CN116582919A - Communication method and base station - Google Patents

Abstract

The embodiment of the present application discloses a communication method, the method includes: determining historical frequency offset data; wherein, the historical frequency offset data represents the frequency offset that exists when transmitting data with the terminal; according to the historical frequency offset data Predicting the current frequency offset situation to obtain target frequency offset data; performing frequency offset compensation for signals transmitted with the terminal based on the target frequency offset data. The embodiment of the present application also discloses a base station.

Description

A communication method and base station

technical field

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

Background technique

If the terminal is in a moving state during the communication with the base station, in order to ensure the reliability of data transmission at this time, frequency offset compensation will be performed on the transmitted data. In the prior art, after receiving the uplink data sent by the terminal through the uplink, the base station predicts the estimated frequency offset value of the downlink data sent to the terminal according to the uplink data, and performs frequency offset compensation for the downlink data based on the estimated frequency offset value; but , when there is no uplink data for a long time, the estimated frequency offset will not be updated. At this time, the frequency offset compensation is performed on the current downlink data based on the previous estimated frequency offset. The compensation effect is poor and the downlink demodulation performance of the terminal is reduced. .

Contents of the invention

In order to solve the above technical problems, the embodiment of the present application expects to provide a communication method and a base station, which solve the problems in the related art that the compensation effect is poor and the downlink demodulation performance of the terminal is reduced.

The technical scheme of the present application is realized like this:

A communication method, wherein the method includes:

Determine historical frequency offset data; wherein, the historical frequency offset data represents the frequency offset that exists when transmitting data with the terminal;

Predicting the current frequency offset situation according to the historical frequency offset data to obtain target frequency offset data;

Perform frequency offset compensation on signals transmitted with the terminal based on the target frequency offset data.

In the above solution, the determination of historical frequency offset data includes:

cyclically storing a frequency offset estimation data sequence of a target length, the frequency offset estimation data sequence including a target quantity of frequency offset estimation data;

The historical frequency offset data is determined based on the frequency offset estimation data sequence of the target length.

In the above solution, the determination of the historical frequency offset data includes:

When the data transmitted by the terminal is not received within the target time period but the data needs to be sent to the terminal, determine the historical frequency offset data; the target time period is based on the carrier scheduled when the terminal transmits data Determined by the period of the time slot duration and the threshold of the number of scheduling times;

The method of predicting the current frequency offset situation according to the historical frequency offset data, and after obtaining the target frequency offset data, further includes:

The target frequency offset data is updated into the frequency offset estimation data sequence.

In the above solution, the determination of the historical frequency offset data based on the frequency offset estimation data sequence of the target length includes:

determining the sequence accuracy rate of the frequency offset estimation data sequence;

In the case that the sequence accuracy meets the accuracy threshold requirement, the frequency offset estimation data sequence is used as the historical frequency offset data.

In the above solution, the frequency offset estimation data includes a frequency offset value and a corresponding data check mark; the data check mark is used to indicate whether the data is decoded correctly under the corresponding frequency offset value;

The determining the sequence accuracy rate of the frequency offset estimation data sequence includes:

Determine the sequence accuracy rate based on the data verification identifier and the number of frequency offset values in the frequency offset estimation sequence.

In the above scheme, the method also includes:

In the case that the sequence accuracy rate does not meet the accuracy rate threshold requirement, the frequency offset data used when performing frequency offset compensation on the signal transmitted between the terminal and the terminal last time is used as the target frequency offset data, based on the The target frequency offset data performs frequency offset compensation on signals transmitted between the terminal and the terminal.

In the above scheme, the prediction of the current frequency deviation situation according to the historical frequency deviation data includes:

Processing the historical frequency offset data based on the target frequency offset prediction model to predict the current frequency offset situation;

The training method of the target frequency offset prediction model includes:

Determining the frequency offset data to be trained in different scenarios and the confidence corresponding to the frequency offset data to be trained;

Perform model training on an initial frequency offset prediction model based on the frequency offset data to be trained and the confidence level to obtain the target frequency offset prediction model.

In the above solution, the determination of the frequency offset data to be trained in different scenarios includes:

Obtain initial historical record data in different scenarios;

Screening the initial historical record data based on the operating parameters corresponding to the terminal to obtain target historical record data;

The frequency offset data to be trained is determined based on the frequency offset value in the target historical record data.

In the above solution, the initial historical record data is screened based on the operating parameters corresponding to the terminal to obtain target historical record data, including:

From the initial historical record data, filter out the intermediate historical record data generated when the terminal transmits data at different operating speeds;

From the intermediate historical record data, the target historical record data whose data decoding is correct for N consecutive data frames is screened out.

A base station, the base station comprising: a processor and a memory;

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

The processor is used to execute the communication program stored in the memory, so as to realize the following steps:

Determine historical frequency offset data; wherein, the historical frequency offset data represents the frequency offset that exists when transmitting signals with the terminal;

Predicting the current frequency offset situation according to the historical frequency offset data to obtain target frequency offset data;

Perform frequency offset compensation on signals transmitted with the terminal based on the target frequency offset data.

Description of drawings

FIG. 1 is a schematic flowchart of a communication method according to an embodiment of the present application;

Fig. 2 is a schematic flow chart of a model method provided by the embodiment of the present application;

FIG. 3 is a schematic diagram of a network architecture of an initial frequency domain prediction model in a communication method provided by an embodiment of the present application;

FIG. 4 is a schematic diagram of a network architecture of an initial frequency domain prediction model in another communication method provided by an embodiment of the present application;

FIG. 5 is a schematic flow chart of acquiring target historical record data in a communication method provided by an embodiment of the present application;

FIG. 6 is a schematic flowchart of another communication method provided by the embodiment of the present application;

FIG. 7 is a schematic flowchart of another communication method provided by the embodiment of the present application;

FIG. 8 is a schematic diagram of a frequency offset estimation data sequence in a communication method provided by an embodiment of the present application;

FIG. 9 is a schematic flowchart of determining a target time period in a communication method provided in an embodiment of the present application;

FIG. 10 is a schematic structural diagram of a communication device provided by an embodiment of the present application;

FIG. 11 is a schematic structural diagram of a base station provided by an embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

It should be understood that the specific embodiments described here are only used to explain the present application, and are not intended to limit the present application.

