A method for predicting a channel time series and related devices

Abstract

This application discloses a method and related apparatus for predicting channel time series at future times. The method includes: firstly, acquiring the channel time series at M times to obtain X(t) m )={x n (t m Let X(t) = |m=1,2,…,M} (n=1,2,…,N), where N is the number of channel time series within a given time moment, and M and N are both positive integers. Next, based on the stated X(t)... m )={x n (t m Construct a mapping Ψ for each m = 1, 2, ..., M (n = 1, 2, ..., N). p , among which, Ψ p (X(t m ))=x n (t m+p‑1 (p=2,3,…,L), where L-1 is the number of time steps to be predicted, and X(t) is used as the time step to be predicted. m (m=1,2,…,M-p+1) as input, with x n (t m+p‑1 Using (m=1,2,…,M-p+1) as the output, we perform a Gaussian process regression (GPR) fitting to obtain Ψ. p (p = 2, 3, ..., L). Finally, X(t) m (m = M - p + 2, M - p + 3, ..., M) respectively input the Ψ p (p = 2, 3, ..., L), we get x n (t m+p‑1 This application generates predicted values ​​for (n = 1, 2, ..., N; m = 1, 2, ..., M; p = 2, 3, ..., L), thus achieving multi-step prediction of channel time series. Compared with existing single-step prediction, this application does not accumulate or amplify errors during the prediction process, resulting in more accurate prediction results.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a method and related equipment for predicting channel time series. Background Technology

[0002] Communication between a terminal device and a base station in the airspace can consist of 4*48 channels, each with 168 sub-channels. At any given time, the communication channels between the terminal device and the base station can be described by 4*48*168 data points. Based on the currently observed channel time series, the receiver can predict future changes in that time series, which can help improve the performance of various operations performed at the transmitter or receiver. Examples include adaptive coding and modulation, decoding processing, channel equalization, and antenna beamforming.

[0003] Currently, common methods for predicting channel time series include autoregressive (AR) methods, exponential smoothing methods, and moving average methods. Taking AR as an example, its prediction process can be... Description. Where x t Let c be the value of the variable to be predicted at time t, and ε be a constant. t (epsilon) is a random error value with a mean of 0 and a standard deviation of σ(sigma), where i = {1, 2, ..., p} are the indices of the p time points before time t. Let x be the autocorrelation coefficient of this variable. t-i Let be the variable value observed at time ti.

[0004] However, the above method can only achieve single-step prediction, that is, predicting the channel time series at the next time point based on the currently observed channel time series. If we want to obtain the channel time series at several times after the current time, we need to iterate step by step based on the channel time series obtained from each prediction. In this case, the prediction error will accumulate and amplify continuously during the iteration process, resulting in an inaccurate final prediction result. Summary of the Invention

[0005] This application provides a method and related apparatus for predicting channel time series, used to predict channel time series at future times.

[0006] The first aspect of this application provides a method for predicting channel time series. In this application, the channel time series at M time points are first obtained to obtain X(t) m )={x n (t m Let X(t) = |m=1,2,…,M} (n=1,2,…,N), where N is the number of channel time series within a given time moment, and M and N are both positive integers. Next, based on the stated X(t)...m )={x n (t m Construct a mapping Ψ for each m = 1, 2, ..., M (n = 1, 2, ..., N). p , among which, Ψ p (X(t m ))=x n (t m+p-1 (p=2,3,…,L), where L-1 is the number of time steps to be predicted, and X(t) is used as the time step to be predicted. m (m=1,2,…,M-p+1) as input, with x n (t m+p-1 Using (m=1,2,…,M-p+1) as the output, we perform a Gaussian process regression (GPR) fitting to obtain Ψ. p (p = 2, 3, ..., L). Finally, X(t) m (m = M - p + 2, M - p + 3, ..., M) respectively input the Ψ p (p = 2, 3, ..., L), we get x n (t m+p-1 This application generates predicted values ​​for (n = 1, 2, ..., N; m = 1, 2, ..., M; p = 2, 3, ..., L), thus achieving multi-step prediction of channel time series. Compared with existing single-step prediction, this application does not accumulate or amplify errors during the prediction process, resulting in more accurate prediction results.

[0007] In some feasible implementations, the statement based on X(t) m )={x n (t m Construct a mapping Ψ for each m = 1, 2, ..., M (n = 1, 2, ..., N). p include:

[0008] Construct the Space-Time Information (STI) equation, which is expressed as:

[0009]

[0010] Therefore, Ψ can be obtained by solving the STI equation above. p .

[0011] In some feasible implementations, the statement based on X(t) m )={x n (t m Construct a mapping Ψ for each m = 1, 2, ..., M (n = 1, 2, ..., N). p Afterwards, the Ψ can be obtained. p Length dimension parameter l p , where l p ={lp1 ,l p2 ,…,l pN}, l pn With x n (t m Corresponding to, and taking the l p ={l p1 ,l p2 ,…,l pN Find the largest P values ​​in} to obtain l p '={l p '1,l p '2,…,l p ' P}, determine the l p '={l p '1,l p '2,…,l p ' P x corresponding to each value in} n (t m ), thus obtaining the characteristic variable X p '(t m )={x p '1(t m ),x p '2(t m ),…,x p ' P (t m Then, with Ψ p (X p '(t m (n = 1, 2, ..., N) is the output, and the characteristic variable X is... p '(t m Using (n = 1, 2, ..., N) as input, fit the GPR to obtain the optimized Ψ. p (p = 1, 2, ..., L), thus achieving the control of Ψ p Optimization yields a better Ψ p .

