memory 112 to determine the historical frequency offset data to implement the following steps:

[0198] In a case where no data transmitted by the terminal is received in a target time period but data needs to be sent to the terminal, the historical frequency offset data is determined; the target time period is determined based on a period of a time slot duration of a carrier scheduled when the terminal transmits data and a threshold of scheduling times;

[0199] Correspondingly, the processor 111 is configured to execute the communication program in the memory 112 to process the historical frequency offset data by using the target frequency offset prediction model to obtain the target frequency offset data, and the following steps can also be implemented:

[0200] updating the target frequency offset data into the frequency offset estimation data sequence.

[0201] In other embodiments of the present application, the processor 111 is configured to execute the communication program in the memory 112 to determine the historical frequency offset data based on the frequency offset estimation data sequence of the target length to implement the following steps:

[0202] determining a sequence accuracy rate of the frequency offset estimation data sequence;

[0203] in a case where the sequence accuracy rate meets a correct rate threshold requirement, taking the frequency offset estimation data sequence as the historical frequency offset data.

[0204] In other embodiments of the present application, the frequency offset estimation data includes a frequency offset value and a corresponding data check identifier; the data check identifier is used to represent whether the data is correctly decoded under the corresponding frequency offset value; the processor 111 is configured to execute the communication program in the memory 112 to determine the sequence accuracy rate of the frequency offset estimation data sequence to implement the following steps:

[0205] determining the sequence accuracy rate based on the data check identifier and a number of the frequency offset values in the frequency offset estimation sequence.

[0206] In other embodiments of the present application, the processor 111 is configured to execute the communication program in the memory 112 to implement the following steps:

[0207] In a case where the sequence accuracy does not satisfy the accuracy threshold requirement, using frequency offset data used for frequency offset compensation of the signal transmitted between the terminal last time as target frequency offset data, and performing frequency offset compensation on the signal transmitted between the terminal based on the target frequency offset data.

[0208] In other embodiments of the present application, the processor 111 is configured to execute the communication program in the memory 112 to predict the current frequency offset condition according to the historical frequency offset data, and the following steps can be implemented:

[0209] processing the historical frequency offset data based on the target frequency offset prediction model to predict the current frequency offset condition;

[0210] In other embodiments of the present application, the processor 111 is configured to execute the communication program in the memory 112 to further implement the following steps:

[0211] determining the to-be-trained frequency offset data and the confidence degree corresponding to the to-be-trained frequency offset data in different scenarios;

[0212] training the initial frequency offset prediction model based on the to-be-trained frequency offset data and the confidence degree to obtain the target frequency offset prediction model.

[0213] In other embodiments of the present application, the processor 111 is configured to execute the communication program in the memory 112 to determine the to-be-trained frequency offset data in different scenarios, and the following steps can be implemented:

[0214] obtaining initial historical record data in different scenarios;

[0215] filtering the initial historical record data based on the running parameters corresponding to the terminal to obtain target historical record data;

[0216] determining the to-be-trained frequency offset data based on the frequency offset values in the target historical record data.

[0217] In other embodiments of the present application, the processor 111 is configured to execute the communication program in the memory 112 to filter the initial historical record data based on the running parameters corresponding to the terminal to obtain target historical record data, and the following steps can be implemented:

[0218] from the initial historical record data, filtering intermediate historical record data generated when the terminal transmits data at different running speeds;

[0219] from the intermediate historical record data, filtering target historical record data in which data decoding of N consecutive data frames is correct.

[0220] It should be noted that the specific description of the steps performed by the processor can be referred to Figure 1 and 6The communication method provided by the corresponding embodiments.

[0221] The base station provided by the embodiments of the present application can process the historical frequency offset data by using the target frequency offset prediction model to predict the frequency offset data (target frequency offset data) at the current time point, regardless of whether there is uplink data at the current time point. The historical frequency offset data represents the frequency offset that exists when data is transmitted between the base station and the terminal, that is, the historical frequency offset data is the frequency offset that exists when the base station and the terminal transmit data at the historical time point. Therefore, when the target frequency offset prediction model is used to process the historical frequency offset data to predict the frequency offset data at the current time point, the frequency offset at the historical time point is comprehensively considered during the prediction. The predicted target frequency offset data is more accurate and more suitable for the current data transmission scenario, thereby improving the downlink demodulation performance of the terminal and solving the problem of poor compensation effect and reduced downlink demodulation performance of the terminal in the related art.

[0222] Based on the foregoing embodiments, the embodiments of the present application provide a computer-readable storage medium storing one or more programs, which can be executed by one or more processors to implement Figure 1 and 6 The steps of the communication method provided by the corresponding embodiments.

[0223] It should be noted that the computer-readable storage medium described above can be a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), an Electrically Erasable Programmable Read-Only Memory (EEPROM), a Ferromagnetic Random Access Memory (FRAM), a Flash Memory, a magnetic surface memory, an optical disc, or a Compact Disc Read-Only Memory (CD-ROM) memory, etc. It can also be various electronic devices including one or any combination of the above memories, such as a mobile phone, a computer, a tablet device, a personal digital assistant, etc.

[0224] It should be noted that, in the present document, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises a", "comprising", or "comprises" does not, without more constraints, preclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0225] The above-mentioned sequence numbers of the embodiments of the present application are only for description, and do not represent advantages or disadvantages of the embodiments.