Before further describing the embodiments of the present application in detail, the nouns and terms involved in the embodiments of the present application will be described first, and the nouns and terms involved in the embodiments of the present disclosure are applicable to the following explanations.

(1) Doppler frequency shift

Orthogonal Frequency Division Multiplexing (OFDM) is a long-term evolution (Long Term Evolution, LTE), wireless air interface technology of the fifth generation mobile communication technology (5thGeneration Mobile Communication Technology New Radio, 5G NR) and other communication systems The key technology in this method, this OFDM method has high spectral efficiency, can support high-speed data transmission, and can support various bandwidth flexible configurations and many other advantages. OFDM is very sensitive to phase noise and carrier frequency offset, and carrier frequency offset will destroy the orthogonality between subcarriers.

When the mobile terminal is in motion, especially when communicating at high speed, the frequency of the signal between the mobile terminal and the receiving end of the base station will change, which is called the Doppler effect. The frequency shift caused by the Doppler effect is called the Doppler frequency shift, and its formula is: Among them, θ is the angle between the moving direction of the mobile terminal and the direction of the incident wave, v is the moving speed of the mobile terminal, c is the propagation speed of the electromagnetic wave, and f is the carrier frequency.

(2) Uplink and Downlink

A wireless communication system can simultaneously support communication for multiple wireless communication devices. A wireless communication device may communicate with one or more base stations (also referred to as access points, or nodes) via transmissions on the uplink and downlink. The uplink (or reverse link) refers to the communication link from the wireless communication device (terminal) to the base station, and the downlink (or forward link) refers to the communication link from the base station to the wireless communication device.

(3) Wireless channel

In the process of wireless signal transmission, after reflection, scattering, diffraction, etc., the signal arriving at the receiving side is the superposition of multiple signal components with different delays and phases, and changes in the communication environment will cause the signal at the receiving end to change with time. The time changes, so interference and distortion occur, causing fading of the received signal.

During the movement of the mobile terminal, for example, during the operation of the high-speed rail, the terminal user experiences complex and diverse scenes, such as viaducts, open areas, mountains, stations, etc. The geographical structure and distribution of scatterers in different scenes are different, which will lead to high-speed rail The multipath effect and large-scale fading of wireless channels are different.

The related technologies involved in the embodiments of the present application will be described below.

When the terminal is in motion, especially in high-speed scenarios such as high-speed railway scenarios and flight scenarios, the base station receiving system needs to accurately estimate the frequency offset and compensate the received uplink data in order to ensure system performance; similarly, the downlink When data is sent, frequency offset compensation also needs to be performed to ensure the demodulation performance of the terminal. At present, most methods of frequency offset compensation are based on uplink traffic channel (Physical Uplink Shared Channel, PUSCH) data for frequency offset estimation. , SRS) to perform frequency offset estimation, but the periodicity of SRS determines that it cannot perform frequency offset estimation at every scheduling time, so the current common solution is to perform frequency offset compensation for downlink data based on historical frequency offset estimation values.

Solution 1 in the related art: In the cell corresponding to any remote radio unit (Remote Radio Unit, RRU) of the base station, according to the demodulation reference signal estimation of the uplink between the RRU and each user equipment (terminal) The estimated frequency offset value of each user equipment is obtained and stored, and the frequency offset pre-compensation value is determined based on the estimated frequency offset 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. In a scenario where there is no downlink data for a period of time, the frequency offset estimation value will not be continuously updated. Using the historical value of frequency offset estimation to perform frequency offset compensation on downlink data will cause problems. Reduce the downlink demodulation performance of the terminal.

Solution 2 in the related art: obtain reference signals respectively at consecutive sampling points, extract the cyclic prefix and end data of each reference signal, and calculate the phase difference and all of the reference signals through the cyclic prefix and end data of each reference signal The average phase difference of the reference signals, the frequency offset value of each reference signal is obtained through the average phase difference and the phase difference of each reference signal; the frequency offset compensation is performed on each reference signal through the frequency offset value of each reference signal. The disadvantage of this solution is that it still needs to rely on uplink data to estimate the frequency offset of the current frame, and cannot handle scenarios without uplink data.

To this end, an embodiment of the present application provides a communication method, which can be applied to a base station. Referring to FIG. 1, the method includes the following steps:

Step 101. Determine historical frequency offset data.

Wherein, the historical frequency offset data represents the frequency offset that exists when data is transmitted between the base station and the terminal. The historical frequency offset data may be frequency offset data generated when the base station and the terminal transmit data at historical time points. Optionally, the historical frequency offset data may include at least one piece of frequency offset estimation data, and each frequency offset estimation data may include a frequency offset value and a data check code of the frequency offset value. If the terminal is in motion during signal transmission with the base station, the vibration frequency of the bearer signal received by the base station will be different from the frequency sent by the terminal, and the resulting frequency deviation is the frequency offset. By processing the signal through the frequency offset value, the frequency offset that occurs during the signal transmission process can be eliminated, thereby improving the reliability of communication.

In a feasible implementation, when the base station is in normal operation and there is uplink data, multiple frequency offset estimation data corresponding to the uplink data can be cyclically stored, and the sequence composed of these multiple frequency offset estimation data can be used as a history Frequency offset data. Or, in the process of data interaction between the base station and the terminal, when uplink data and downlink data exist at the same time, both 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 historical frequency Partial data.

Step 102: Predict the current frequency offset situation according to the historical frequency offset data to obtain target frequency offset data.

Wherein, the target frequency offset data is the frequency offset data obtained after predicting the current frequency offset situation according to the historical frequency offset data.

Since there is a certain relationship between the frequency offset of the uplink data and the frequency offset of the downlink data at the same time, the frequency offset estimation data corresponding to the uplink data and the frequency offset estimation data corresponding to the downlink data can be included. The current frequency offset situation is predicted to obtain the target frequency offset data.