[0012] In some feasible implementations, X(t) m (m = M - p + 2, M - p + 3, ..., M) respectively input the Ψ p (p = 2, 3, ..., L), respectively, we obtain Ψ p x corresponding to (p = 2, 3, ..., L) n (t M+p-1 ), calculate the Ψ p x corresponding to (p = 2, 3, ..., L) n (t M+p-1 The average value of ) is used as the x n (tM+p-1 This allows for multi-step prediction by obtaining the predicted value of ).

[0013] In some feasible implementations, the step of X(t) m (m = M - p + 2, M - p + 3, ..., M) respectively input the Ψ p (p = 2, 3, ..., L), we get x n (t m+p-1 The predicted values ​​for (n = 1, 2, ..., N; m = 1, 2, ..., M; p = 2, ..., L) include:

[0014] S1. Let p = 2;

[0015] S2. With X(t) m (m=1,2,…,M-p'+1) as input, with x n (t m+p’-1 Using (m=1,2,…,M-p'+1) as the output, fit it to GPR to obtain Ψ p’ Where p' = p, p+1, p+2, ..., L;

[0016] S3. X(t) m (m = M - (p - 2), M - (p - 1), ..., M) input Ψ respectively p ,Ψ p+1 ,…,Ψ L We get L-p+1 x n (t M+p-1 );

[0017] S4. Calculate L-p+1 x n (t M+p-1 The average of ) is used as x n (t M+p-1 The predicted value of );

[0018] S5. If p is not equal to L, then p = p + 1 and execute step S2; otherwise, the execution is complete.

[0019] Since the previous prediction data is used as known data to calculate the next prediction data, the accuracy and performance of the prediction can be improved.

[0020] A second aspect of this application provides a channel time series prediction device, comprising:

[0021] The transceiver module is used to acquire the channel time series at M time points to obtain X(t) m )={x n (t m)|m=1,2,…,M}(n=1,2,…,N), where N is the number of channel time series within a given time moment, and M and N are both positive integers;

[0022] Processing module, used for processing based on X(t) m )={x n (t m Construct a mapping Ψ for each m = 1, 2, ..., M (n = 1, 2, ..., N). p , among which, Ψ p (X(t m ))=x n (t m+p-1 (p = 2, 3, ..., L), where L-1 is the number of time steps to be predicted;

[0023] The processing module is also used to use X(t) m (m=1,2,…,M-p+1) as input, with x n (t m+p-1 Using (m=1,2,…,M-p+1) as the output, we perform a Gaussian process regression (GPR) fitting to obtain Ψ. p (p = 2, 3, ..., L);

[0024] The processing module is also used to process X(t) m (m = M - p + 2, M - p + 3, ..., M) respectively input the Ψ p (p = 2, 3, ..., L), we get x n (t m+p-1 Predicted values ​​for (n = 1, 2, ..., N; m = 1, 2, ..., M; p = 2, 3, ..., L).

[0025] In some feasible implementations, the processing module is specifically used for:

[0026] Construct the Space-Time Information (STI) equation, which is expressed as:

[0027]

[0028] In some feasible implementations, the processing module is further configured to:

[0029] Obtain the Ψ p Length dimension parameter l p , where l p ={l p1 ,l p2 ,…,l pN}, l pn With x n (t m )correspond;

[0030] Take the l p ={l p1 ,l p2 ,…,l pN Find the largest P values ​​in} to obtain l p '={l p '1,l p '2,…,l p ' P};

[0031] Determine the l p '={l p '1,l p '2,…,l p ' P x corresponding to each value in} n (t m ), thus obtaining the characteristic variable X p '(t m )={x p '1(t m ),x p '2(t m ),…,x p ' P (t m )};

[0032] With Ψ p (X p '(t m (n = 1, 2, ..., N) is the output, and the characteristic variable X is... p '(t m Using (n = 1, 2, ..., N) as input, fit the GPR to obtain the optimized Ψ. p (p = 1, 2, ..., L).

[0033] In some feasible implementations, the processing module is specifically used for:

[0034] X(t) m (m = M - p + 2, M - p + 3, ..., M) respectively input the Ψ p (p = 2, 3, ..., L), respectively, we obtain Ψ p x corresponding to (p = 2, 3, ..., L) n (t M+p-1 );

[0035] Calculate the Ψ p x corresponding to (p = 2, 3, ..., L) n (t M+p-1 The average value of ) is used as the x n (t M+p-1 The predicted value of ).

[0036] In some feasible implementations, the processing module is specifically used for:

[0037] S1. Let p = 2;

[0038] S2. With X(t) m (m=1,2,…,M-p'+1) as input, with x n (t m+p’-1 Using (m=1,2,…,M-p'+1) as the output, fit it to GPR to obtain Ψ p’ Where p' = p, p+1, p+2, ..., L;

[0039] S3. X(t) m (m = M - (p - 2), M - (p - 1), ..., M) input Ψ respectively p ,Ψ p+1 ,…,Ψ L We get L-p+1 x n (t M+p-1 );

[0040] S4. Calculate L-p+1 x n (t M+p-1 The average of ) is used as x n (t M+p-1 The predicted value of );

[0041] S5. If p is not equal to L, then p = p + 1 and execute step S2; otherwise, the execution is complete.