[0226] Those skilled in the art can clearly understand the above-mentioned embodiment method by means of software and necessary general hardware platform, of course, it can also be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes a plurality of instructions for causing a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device) to execute the methods described in various embodiments of the present application.

[0227] The present application is described with reference to flowcharts and / or block diagrams according to the methods, devices (systems), and computer program products of the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to produce a machine, so that the instructions executed by the computer or other programmable data processing devices produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks Figure 1 The functions specified in one or more flows and / or blocks

[0228] These computer program instructions can also be stored in a computer-readable storage medium that can guide the computer or other programmable data processing devices to work in a specific manner, so that the instructions stored in the computer-readable storage medium produce a manufactured product including instruction devices that implement the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks Figure 1 The functions specified in one or more flows and / or blocks

[0229] These computer program instructions can also be loaded into a computer or other programmable data processing devices, so that a series of operational steps are performed on the computer or other programmable data processing devices to generate a computer-implemented process, so that the instructions executed on the computer or other programmable devices provide a process for implementing the flow Figure 1 One flow or multiple flows and / or the functions specified in the block ​ One block or multiple blocks.

[0230] The above is only the preferred embodiment of the present application, and does not limit the patent scope of the present application, and any equivalent structure or equivalent flow transformation using the content of the specification and drawings of the present application, or direct or indirect application in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A communication method, wherein, The method includes: If no data is received from the terminal during the target time period but data needs to be sent to the terminal, historical frequency offset data is determined; wherein, the historical frequency offset data represents the frequency offset that exists when transmitting data with the terminal; the target time period is determined based on the period of the time slot duration of the carrier scheduled when the terminal transmits data and the scheduling number threshold. The historical frequency offset data is processed according to the target frequency offset prediction model to predict the current frequency offset situation and obtain the target frequency offset data; the target frequency offset prediction model is a neural network model. Frequency offset compensation is performed on the signal transmitted between the target frequency offset data and the terminal.

2. The method according to claim 1, wherein, The determination of historical frequency offset data includes: A frequency offset estimation data sequence of target length is stored cyclically, and the frequency offset estimation data sequence includes frequency offset estimation data of the target number; Based on the frequency offset estimation data sequence of the target length, the historical frequency offset data is determined.

3. The method according to claim 2, wherein, After processing the historical frequency offset data according to the target frequency offset prediction model to predict the current frequency offset and obtain the target frequency offset data, the process further includes: The target frequency offset data is updated into the frequency offset estimation data sequence.

4. The method according to claim 2, wherein, The determination of the historical frequency offset data based on the frequency offset estimation data sequence of the target length includes: Determine the sequence accuracy of the frequency offset estimation data sequence; If the sequence accuracy meets the accuracy threshold requirement, the frequency offset estimation data sequence is used as the historical frequency offset data.

5. The method according to claim 4, wherein, The frequency offset estimation data includes the frequency offset value and the corresponding data verification identifier; The data verification identifier is used to characterize whether the data decoding is correct under the corresponding frequency offset value; Determining the sequence accuracy of the frequency offset estimation data sequence includes: The sequence accuracy is determined based on the data verification identifier and the number of frequency offset values ​​in the frequency offset estimation data sequence.

6. The method according to claim 4, wherein, The method further includes: If the sequence accuracy does not meet the accuracy threshold requirement, the frequency offset data used in the most recent frequency offset compensation of the signal transmitted with the terminal is used as the target frequency offset data, and frequency offset compensation of the signal transmitted with the terminal is performed based on the target frequency offset data.

7. The method according to claim 1, wherein, The training method for the target frequency offset prediction model includes: Determine the training frequency offset data and the confidence level corresponding to the training frequency offset data in different scenarios; The initial frequency offset prediction model is trained based on the frequency offset data to be trained and the confidence level to obtain the target frequency offset prediction model.

8. The method according to claim 7, wherein, The determination of the training frequency offset data under different scenarios includes: Obtain initial historical data under different scenarios; The initial historical data is filtered based on the operating parameters corresponding to the terminal to obtain the target historical data; The frequency offset data to be trained is determined based on the frequency offset value in the target historical data.

9. The method according to claim 8, wherein, The step of filtering the initial historical data based on the operating parameters corresponding to the terminal to obtain the target historical data includes: From the initial historical data, intermediate historical data generated when the terminal transmits data at different operating speeds are filtered out; From the intermediate historical data, select target historical data in which the data decoding of N consecutive data frames is correct.

10. A base station, wherein, The base station includes: a processor and a memory; The communication bus is used to establish a communication connection between the processor and the memory; The processor is used to execute the communication program stored in the memory to perform the following steps: If no data is received from the terminal during the target time period but data needs to be sent to the terminal, historical frequency offset data is determined; wherein, the historical frequency offset data represents the frequency offset that exists when transmitting data with the terminal; the target time period is determined based on the period of the time slot duration of the carrier scheduled when the terminal transmits data and the scheduling number threshold. The historical frequency offset data is processed according to the target frequency offset prediction model to predict the current frequency offset situation and obtain the target frequency offset data; the target frequency offset prediction model is a neural network model. Frequency offset compensation is performed on the signal transmitted between the target frequency offset data and the terminal.

Images