In a feasible implementation manner, the target frequency offset prediction model may be used to process the historical frequency offset data to obtain the target frequency offset data.

Wherein, the target frequency offset prediction model is a pre-trained model for predicting frequency offset data existing between the base station and the terminal. The target frequency offset data is a frequency offset value obtained by processing historical frequency offset data using a target frequency offset prediction model.

In some embodiments, the target frequency offset prediction model can be obtained by training the frequency offset data to be trained and the confidence corresponding to the frequency offset data to be trained; thus, in the process of training the target frequency offset prediction model, not only different time The frequency offset data under different situations such as points and different scenarios, and the credibility of the frequency offset data are also considered, which can make the trained target frequency offset prediction model more accurate, and then make the predicted target frequency offset data more accurate. Accurate and more suitable for the current data transmission scenario.

Step 103 , perform frequency offset compensation on the signal transmitted with the terminal based on the target frequency offset data.

In the embodiment of the present application, frequency offset compensation may be performed on signals sent from the base station to the terminal according to the target frequency offset data, and frequency offset compensation may be performed on signals sent from the terminal to the base station according to the target frequency offset data.

In the embodiment of the present application, regardless of whether there is uplink data at the current time point, the current frequency offset situation can be predicted according to the historical frequency offset data to obtain the target frequency offset data; and the historical frequency offset data represents the transmission data between terminals The frequency offset situation that exists at any time, that is, the historical frequency offset data is the frequency offset situation that exists when the base station and the terminal transmit data at historical time points, then the current frequency offset situation is predicted according to the historical frequency offset data, and the comprehensive Considering the frequency offset generated at historical time points, the predicted target frequency offset data is more accurate and more suitable for the current data transmission scenario, which can improve the downlink demodulation performance of the terminal and solve the poor compensation effect in related technologies Moreover, the downlink demodulation performance of the terminal is reduced.

Based on the foregoing embodiments, the embodiments of the present application provide a model method applied to electronic devices, which may refer to servers, notebook computers, tablet computers, desktop computers, smart TVs, set-top boxes, mobile devices, etc. that have data processing capabilities device of. , shown in Fig. 2, this method comprises the following steps:

Step 201. Determine frequency offset data to be trained and confidence levels corresponding to the frequency offset data to be trained in different scenarios.

Wherein, the frequency offset data to be trained is used to train a target frequency offset prediction model; the frequency offset data to be trained may include multiple frequency offset data. In a feasible implementation manner, the frequency offset data to be trained may include frequency offset data of the terminal in different scenarios; different scenarios may include but not limited to: viaducts, open land, mountains, and plains. Further, the frequency offset data to be trained may also include frequency offset data corresponding to different operating parameters of the terminal in different scenarios.

During implementation, the frequency offset data to be trained can be determined from the log data generated when the terminal and the base station transmit data at historical time points.

Step 202: Perform model training on the initial frequency offset prediction model based on the frequency offset data to be trained and the confidence level, to obtain a target frequency offset prediction model.

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

In other embodiments, in order to further improve model performance, a long short-term memory neural network may be used to construct an initial frequency offset prediction model. Furthermore, the improved long-short-term memory neural network can also be used to construct the initial frequency offset prediction model; as shown in Figure 4, in addition to the basic input gate, forgetting gate and output gate, a confidence module is also added; through The confidence module assigns different weights to the frequency offset values in the input X _n to further improve the model performance. The input gate is used to control whether the data is input and processed when there is data input; the memory cell (MemoryCell) is used to store the processed data for the next memory unit to use; the output gate , used to control whether to output the data calculated this time; forget gate, used to control whether to clear the data in the memory unit.

The initial frequency offset prediction model in Figure 4 can process any frequency offset data to be trained as follows: determine the confidence of the frequency offset data to be trained through the confidence module; combine the frequency offset data to be trained with the frequency offset data to be trained The confidence of the training frequency offset data is input to the input gate for processing to obtain the processed data; after the memory unit stores the processed data, the forget gate judges whether it needs to be discarded based on the newly stored data and the previously stored data. Some data; then, the processed data is output by the output gate to obtain the prediction result.

During implementation, the frequency offset estimation sequence initially used when training the target frequency offset prediction model can be obtained using traditional algorithms, for example, the frequency offset estimation method based on the uplink demodulation reference signal in the aforementioned related art can be used to obtain Alternatively, the frequency offset estimation method based on the phase difference of the reference signal in the aforementioned related art may also be used.

It should be noted that when the frequency offset estimation sequence obtained by the 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 decode the data correctly, then the It is added to the frequency offset estimation sequence as a new training sample to continue model training; in this way, the training sample is continuously expanded, and then the performance of the model is improved based on the expanded training sample.

In a feasible implementation, the length of the frequency offset value sequence in the frequency offset data to be trained 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+th One frequency offset value is used as the predicted value of the model.

In a feasible implementation, the confidence level of the frequency offset data to be trained can be set according to the decoding situation of the frequency offset value in the frequency offset data to be trained; specifically, it can be set according to the frequency offset value in the frequency offset data to be trained The number of offset values and the decoding situation of each frequency offset value are set; for example, if the total number of frequency offset values in the frequency offset data to be trained is a, the result of the data in the frequency offset data to be trained is that the decoding is successful The number of frequency offset values is b, then the confidence level can be b/a.

In another feasible implementation, the confidence level of the frequency offset data to be trained can also be set according to the source of the frequency offset value in the frequency offset data to be trained; specifically, since the frequency offset value obtained by the traditional algorithm is relatively accurate, therefore The frequency offset value obtained by the traditional algorithm can be set higher; since the frequency offset value predicted by the model is not completely accurate, if it cannot be verified, the confidence of the frequency offset value predicted by the model can be set lower , and in the case of verification, the confidence level of the frequency offset value predicted by the model can be set higher.

In some embodiments, step 201 may be implemented through steps 201a to 201b:

Step 201a, acquiring initial historical record data in different scenarios.

Wherein, the initial historical record data may be initial log data of the terminal in different scenarios.