[0042] A third aspect of this application provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method described in any one of the first, second, or third aspects above.

[0043] A fourth aspect of this application provides a computer program product including computer-executable instructions stored in a computer-readable storage medium; at least one processor of the device can read the computer-executable instructions from the computer-readable storage medium, and the at least one processor executes the computer-executable instructions to cause the device to implement the method provided by the first aspect or any possible implementation thereof.

[0044] A fifth aspect of this application provides a communication device that may include at least one processor, a memory, and a communication interface. The at least one processor is coupled to the memory and the communication interface. The memory is used to store instructions, the at least one processor is used to execute the instructions, and the communication interface is used to communicate with other communication devices under the control of the at least one processor. When executed by the at least one processor, the instructions cause the at least one processor to perform a method of the first aspect or any possible implementation thereof.

[0045] The sixth aspect of this application provides a chip system including a processor for supporting the implementation of the functions involved in the first aspect or any possible implementation thereof.

[0046] In one possible design, the chip system may also include a memory for storing necessary program instructions and data. The chip system can be composed of chips or may include chips and other discrete components.

[0047] The technical effects of the second to sixth aspects or any of their possible implementations can be found in the first aspect or the technical effects of different possible implementations of the first aspect, and will not be repeated here. Attached Figure Description

[0048] Figure 1 This is a schematic diagram of the composition structure of a communication system provided in an embodiment of this application;

[0049] Figure 2-1 A flowchart illustrating a channel time series prediction method provided in an embodiment of this application;

[0050] Figure 2-2 This is a schematic diagram illustrating the implementation effect of multi-step prediction of channel time series in this application;

[0051] Figure 2-3 This is another schematic diagram illustrating the implementation effect of multi-step prediction of channel time series in this application;

[0052] Figure 3 This is a schematic diagram of the structure of a prediction device provided in an embodiment of this application;

[0053] Figure 4 This is a schematic diagram of the structure of a communication device provided in an embodiment of this application. Detailed Implementation

[0054] This application provides a method for predicting channel time series, used to predict channel time series at future times.

[0055] The embodiments of this application will now be described with reference to the accompanying drawings.

[0056] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such terms are interchangeable where appropriate; this is merely a way of distinguishing objects with the same attributes in the embodiments of this application. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, so that a process, method, system, product, or apparatus that comprises a series of elements is not necessarily limited to those elements, but may include other elements not explicitly listed or inherent to those processes, methods, products, or apparatuses.

[0057] Please see Figure 1 The diagram shown is a schematic representation of the structural composition of a communication system according to an embodiment of this application. This application provides a communication system 100, including a base station 110 and multiple terminal devices 120.

[0058] The base station 110 in this application embodiment is used as an access device to access the mobile communication system wirelessly. Examples of devices 110 include: next-generation Node B (gNodeB), transmission reception point (TRP), transmission reception point (TRP), evolved Node B (eNB), radio network controller (RNC) in 5G communication systems, base stations in future mobile communication systems, or access nodes, Node B (NB), base station controller (BSC), base transceiver station (BTS) in WiFi systems, home base stations (e.g., home evolved Node B, or home Node B (HNB), base band unit (BBU), or wireless fidelity (Wi-Fi) access point (AP), etc. The embodiments of this application do not limit the specific technology or device form used in the network equipment.

[0059] Terminal equipment, also known as user equipment (UE), mobile station (MS), mobile terminal (MT), or simply terminal, is a device that provides voice and / or data connectivity to a user, or a chip embedded within that device. Examples include handheld devices and in-vehicle devices with wireless connectivity. Currently, some examples of terminal equipment include: mobile phones, desktop computers, tablets, laptops, PDAs, mobile internet devices (MID), wearable devices, virtual reality (VR) devices, augmented reality (AR) devices, wireless terminals in industrial control, wireless terminals in self-driving cars, wireless terminals in remote medical surgery, wireless terminals in smart grids, wireless terminals in transportation safety, wireless terminals in smart cities, wireless terminals in smart homes, and 5G-residential gateways (5G-RG) that support 5G integration.

[0060] Communication between a terminal device and a base station in the airspace can consist of 4*48 channels, each with 168 sub-channels. At any given time, the communication channels between the terminal device and the base station can be described by 4*48*168 data points. Based on the currently observed channel time series, the receiver can predict future changes in that time series, which can help improve the performance of various operations performed at the transmitter or receiver. Examples include adaptive coding and modulation, decoding processing, channel equalization, and antenna beamforming.

[0061] Currently, common methods for predicting channel time series include autoregressive (AR) methods, exponential smoothing methods, and moving average methods. Taking AR as an example, its prediction process can be... Description. Where x t Let c be the value of the variable to be predicted at time t, and ε be a constant. t (epsilon) is a random error value with a mean of 0 and a standard deviation of σ(sigma), where i = {1, 2, ..., p} are the indices of the p time points before time t. Let x be the autocorrelation coefficient of this variable. t-i Let be the variable value observed at time ti.