In a feasible implementation manner, log data generated when the terminal and the base station transmit data at historical time points may be acquired from multiple network managers, and the acquired log data may be determined as initial historical record data.

Exemplarily, as shown in FIG. 5 , the network management system in FIG. 5 may include multiple network managers such as network manager 1 to network manager m, and each network manager may manage multiple base stations such as base station 1 to base station n. Specifically, through the screening module 1, multiple base stations in typical high-speed scenarios such as mountain scene, viaduct scene, open field scene, and plain scene can be selected from the network management system, and the base station logs of these multiple base stations can be obtained. Logs serve as initial history data. Wherein, the screening module 1 mainly performs base station selection; in a feasible implementation manner, the screening module 1 may be a script corresponding to a screening logic for screening historical log data in different scenarios.

Step 201b: Screen the initial historical record data based on the operating parameters corresponding to the terminal to obtain target historical record data.

Wherein, the target historical record data is the historical record data filtered out from the initial historical record data; in other words, the target historical record data may be the target log data filtered out from the initial log data.

In a feasible implementation manner, as shown in FIG. 3 , the screening module 2 may be used to screen initial historical record data based on terminal operating parameters to obtain target historical record data. Further, through the screening module 2, based on the terminal's running speed and the decoding situation of the frequency offset value in the initial historical record data, the initial historical record data can be screened to obtain the target historical record data; Target historical record data, construct a data set for training target frequency offset prediction model. Among them, the screening module 2 mainly performs period selection, and the period selection may include but not limited to: the period during which the operating speed changes, the period between different time points, etc.; in a feasible implementation, the screening module 2 may be used for screening The script corresponding to the filtering logic of log data corresponding to different operating parameters.

Step 201c, based on the frequency offset value in the target historical record data, determine the frequency offset data to be trained.

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

In some embodiments, step 201b may be implemented through steps A1 to A2:

Step A1. From the initial historical record data, filter out the intermediate historical record data generated when the terminal transmits data at different operating speeds.

Among them, the intermediate historical record data is the historical record data generated by terminals filtered out from the initial historical record data at different operating speeds; Base station log generated at speed.

In a feasible implementation, from the initial historical record data, the log data generated by the train terminal at a speed of 250km/h to 350km/h can be selected as the intermediate historical record data; further, the train can also be selected at The log data generated by the terminal in the acceleration phase, deceleration phase, and constant speed phase are used as intermediate historical record data; specific settings can be made according to actual business needs, which is not limited in this embodiment of the application.

Step A2. From the intermediate historical record data, select the target historical record data whose data decoding is correct for N consecutive data frames.

Wherein, the base station log may include a scheduler log; the scheduler log may include results of uplink decoding and downlink decoding.

In a feasible implementation manner, the scheduler log may be selected from the base station log (intermediate historical record data), and the log data whose downlink decoding is correct for consecutive N+1 frames are screened out to obtain the target historical record data.

In the embodiment of the present application, the target frequency offset prediction model is trained by using the frequency offset values in the historical record data of different operating parameters in different scenarios, so that the target frequency offset prediction model can learn the values of different operating parameters in different scenarios. Frequency offset to more accurately predict the frequency offset at the current time point. Moreover, when training the target frequency offset prediction model, the decoding situation of each frequency offset value is also added, which further improves the model performance and model accuracy of the target frequency offset prediction model.

Based on the foregoing embodiments, after the target frequency offset prediction model is trained, the target frequency offset prediction model can be deployed in the base station to be used when the base station communicates with the terminal. The specific usage method can refer to the communication method shown in FIG. 6, the method includes the following steps:

Step 601. Determine historical frequency offset data when no data transmitted by the terminal is received within the target time period but data needs to be sent to the terminal.

Wherein, the target time period is determined based on the cycle of the time slot duration of the carrier scheduled when the terminal transmits data and the scheduling times threshold.

In this embodiment of the present application, the target time period is a trigger condition for predicting target frequency offset data using a target frequency offset prediction model. In a feasible implementation manner, a product operation may be performed on the period of the time slot duration of the carrier scheduled when the terminal transmits data and the scheduling times threshold to obtain the target time period. The calculation formula of the target time period can be: target time period=scheduling period*scheduling times threshold; wherein, the scheduling period refers to the period of the slot duration; taking 5G NR as an example, the central The time slot duration in the processor (Central Processor, CP) parameter set is as follows: 1ms, 0.5ms, 0.25ms, 0.125ms, 0.0625ms; correspondingly, the scheduling period is equal to the period of the time slot duration, and the scheduling times threshold is adjustable , the default value is 64, for example, when the period of the slot duration is 0.5ms, the target time period is 32ms.

In a feasible implementation, the historical frequency offset data will only be obtained when the data transmitted by the terminal is not received within the target time period but the data needs to be sent to the terminal, so as to use the target frequency offset prediction model to analyze the historical frequency offset The data is processed to obtain the predicted target frequency offset data.

Step 602, using the target frequency offset prediction model to process the historical frequency offset data to obtain the target frequency offset data.

In a feasible implementation manner, historical frequency offset data may be input into a target frequency offset prediction model for prediction to obtain target frequency offset data.

Step 603, perform frequency offset compensation on the signal transmitted with the terminal based on the target frequency offset data.

Step 604. Update the target frequency offset data into the frequency offset estimation data sequence.

In the embodiment of the present application, after the target frequency offset data is acquired, 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 frequency offset estimation data based on The historical frequency offset data determined by the frequency offset estimation data sequence is more in line with the current scene, so that the target frequency offset data obtained after the historical frequency offset data is processed by the target frequency offset prediction model is more accurate, thereby improving the downlink demodulation of the terminal performance.

In some embodiments, "determining historical frequency offset data" in step 601 may be implemented through the following steps 601a to 601b:

Step 601a, cyclically store a frequency offset estimation data sequence of a target length, and the frequency offset estimation data sequence includes a target amount of frequency offset estimation data.