[0062] However, the above method can only achieve single-step prediction, that is, predicting the channel time series at the next time point based on the currently observed channel time series. If we want to obtain the channel time series at several times after the current time, we need to iterate step by step based on the channel time series obtained from each prediction. In this case, the prediction error will accumulate and amplify continuously during the iteration process, resulting in an inaccurate final prediction result.

[0063] In this application, the channel time series at M time points are first obtained to obtain X(t) m )={x n (t m Let X(t) = |m=1,2,…,M} (n=1,2,…,N), where N is the number of channel time series within a given time moment, and M and N are both positive integers. Next, based on the stated X(t)... m )={x n (t m Construct a mapping Ψ for each m = 1, 2, ..., M (n = 1, 2, ..., N). p , among which, Ψ p (X(t m ))=x n (t m+p-1 (p=2,3,…,L), where L-1 is the number of time steps to be predicted, and X(t) is used as the time step to be predicted. m (m=1,2,…,M-p+1) as input, with x n (t m+p-1 Using (m=1,2,…,M-p+1) as the output, we perform a Gaussian process regression (GPR) fitting to obtain Ψ. p (p = 2, 3, ..., L). Finally, X(t) m (m = M - p + 2, M - p + 3, ..., M) respectively input the Ψ p (p = 2, 3, ..., L), we get x n (t m+p-1 This application generates predicted values ​​for (n = 1, 2, ..., N; m = 1, 2, ..., M; p = 2, 3, ..., L), thus achieving multi-step prediction of channel time series. Compared with existing single-step prediction, this application does not accumulate or amplify errors during the prediction process, resulting in more accurate prediction results.

[0064] Please see Figure 2-1 As shown in the embodiments of this application, a method for predicting channel time series mainly includes the following steps:

[0065] 201. The prediction device acquires the channel time series at M time points to obtain X(t). m )={x n (tm Let M = |m = 1, 2, ..., M} (n = 1, 2, ..., N), where N is the number of channel time series within a given time moment, and M and N are both positive integers.

[0066] In this embodiment, a preset device is used to predict channel time series at one or more future times. The prediction device can be a receiver, a transmitter, or a third-party device; no limitation is made here. The receiver can be a base station, and the transmitter can be a terminal device; or, the receiver can be a terminal device, and the transmitter can be a base station; no limitation is made here. For example, the receiver receives channel time series at multiple times transmitted by the transmitter to obtain observation data; or, the transmitter acquires channel time series at multiple times transmitted to obtain observation data; or, a third-party prediction device acquires channel time series at multiple times forwarded by the receiver or transmitter to obtain observation data. No limitation is made here.

[0067] In some feasible implementations, the channel time series at a given moment can be set as X(t). m ), where X(t) m )={x n (t m )|m=1,2,…,M}(n=1,2,…,N). For example, N=4*48*168, t={t1,t2,…,t M}, then X(t) is expressed as t1~t M The channel time series of 4*48*168 data points at various times, i.e., the observed data is X={X(t1),X(t2),…,X(t3)} M This forms an N x M matrix. In some feasible implementations, t i and t i+1 The time difference represented is a constant (i = 1, 2, ..., M-1). For example, t i and t i+1 The time difference represented is 1 millisecond.

[0068] 202. Predict the initialization parameters of the device.

[0069] In this embodiment of the application, the initialization parameters may include the number of time steps to be predicted, L-1, that is, the prediction device can predict time t. M+1 ,t M+2 ,…,t M+L-1 The channel time series, i.e., X(t) M+1 ), X(t) M+2 ), ..., X(t) M+L-1 ), that is, X(t) M+p-1(p = 2, ..., L). In some possible implementations, the initialization parameters can also be Gaussian and the initial signal strength parameter σ of the function. 2 f and the initial length scale parameter l, where l = {l1, l2, ..., l N}, l in l n X in X(t) n (t) One-to-one correspondence. It should be noted that the signal strength parameter σ 2 f The initial length scale parameter l can uniquely determine a Gaussian kernel function.

[0070] For example, the observed data is X = {X(t1), X(t2), ..., X(t...}}. M )}, where X(t) m )={x1(t m ),x1(t m ),…,x N (t m (m=1,2,…,M), for example, the observation length can be M=64, and N can be 4*48*168. The number of time steps to be predicted is L-1=24, then the predicted time is t. M+1 ,t M+2 ,…,t M+L-1 That is, t 65 ,t 66 ,…,t 88 .

[0071] For example, the initial signal strength parameter is set to σ. 2 f =1500, initial length scale parameter l = {l1,l2,…,l N} = uniform(σ 2 f Let l = {l1, l2, ..., ln}, where uniform(·) represents a function of uniform distribution. It should be noted that l = {l1, l2, ..., ln} N} and X(t m )={X1(t m ),X2(t m ),…,X N (t m One-to-one correspondence, i.e., l n Corresponding to X n (t m ), m=1,2,…,M; n=1,2,…,N.

[0072] 203. Prediction equipment based on X(t) m )={x n (tm Construct a mapping Ψ for each m = 1, 2, ..., M (n = 1, 2, ..., N). p , among which, Ψ p (X(t m ))=x n (t m+p-1 (p = 2, 3, ..., L), where L-1 is the number of time steps to be predicted.

[0073] For example, the prediction device is based on X(t) m )={x n (t m Construct the space-time information (STI) equations for |m=1,2,…,M} (n=1,2,…,N), where the STI equations can be expressed as follows:

[0074]

[0075] It should be noted that the STI equation is used to pass through Ψ p Used to convert X(t) m Mapping to x n (t m+p-1 ), that is, Ψ p (X(t m ))=x n (t m+p-1 (p = 2, 3, ..., L).

[0076] For example: Ψ1(X(t1))=x n (t1), Ψ1(X(t2))=x n (t2), …, Ψ1(X(t) M ))=x n (t M ); Ψ2(X(t1))=x n (t2), Ψ2(X(t2))=x n (t3), …, Ψ2(X(t) M ))=x n (t M+1 );……;Ψ L (X(t1))=x n (t L ), Ψ L (X(t2))=x n (t L+1 ), ..., Ψ L (X(t M ))=x n (t M+L-1 ), where n = 1, 2, ..., N.

[0077] In some possible implementations, Ψ p (p = 1, 2, ..., L) can be obtained by fitting a Gaussian kernel function. That is, the prediction device can be based on X(t) m (m=1,2,…,M) and Gaussian process regression (GPR) yield Ψ p (p = 1, 2, ..., L).

[0078] 204. Prediction equipment uses X(t) m (m=1,2,…,M-p+1) as input, with x n (t m+p-1 Using (m=1,2,…,M-p+1) as the output, we perform a Gaussian process regression (GPR) fitting to obtain Ψ. p (p = 2, 3, ..., L).

[0079] In the first line of the above STI equation, x n (t1)~x n (t M If t is known, then X(t) m (m=1,2,…,M) as input, with x n (t1)~x n (t M As the output, we perform Gaussian process regression (GPR) fitting to obtain Ψ1, as well as the signal intensity parameter and length scale parameter corresponding to Ψ1.

[0080] In the second line of the STI equation above, x n (t2)~x n (t M If t is known, then X(t) m (m=2,3,…,M) as input, with x n (t2)~x n (t M As the output, we perform Gaussian process regression (GPR) fitting to obtain Ψ2, and the corresponding signal intensity parameters and length scale parameters of Ψ1.

[0081] In the third line of the STI equation above, x n (t3)~x n (t M If t is known, then X(t) m (m=3,4,…,M) as input, with x n (t3)~x n (t MAs the output, a Gaussian process regression (GPR) is performed to obtain Ψ3, as well as the signal intensity parameter and length scale parameter corresponding to Ψ3.

[0082] In the p-th row of the above STI equation, x n (t p )~x n (t M If t is known, then X(t) m (m = p, p+1, ..., M) as input, with x n (t p )~x n (t M As the output, Ψ is fitted to Gaussian process regression (GPR) to obtain Ψ. p , and Ψ p The corresponding signal strength parameters and length scale parameters. Where p = 1, 2, ..., L, ultimately yielding Ψ1~Ψ L .

[0083] 205. The prediction device will X(t) m Input Ψ(m=M-p+2,M-p+3,…,M) respectively. p (p = 2, 3, ..., L), we get x n (t m+p-1 Predicted values ​​for (n = 1, 2, ..., N; m = 1, 2, ..., M; p = 2, 3, ..., L).

[0084] The predictive device obtains the mapping Ψ1~Ψ L Then, the X(t) can be... m (m = Mp-2, Mp-1, ..., M) Input Ψ1 to Ψ respectively L , to obtain x n (t M+p-1 (p=2,3,…,L), thus achieving the x in the p-th row of the STI equation. n (t M+p-1 Prediction of (p=2,3,…,L).

[0085] For example, in the second line of the STI equation above, after obtaining Ψ2 through step 204, the prediction device can then predict X(t). M Input Ψ2, get Ψ2(X(t) M ))=x n (t M+1 This means that the prediction of x has been achieved. n (t M+1 ).

[0086] In the third line of the STI equation above, after obtaining Ψ3 through step 204, the prediction device can then predict X(t).M-1 ) and X(t M Input Ψ3, get Ψ3(X(t) M-1 ))=x n (t M+1 ), Ψ3(X(t) M ))=x n (t M+2 This means that the prediction of x has been achieved. n (t M+1 ) and x n (t M+2 ).

[0087] In the p-th row of the above STI equation, Ψ is obtained through step 204. p Then, the prediction device can determine X(t) m (m=1,2,…,M-p+1) as input, with x n (t m (m=p,p+1,…,M;n=1,2,…,N) is used as the output, and Ψ is obtained by fitting based on GPR. p (p = 1, 2, ..., L).

[0088] 206. Predictive equipment extracts feature variables.

[0089] In this embodiment, the extracted feature variables can be used as training data to further optimize Ψ. p (p = 2, 3, ..., L).