Wherein, the target length is the total length of the cyclically stored frequency offset estimation data sequence; the target length can be preset, and the target length can be set according to actual service requirements, which is not limited in this 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 set in advance, and the target number can be set according to actual service requirements, which is not limited in this embodiment of the present application.

In a feasible implementation, the corresponding frequency offset data can be cyclically stored according to the time when the base station receives uplink data to obtain a frequency offset estimation data sequence of a target length; wherein the frequency offset estimation data sequence includes the frequency offset obtained by a traditional algorithm Offset data, and frequency offset data predicted by the model.

Step 601b, based on the frequency offset estimation data sequence of the target length, determine historical frequency offset data.

Wherein, the historical frequency offset data represents the frequency offset existing when transmitting data with the terminal.

In the embodiment of the present application, the frequency offset estimation data sequence of 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 filtered out from the at least one frequency offset estimation data sequence; Wherein, the quantity of frequency offset estimation data included in each frequency offset estimation data sequence is the target quantity.

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

Step B1. Determine the sequence accuracy rate of the frequency offset estimation data sequence.

Wherein, the frequency offset estimation data includes a frequency offset value and a corresponding data check mark; the data check mark is used to indicate whether the data is decoded correctly under the frequency offset value in the corresponding frequency offset estimate data.

In the embodiment of the present application, the sequence accuracy rate represents the correct data decoding under the frequency offset value in the frequency offset estimation data sequence.

Step B1 may be implemented by: determining the sequence accuracy rate based on the data verification identifier and the number of frequency offset values in the frequency offset estimation sequence.

In a feasible implementation, the data verification identifier in the frequency offset estimation sequence can be used to determine the number of correct frequency offset values for data decoding in the frequency offset estimation sequence; according to the correct frequency offset value for data decoding The number of frequency offset values and the total number of frequency offset values included in the frequency offset estimation data sequence determine the correct rate of the sequence. Specifically, the number of frequency offset values correctly decoded by the data may be divided by the total number of frequency offset values included in the frequency offset estimation data sequence to obtain the sequence accuracy rate.

Step B2, when the sequence accuracy meets the accuracy threshold requirement, use the frequency offset estimation data sequence as the historical frequency offset data.

In this embodiment of the application, the accuracy threshold requirement is the basis for judging whether the frequency offset estimation data sequence is historical frequency offset data; the accuracy threshold requirement can be set in advance, and can be set according to actual business needs, for example, the accuracy threshold requirement Can be set to 0.8. In a feasible implementation manner, the frequency offset estimation data sequence may be used as historical frequency offset data when the sequence accuracy rate is greater than or equal to the accuracy rate threshold requirement.

Based on the foregoing embodiments, in other embodiments of the present application, the communication method may further include the following steps:

Step 605. In the case that the sequence accuracy rate does not meet the accuracy rate threshold requirement, the frequency offset data used for the latest frequency offset compensation for the signal transmitted between the terminal and the terminal is used as the target frequency offset data, and based on the target frequency offset data, the The frequency offset compensation is performed on the signal transmitted between the terminal and the terminal.

In the embodiment of the present application, when the sequence accuracy of the frequency offset estimation data sequence is lower than the accuracy threshold requirement, the frequency offset data used for the latest frequency offset compensation for the signal transmitted between the terminal and the terminal can be directly used as The target frequency offset data, based on the target frequency offset data, performs frequency offset compensation for signals transmitted between the terminal and the terminal.

It should be noted that, in the communication method provided by the embodiment of the present application, regardless of whether there is uplink data or downlink data in the current time period, the target frequency offset pre-stored model can be used to predict the current frequency offset situation and obtain the target frequency offset data. Obtain the target frequency offset data; 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 calculate the current The frequency deviation is predicted to obtain the target frequency deviation data.

The communication method provided by the embodiment of the present application can be applied to many different high-speed scenarios such as high-speed railway scenarios and flight scenarios. However, in different high-speed scenarios, the involved environments and frequency offsets will be different, so it needs to be based on the actual application scenario. Adaptively adjust the construction method of the data set to make the constructed data set more in line with the actual application scenario.

It should be noted that, for descriptions of the same steps and content in this embodiment as in other embodiments, reference may be made to the descriptions in other embodiments, and details are not repeated here.

In the communication method provided by the embodiments of the present application, regardless of whether there is uplink data at the current time point, the historical frequency offset data can be processed through the target frequency offset prediction model to predict the frequency offset data at the current time point (target frequency offset data); and, the historical frequency offset data represents the frequency offset that exists when data is transmitted between the terminal 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 historical time points, then the target The frequency offset prediction model processes the historical frequency offset data to predict the frequency offset data at the current time point. During the prediction, the frequency offset situation generated at the historical time point is comprehensively considered, and the predicted target frequency offset data is more accurate and closer. The current data transmission scenario can improve the downlink demodulation performance of the terminal, and solve the problem that the compensation effect is poor and the downlink demodulation performance of the terminal is reduced in the related art.

The following describes the application of the communication method provided by the embodiment of the present application in an actual scene, taking a multi-terminal in a train scene as an example for illustration.

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

Step 701, cyclically store the frequency offset estimation data sequence of the target length according to the terminal.

In a feasible implementation manner, frequency offset estimation data sequences of a target length may be cyclically stored according to terminal identifiers. The purpose of cyclically storing the frequency offset estimation data sequence of the target length according to the terminal is to efficiently compensate multiple terminals in the same environment according to the environment where the terminal is located, so as to improve processing efficiency.