[0090] In some feasible implementations, the predictive device can obtain Ψ p Length dimension parameter l p , where l p ={l p1 ,l p2 ,…,l pN}, l pn With x n (t m ) Corresponding. Then, the prediction device takes l p ={l p1 ,l p2 ,…,l pN Find the largest P values ​​in} to obtain l p '={l p '1,l p '2,…,l p ' P Next, the predictive device determines l. p '={l p '1,l p '2,…,l p ' Px corresponding to each value in} n (t m ), thus obtaining the characteristic variable X p '(t m )={x p '1(t m ),x p '2(t m ),…,x p ' P (t m )}.

[0091] For example, N = 4 * 48 * 168 = 32256, then P = └N * 10% ┘ = 3225, and the characteristic variable is X. p '(t m )={x p '1(t m ),x p '2(t m ),…,x p ' 3225 (t m )}.

[0092] 207. Predictive equipment with Ψ p (X p '(t m The output is n = 1, 2, ..., N, with the characteristic variable X. p '(t m Using (n = 1, 2, ..., N) as input, fit the GPR to obtain the optimized Ψ. p (p = 1, 2, ..., L).

[0093] In some feasible implementations, the prediction device can first set the number of optimization process runs required to fit the GPR and the maximum number of iterations for the maximum likelihood estimation optimization process. For example, the number of optimization process runs n_run = 10, and the maximum number of iterations for the maximum likelihood estimation optimization process iter = 20000. Then, the prediction device can then use the feature variable X... p '(t m )={x p '1(t m ),x p '2(t m ),…,x p ' P (t m Enter Ψ respectively p (p = 1, 2, ..., L), we obtain Ψ p (X p '(t m Then, the predictive device uses Ψ p (Xp '(t m The output is n = 1, 2, ..., N, with the characteristic variable X. p '(t m Using (n = 1, 2, ..., N) as input, fit the GPR to obtain the optimized Ψ'. p (p = 1, 2, ..., L).

[0094] 208. Prediction equipment based on Ψ' p (p = 2, 3, ..., L) Predict x n (t M+p-1 ).

[0095] In some feasible implementations, the prediction device can predict x in one go. n (t M+1 )~x n (t M+L-1 ).

[0096] For example, the prediction device can predict X(t) m (m = M - p + 2, M - p + 3, ..., M) respectively input the Ψ p (p = 2, 3, ..., L), respectively, we obtain Ψ p x corresponding to (p = 2, 3, ..., L) n (t M+p-1 ), and calculate Ψ p x corresponding to (p = 2, 3, ..., L) n (t M+p-1 The average of ) is used as x n (t M+p-1 The predicted value of x is thus achieved. n (t M+p-1 Prediction of (p=2,3,…,L).

[0097] For example, for x n (t M+L-1 The prediction device will predict X(t) m (m=M-L+2,M-L+1,…,M) Input the optimized Ψ' respectively p (p = 2, 3, ..., L), thus obtaining Ψ' p (X(t m (p = 2, 3, ..., L), then calculate the average value [Ψ'2(X(t)] m ))+Ψ'3(X(t m ))+…+Ψ' L (X(t m ))] / (L-1), using this as x n (t M+L-1 The predicted value of ).

[0098] In some possible implementations, the prediction device can first predict x. n (t M+p-1 ), then predict x n (t M+p ), p=2,3,…,L.

[0099] For example, the prediction device may perform the following steps:

[0100] S1. Let p = 2;

[0101] S2. With X(t) m (m=1,2,…,M-p'+1) as input, with x n (t m+p’-1 Using (m=1,2,…,M-p'+1) as the output, fit it to GPR to obtain Ψ p Where p' = p, p+1, p+2, ..., L;

[0102] S3. X(t) m (m = M - (p - 2), M - (p - 1), ..., M) input Ψ respectively p ,Ψ p+1 ,…,Ψ L We get L-p+1 x n (t M+p-1 );

[0103] S4. Calculate L-p+1 x n (t M+p-1 The average of ) is used as x n (t M+p-1 The predicted value of );

[0104] S5. If p is not equal to L, then p = p + 1 and execute step S2; otherwise, the execution is complete.

[0105] For example, the prediction device first predicts x n (t M+1 Then predict x. n (t M+2 ), predicting up to x n (t M+L-1 ).

[0106] The following example illustrates the prediction of channel time series in communication between terminal devices and base stations using 5G technology. Steps 201-208 above demonstrate the implementation effect of multi-step prediction of channel time series.

[0107] For example, such as Figure 2-2 and Figure 2-3 As shown, the horizontal axis represents the time axis, and the vertical axis represents the predicted values. The length of the observed data is M = 64, i.e., X(t1) ~ X(t2). 64 The observed data are X(t), and the number of time steps to be predicted is L-1 = 24, meaning the data to be predicted is X(t). 65 )~X(t 88 ).in, Figure 2-2 The vertical axis represents the real part of the predicted value. Figure 2-3 The vertical axis represents the imaginary part of the predicted value.

[0108] in, Figure 2-2 and Figure 2-3 The data for the 64 time points before the dashed line are observed data, and the data after that are the prediction results. Solid dots represent the actual future values, and hollow dots represent the predicted values ​​obtained by application. The correlation coefficients between the actual values ​​and the predicted values ​​reached 0.843 and 0.701, respectively.

[0109] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0110] To facilitate better implementation of the above-described solutions in the embodiments of this application, related apparatus for implementing the above-described solutions is also provided below.