During normal operation of the base station, when there is uplink data, the frequency offset estimation data sequence of the target length is cyclically stored by the terminal. As shown in Figure 8, the frequency offset estimation data sequence 1 in Figure 8 includes a frequency offset value sequence composed of n frequency offset values such as frequency offset value 1, frequency offset value 2, ..., frequency offset value n, and corresponding data A data check indicator bit sequence (data check mark) composed of check indicator bits; sequence 2 includes a frequency offset value composed of n frequency offset values such as frequency offset value 2, frequency offset value 3, ..., frequency offset value n+1 Offset value sequence and corresponding data check indication bit sequence composed of data check indication bits; the data check code identification is used to indicate whether the data is decoded correctly under the corresponding frequency offset value; if the downlink data is decoded correctly (ie ACK), the corresponding bit is 1; if the downlink data is decoded incorrectly (that is, NACK), the corresponding bit is 0. When implemented, n frequency offset values will be stored first to obtain frequency offset estimation data sequence 1; when there is a new frequency offset value, the new frequency offset value will be added to the frequency offset estimation data as the n+1th frequency offset value In the sequence, based on frequency offset value 2, frequency offset value 3, ..., frequency offset value n+1 and other n frequency offset values to form a new frequency offset estimation data sequence (ie, frequency offset estimation data sequence 2), this will always The frequency offset estimation data sequence is updated based on the new frequency offset value to cyclically store the frequency offset estimation data sequence with a length of n, so that the n frequency offset values in the frequency offset estimation data sequence are the latest.

It should be noted that in the process of cyclic storage, if the traditional algorithm is used to obtain the frequency offset value when there is uplink data, and the frequency offset value is obtained by model prediction when there is no uplink data, then the frequency offset value in the frequency offset estimation data sequence There may be the following three situations: 1. When uplink data exists for a long period of time, the frequency offset values in the frequency offset estimation data sequence at this time are all frequency offset values obtained by using traditional algorithms; 2. In a period of time If there is uplink data within a period of time and there is no uplink data for a period of time, part of the frequency offset value in the frequency offset estimation data sequence comes from the traditional algorithm and part from the model prediction; 3. There is no uplink data for a long period of time At this time, the frequency offset values in the frequency offset estimation data sequence are all obtained by model prediction.

Step 702. Determine whether the time when the terminal has downlink scheduling and no 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 terminal has downlink scheduling and no uplink scheduling time If the target time period is exceeded, and the sequence accuracy of the frequency offset estimation data sequence meets the accuracy threshold requirement, step 703 is executed.

In the embodiment of the present application, as shown in FIG. 9 , the implementation of determining the target time period may be as follows: step 901, obtain the period of the time slot duration configured by the base station; step 902, obtain the threshold of the number of scheduling times; The period of the slot duration and the threshold of scheduling times are multiplied to obtain the target time period.

The judging module judges whether the terminal satisfies condition 1 (that is, there is no scheduling of uplink traffic channel while there is downlink scheduling within the target time period) in each scheduling period, and satisfies condition 2 (data verification of the frequency offset estimated value sequence of the terminal) at the same time. The correct rate meets the correct rate threshold requirement); if both condition 1 and condition 2 are met, then execute step 703; if not simultaneously meet condition 1 and condition 2, then execute step 706.

Step 703, using the target frequency offset prediction model to process the frequency offset estimation data sequence of each terminal to obtain target frequency offset data.

In a feasible implementation, the network architecture of the target frequency offset prediction model can be the network architecture shown in Figure 5 to ensure the model performance of the target frequency offset prediction model, thereby ensuring the accuracy of the predicted target frequency offset data .

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

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

Step 705: Synthesize target frequency offset data of all terminals and select an appropriate frequency offset value to perform frequency offset compensation for downlink data.

In a feasible implementation manner, an average value of target frequency offset data of all terminals may be determined, and frequency offset compensation is performed on downlink data according to the average value.

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

Step 706: Perform frequency offset compensation for the signal transmitted between the base station and the terminal this time by using the frequency offset data used when performing frequency offset compensation for the signal transmitted between the base station and the terminal last time.

It should be noted that both the base station and the terminal can perform frequency offset compensation for the transmitted data, and generally the base station side performs a large-scale frequency offset compensation; for example, if the frequency offset when transmitting data between the base station and the terminal is 1010, Then the base station side will make up 1000, and the terminal side will make up 10. The embodiment of the present application is applied to the base station side as an example to illustrate the communication method provided by the embodiment of the present application, but the communication method provided in the embodiment of the present application can also be extended to the terminal side; The historical frequency offset exists in the data, and the target frequency offset prediction model is used to process the historical frequency offset to obtain the predicted frequency offset data, and the pre-stored frequency offset data is used to compensate the frequency offset of the signal transmitted between the base station.

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

Based on the foregoing embodiments, this embodiment of the present application provides a communication device, which can be applied to the communication method provided in the embodiments corresponding to Figures 1 and 6, as shown in Figure 10, the communication device includes:

The determining unit 1001 is configured to determine historical frequency offset data; wherein, the historical frequency offset data represents the frequency offset that exists when transmitting data with the terminal;

The processing unit 1002 is configured to predict the current frequency offset situation according to historical frequency offset data, and obtain target frequency offset data;

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

In the embodiment of this application, the determining unit 1001 is also configured to perform the following steps:

cyclically storing a frequency offset estimation data sequence of a target length, the frequency offset estimation data sequence including a target quantity of frequency offset estimation data;

Based on the frequency offset estimation data sequence of the target length, historical frequency offset data is determined.

In the embodiment of the present application, the determination unit 1001 is also used to determine the historical frequency offset data when the data transmitted by the terminal is not received within the target time period but the data needs to be sent to the terminal; the target time period is based on the terminal transmission data Determined by the period of the time slot duration of the scheduled carrier and the scheduling times threshold;

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

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

Determine the sequence accuracy rate of the frequency offset estimation data sequence;

When the sequence accuracy meets the accuracy threshold requirement, the frequency offset estimation data sequence is used as the historical frequency offset data.

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

In the embodiment of the present application, the processing unit 1002 is further configured to use the frequency offset data used when performing frequency offset compensation for the signal transmitted between the terminal and the terminal last time as The target frequency offset data, based on the target frequency offset data, performs frequency offset compensation for signals transmitted between the terminal and the terminal.

In the embodiment of the present application, the processing unit 1002 is further configured to process historical frequency offset data based on the target frequency offset prediction model, so as to predict the current frequency offset situation;

The training methods of the target frequency offset prediction model include:

Determine the confidence level corresponding to the frequency offset data to be trained and the frequency offset data to be trained in different scenarios;

Based on the frequency offset data to be trained and the confidence level, the initial frequency offset prediction model is trained to obtain the target frequency offset prediction model.