[0111] Please see Figure 3 As shown in the embodiment of this application, a channel time series prediction device 300 may include:

[0112] The transceiver module 301 is used to acquire the channel time series at M time points to obtain X(t) m )={x n (t m )|m=1,2,…,M}(n=1,2,…,N), where N is the number of channel time series within a given time moment, and M and N are both positive integers;

[0113] Processing module 302, used for processing based on the X(t) m )={x n (t m Construct a mapping Ψ for each m = 1, 2, ..., M (n = 1, 2, ..., N). p , among which, Ψ p (X(tm ))=x n (t m+p-1 (p = 2, 3, ..., L), where L-1 is the number of time steps to be predicted;

[0114] The processing module 302 is further configured to use X(t) m (m=1,2,…,M-p+1) as input, with x n (t m+p-1 Using (m=1,2,…,M-p+1) as the output, we perform a Gaussian process regression (GPR) fitting to obtain Ψ. p (p = 2, 3, ..., L);

[0115] The processing module 302 is further configured to process X(t) m (m = M - p + 2, M - p + 3, ..., M) respectively input the Ψ p (p = 2, 3, ..., L), we get x n (t m+p-1 Predicted values ​​for (n = 1, 2, ..., N; m = 1, 2, ..., M; p = 2, 3, ..., L).

[0116] In some feasible implementations, the processing module 302 is specifically used for:

[0117] Construct the Space-Time Information (STI) equation, which is expressed as:

[0118]

[0119] In some feasible implementations, the processing module 302 is further configured to: obtain the Ψ p Length dimension parameter l p , where l p ={l p1 ,l p2 ,…,l pN}, l pn With x n (t m )correspond;

[0120] Take the l p ={l p1 ,l p2 ,…,l pN Find the largest P values ​​in} to obtain l p '={l p '1,l p '2,…,l p ' P};

[0121] Determine the l p'={l p '1,l p '2,…,l p ' P x corresponding to each value in} n (t m ), thus obtaining the characteristic variable X p '(t m )={x p '1(t m ),x p '2(t m ),…,x p ' P (t m )};

[0122] With Ψ p (X p '(t m (n = 1, 2, ..., N) is the output, and the characteristic variable X is... p '(t m Using (n = 1, 2, ..., N) as input, fit the GPR to obtain the optimized Ψ. p (p = 1, 2, ..., L).

[0123] In some feasible implementations, the processing module 302 is specifically used for:

[0124] X(t) m (m = M - p + 2, M - p + 3, ..., M) respectively input the Ψ p (p = 2, 3, ..., L), respectively, we obtain Ψ p x corresponding to (p = 2, 3, ..., L) n (t M+p-1 );

[0125] Calculate the Ψ p x corresponding to (p = 2, 3, ..., L) n (t M+p-1 The average value of ) is used as the x n (t M+p-1 The predicted value of ).

[0126] In some feasible implementations, the processing module 302 is specifically used for:

[0127] S1. Let p = 2;

[0128] S2. With X(t) m (m=1,2,…,M-p'+1) as input, with x n (t m+p’-1Using (m=1,2,…,M-p'+1) as the output, fit it to GPR to obtain Ψ p’ Where p' = p, p+1, p+2, ..., L;

[0129] S3. X(t) m (m = M - (p - 2), M - (p - 1), ..., M) input Ψ respectively p ,Ψ p+1 ,…,Ψ L We get L-p+1 x n (t M+p-1 );

[0130] S4. Calculate L-p+1 x n (t M+p-1 The average of ) is used as x n (t M+p-1 The predicted value of );

[0131] S5. If p is not equal to L, then p = p + 1 and execute step S2; otherwise, the execution is complete.

[0132] It should be noted that the information interaction and execution process between the modules / units of the above-mentioned device are based on the same concept as the method embodiments of this application, and the resulting technical effects are the same as those of the method embodiments of this application. For details, please refer to the description in the method embodiments shown above in this application, and will not be repeated here.

[0133] This application also provides a computer storage medium storing a program that performs some or all of the steps described in the above method embodiments.

[0134] The following describes another communication device provided in the embodiments of this application. Please refer to [link to relevant documentation]. Figure 4 As shown, the communication device 400 includes:

[0135] The system includes a receiver 401, a transmitter 402, a processor 403, and a memory 404. In some embodiments of this application, the receiver 401, transmitter 402, processor 403, and memory 404 may be connected via a bus or other means. Figure 4 Taking the example of a connection between China and Israel via a bus.

[0136] Memory 404 may include read-only memory and random access memory, and provides instructions and data to processor 403. A portion of memory 404 may also include non-volatile random access memory (NVRAM). Memory 404 stores operating systems and operating instructions, executable modules or data structures, or subsets thereof, or extended sets thereof. The operating instructions may include various operating instructions for implementing various operations. The operating system may include various system programs for implementing various basic business functions and handling hardware-based tasks.

[0137] Processor 403 controls the operation of communication device 400. Processor 403 can also be called a central processing unit (CPU). In specific applications, the various components of communication device 400 are coupled together through a bus system. This bus system includes not only a data bus but also a power bus, control bus, and status signal bus. However, for clarity, all buses are referred to as a bus system in the diagram.