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

Obtain initial historical record data in different scenarios;

Based on the operating parameters corresponding to the terminal, the initial historical record data is screened to obtain the target historical record data;

The frequency offset data to be trained is determined based on the frequency offset value in the target historical record data.

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

From the initial historical record data, filter out the intermediate historical record data generated when the terminal transmits data at different operating speeds;

From the intermediate historical record data, filter out the target historical record data whose data decoding is correct for N consecutive data frames.

It should be noted that, for the specific implementation process of the steps performed by each unit in this embodiment, reference may be made to the implementation process in the communication method provided in the embodiment corresponding to FIGS. 1 and 6 , which will not be repeated here.

The communication device provided by the embodiments of the present application can process the historical frequency offset data through the target frequency offset prediction model to predict the frequency offset data at the current time point (target frequency offset data); and, the historical frequency offset data represents the frequency offset that exists when data is transmitted between the terminal 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 historical time points, then the target The frequency offset prediction model processes the historical frequency offset data to predict the frequency offset data at the current time point. During the prediction, the frequency offset situation generated at the historical time point is comprehensively considered, and the predicted target frequency offset data is more accurate and closer. The current data transmission scenario can improve the downlink demodulation performance of the terminal, and solve the problem that the compensation effect is poor and the downlink demodulation performance of the terminal is reduced in the related art.

Based on the foregoing embodiments, the embodiments of the present application provide a base station, which can be applied to the communication method provided in the embodiments corresponding to FIGS. 1 and 6. Referring to FIG. 11, the device may include: a processor 111, a memory 112 and communication bus 113;

Determine the historical frequency offset data; wherein, the historical frequency offset data represents the frequency offset that exists when transmitting signals with the terminal;

Predict the current frequency deviation situation according to the historical frequency deviation data, and obtain the target frequency deviation data;

Based on the target frequency offset data, the frequency offset compensation is performed on the signal transmitted between the terminal and the terminal.

In other embodiments of the present application, the processor 111 is used to execute the communication program in the memory 112 to determine the historical frequency offset data, so as to implement the following steps:

cyclically storing a frequency offset estimation data sequence of a target length, the frequency offset estimation data sequence including a target quantity of frequency offset estimation data;

Based on the frequency offset estimation data sequence of the target length, historical frequency offset data is determined.

In other embodiments of the present application, the processor 111 is used to execute the communication program in the memory 112 to determine the historical frequency offset data, so as to implement the following steps:

When the data transmitted by the terminal is not received within the target time period but the data needs to be sent to the terminal, determine the historical frequency offset data; the target time period is based on the period and the number of scheduling times of the time slot duration of the carrier scheduled when the terminal transmits data Threshold determined;

Correspondingly, the processor 111 is used to execute the communication program in the memory 112 by using the target frequency offset prediction model to process the historical frequency offset data, and after obtaining the target frequency offset data, the following steps can also be implemented:

The target frequency offset data is updated into the frequency offset estimation data sequence.

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

Determine the sequence accuracy rate of the frequency offset estimation data sequence;

When the sequence accuracy meets the accuracy threshold requirement, the frequency offset estimation data sequence is used as the historical frequency offset data.

In other embodiments of the present application, the frequency offset estimation data includes a frequency offset value and a corresponding data check mark; the data check mark is used to indicate whether the data is decoded correctly under the corresponding frequency offset value; the processor 111 is used to Execute the sequence correct rate of determining the frequency offset estimation data sequence of the communication program in the memory 112, to realize the following steps:

Based on the data check mark and the number of frequency offset values in the frequency offset estimation sequence, the correct rate of the sequence is determined.

In other embodiments of the present application, the processor 111 is used to execute the communication program in the memory 112 and may also implement the following steps:

In the case that the sequence accuracy rate does not meet the accuracy threshold requirement, the frequency offset data used in the latest frequency offset compensation for the signal transmitted between the terminal and the terminal is used as the target frequency offset data, and based on the target frequency offset data pair with the terminal Frequency offset compensation is performed on the signals transmitted between them.

In other embodiments of the present application, the processor 111 is used to execute the communication program in the memory 112 to predict the current frequency deviation according to the historical frequency deviation data, and the following steps can be implemented:

Process the historical frequency offset data based on the target frequency offset prediction model to predict the current frequency offset situation;

In other embodiments of the present application, the processor 111 is used to execute the communication program in the memory 112 and may also implement the following steps:

Determine the confidence level corresponding to the frequency offset data to be trained and the frequency offset data to be trained in different scenarios;

Based on the frequency offset data to be trained and the confidence level, the initial frequency offset prediction model is trained to obtain the target frequency offset prediction model.

In other embodiments of the present application, the processor 111 is used to execute the communication program in the memory 112 to determine the frequency offset data to be trained in different scenarios, so as to implement the following steps:

Obtain initial historical record data in different scenarios;

Based on the operating parameters corresponding to the terminal, the initial historical record data is screened to obtain the target historical record data;

The frequency offset data to be trained is determined based on the frequency offset value in the target historical record data.

In other embodiments of the present application, the processor 111 is used to execute the communication program in the memory 112 to screen the initial historical record data based on the operating parameters corresponding to the terminal to obtain the target historical record data, so as to implement the following steps:

From the initial historical record data, filter out the intermediate historical record data generated when the terminal transmits data at different operating speeds;

From the intermediate historical record data, filter out the target historical record data whose data decoding is correct for N consecutive data frames.

It should be noted that, for a specific description of the steps performed by the processor, reference may be made to the communication method provided in the embodiments corresponding to FIGS. 1 and 6 , and details are not repeated here.

The base station provided by the embodiments of the present application can process the historical frequency offset data through the target frequency offset prediction model to predict the frequency offset data at the current time point (the target frequency offset data ); and, the historical frequency offset data represents the frequency offset that exists when transmitting data with 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 historical time points, then the target frequency The deviation prediction model processes the historical frequency deviation data to predict the frequency deviation data at the current time point. During the prediction, the frequency deviation generated at the historical time point is comprehensively considered, and the predicted target frequency deviation data is more accurate and more suitable. The current data transmission scenario can further improve the downlink demodulation performance of the terminal, and solve the problem in the related art that the compensation effect is poor and the downlink demodulation performance of the terminal is reduced.

Based on the foregoing embodiments, the embodiments of the present application provide a computer-readable storage medium, where one or more programs are stored in the computer-readable storage medium, and the one or more programs can be executed by one or more processors to The steps of the communication method provided by the embodiments corresponding to FIGS. 1 and 6 are implemented.

It should be noted that the above-mentioned computer-readable storage medium may be a read-only memory (Read Only Memory, ROM), a programmable read-only memory (Programmable Read-Only Memory, PROM), an erasable programmable read-only memory (Erasable Programmable Read-Only Memory, EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Magnetic Random Access Memory (Ferromagnetic Random Access Memory, FRAM), Flash Memory (Flash Memory) , magnetic surface memory, CD, or CD-ROM (Compact Disc Read-Only Memory, CD-ROM) and other storage; it can also be various electronic devices including one or any combination of the above storage, such as mobile phones, computers, tablet devices, personal digital assistants, and more.

It should be noted that, in this document, the term "comprising", "comprising" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article or apparatus comprising a set of elements includes not only those elements, It also includes other elements not expressly listed, or elements inherent in the process, method, article, or device. Without further limitations, an element defined by the phrase "comprising a ..." does not preclude the presence of additional identical elements in the process, method, article, or apparatus comprising that element.

The serial numbers of the above embodiments of the present application are for description only, and do not represent the advantages and disadvantages of the embodiments.

Through the description of the above embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus a necessary general-purpose hardware platform, and of course also by hardware, but in many cases the former is better implementation. Based on such an understanding, the technical solution of the present application can be embodied in the form of a software product in essence or the part that contributes to the prior art, and the computer software product is stored in a storage medium (such as ROM/RAM, disk, CD) contains several instructions to make a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) execute the methods described in various embodiments of the present application.

The present application is described with reference to flowcharts and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and a combination of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing equipment produce a An apparatus for realizing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions The device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby The instructions provide steps for implementing the functions specified in the flow chart or blocks of the flowchart and/or the block or blocks of the block diagrams.

The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structures or equivalent process transformations made by using the description of the application and the accompanying drawings are directly or indirectly used in other related technical fields. , are all included in the patent protection scope of the present application in the same way.

Claims (10)

1. A communication method, wherein the method comprises: Determine historical frequency offset data; wherein, the historical frequency offset data represents the frequency offset that exists when transmitting data with the terminal; Predicting the current frequency offset situation according to the historical frequency offset data to obtain target frequency offset data; Perform frequency offset compensation on signals transmitted with the terminal based on the target frequency offset data. 2. The method according to claim 1, wherein said determining historical frequency offset data comprises: cyclically storing a frequency offset estimation data sequence of a target length, the frequency offset estimation data sequence including a target quantity of frequency offset estimation data; The historical frequency offset data is determined based on the frequency offset estimation data sequence of the target length. 3. The method of claim 1, wherein, The determining the historical frequency offset data includes: When the data transmitted by the terminal is not received within the target time period but the data needs to be sent to the terminal, determine the historical frequency offset data; the target time period is based on the carrier scheduled when the terminal transmits data Determined by the period of the time slot duration and the threshold of the number of scheduling times; The method of predicting the current frequency offset situation according to the historical frequency offset data, and after obtaining the target frequency offset data, further includes: The target frequency offset data is updated into the frequency offset estimation data sequence. 4. The method according to claim 2, wherein said determining the historical frequency offset data based on the frequency offset estimation data sequence of the target length comprises: determining the sequence accuracy rate of the frequency offset estimation data sequence; In the case that 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 of claim 2, wherein, The frequency offset estimation data includes a frequency offset value and a corresponding data check mark; the data check mark is used to indicate whether the data decoding is correct under the corresponding frequency offset value; The determining the sequence accuracy rate of the frequency offset estimation data sequence includes: Determine the sequence accuracy rate based on the data verification identifier and the number of frequency offset values in the frequency offset estimation sequence. 6. The method of claim 4, wherein the method further comprises: In the case that the sequence accuracy rate does not meet the accuracy rate threshold requirement, the frequency offset data used when performing frequency offset compensation on the signal transmitted between the terminal and the terminal last time is used as the target frequency offset data, based on the The target frequency offset data performs frequency offset compensation on signals transmitted between the terminal and the terminal. 7. The method according to claim 1, wherein the predicting the current frequency offset situation according to the historical frequency offset data comprises: Processing the historical frequency offset data based on the target frequency offset prediction model to predict the current frequency offset situation; The training method of the target frequency offset prediction model includes: Determining the frequency offset data to be trained in different scenarios and the confidence corresponding to the frequency offset data to be trained; Perform model training on an initial frequency offset prediction model 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 said determining frequency offset data to be trained in different scenarios comprises: Obtain initial historical record data in different scenarios; Screening the initial historical record data based on the operating parameters corresponding to the terminal to obtain target historical record data; The frequency offset data to be trained is determined based on the frequency offset value in the target historical record data. 9. The method according to claim 8, wherein said initial historical record data is screened based on the operating parameters corresponding to said terminal to obtain target historical record data, comprising: From the initial historical record data, filter out the intermediate historical record data generated when the terminal transmits data at different operating speeds; From the intermediate historical record data, the target historical record data whose data decoding is correct for N consecutive data frames is screened out. 10. A base station, wherein the base station comprises: a processor and a memory; The communication bus is used to realize the communication connection between the processor and the memory; The processor is used to execute the communication program stored in the memory, so as to realize the following steps: Determine historical frequency offset data; wherein, the historical frequency offset data represents the frequency offset that exists when transmitting signals with the terminal; Predicting the current frequency offset situation according to the historical frequency offset data to obtain target frequency offset data; Perform frequency offset compensation on signals transmitted with the terminal based on the target frequency offset data.