[0138] The methods disclosed in the embodiments of this application can be applied to processor 403, or implemented by processor 403. Processor 403 can be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in processor 403 or by instructions in the form of software. The processor 403 can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or can be executed by a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 404. Processor 403 reads the information in memory 404 and, in conjunction with its hardware, completes the steps of the above method.

[0139] The receiver 401 can be used to receive input digital or character information and generate signal inputs related to relevant settings and function control. The transmitter 402 may include display devices such as a display screen and can be used to output digital or character information through an external interface.

[0140] In this embodiment of the application, processor 403 is used to execute the aforementioned channel time series prediction method.

[0141] In another possible design, when the prediction device 300 or communication device 400 is a chip, it includes a processing unit and a communication unit. The processing unit may be, for example, a processor, and the communication unit may be, for example, an input / output interface, pins, or circuits. The processing unit can execute computer-executed instructions stored in a storage unit to cause the chip within the terminal to execute the wireless reporting information transmission method described in any of the first aspects above. Optionally, the storage unit can be a storage unit within the chip, such as a register or cache. Alternatively, the storage unit can be a storage unit located outside the chip within the terminal, such as read-only memory (ROM) or other types of static storage devices capable of storing static information and instructions, such as random access memory (RAM).

[0142] The processor mentioned above can be a general-purpose central processing unit, a microprocessor, an ASIC, or one or more integrated circuits used to control the execution of the program described above.

[0143] It should also be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. In addition, in the device embodiment drawings provided in this application, the connection relationship between modules indicates that they have a communication connection, which can be implemented as one or more communication buses or signal lines.

[0144] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware, or it can be implemented by special-purpose hardware including application-specific integrated circuits, special-purpose CPUs, special-purpose memory, special-purpose components, etc. Generally, any function performed by a computer program can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be diverse, such as analog circuits, digital circuits, or special-purpose circuits. However, for this application, software program implementation is more often the preferred implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium, such as a computer floppy disk, USB flash drive, mobile hard disk, ROM, RAM, magnetic disk, or optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0145] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product.

[0146] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

Claims

1. A method for predicting channel time series, characterized in that, include: Obtain the channel time series at M time points to get X(t) m )={x n (t m )|m=1,2,…,M}(n=1,2,…,N), where N is the number of channel time series within a given time moment, and M and N are both positive integers; Based on the X(t) m )={x n (t m Construct a mapping Ψ for each m=1,2,…,M (n=1,2,…,N). p , among which, Ψ p (X(t m ))=x n (t m+p-1 (p=2,3,…,L), where L-1 is the number of time steps to be predicted; With X(t) m (m=1,2,…,M-p+1) as input, with x n (t m+p-1 Using (m=1,2,…,M-p+1) as the output, we perform a Gaussian process regression (GPR) fitting to obtain Ψ. p (p=2,3,…,L); X(t) m (m=M-p+2,M-p+3,…,M) Input the Ψ respectively p (p=2,3,…,L), obtain x n (t m+p-1 Predicted values ​​for (n=1,2,…,N;m=1,2,…,M;p=2,3,…,L).

2. The method according to claim 1, characterized in that, The basis of X(t) m )={x n (t m Construct a mapping Ψ for each m=1,2,…,M (n=1,2,…,N). p include: Construct the Space-Time Information (STI) equation, which is expressed as: 。 3. The method according to claim 1, characterized in that, The basis of X(t) m )={x n (t m Construct a mapping Ψ for each m=1,2,…,M (n=1,2,…,N). p Following that, it also includes: Obtain the Ψ p Length dimension parameter l p , where l p ={l p1 ,l p2 ,…,l pN }, l pn With x n (t m )correspond; Take the l p ={l p1 ,l p2 ,…,l pN Find the largest P values ​​in} to obtain l p '={l p '1, l p '2, …, l p ' P }; Determine the l p '={l p '1, l p '2, …, l p ' P x corresponding to each value in} n (t m ), thus obtaining the characteristic variable X p '(t m )={x p '1(t m ), x p '2(t m ), …, x p ' P (t m )}; With Ψ p (X p '(t m (n=1,2,…,N) is the output, and the characteristic variable X is the output. p '(t m Using (n=1,2,…,N) as input, fit the GPR to obtain the optimized Ψ. p (p=1,2,…,L).

4. The method according to any one of claims 1-3, characterized in that, The X(t) m (m=M-p+2,M-p+3,…,M) Input the Ψ respectively p (p=2, 3, ...,L), to obtain x n (t m+p-1 (n=1,2,…,N; m = 1, 2, ..., M; The predicted values ​​for p=2,…,L) include: X(t) m (m=M-p+2,M-p+3,…,M) respectively input the Ψ p (p=2,3,…,L), respectively obtain Ψ p x corresponding to (p=2,3,…,L) n (t M+p-1 ); Calculate the Ψ p x corresponding to (p=2,3,…,L) n (t M+p-1 The average value of ) is used as the x n (t M+p-1 The predicted value of ).

5. The method according to any one of claims 1-3, characterized in that, The X(t) m (m=M-p+2,M-p+3,…,M) Input the Ψ respectively p (p=2, 3, ...,L), to obtain x n (t m+p-1 (n=1,2,…,N; m = 1, 2, ..., M; The predicted values ​​for p=2,…,L) include: S1. Let p=2; S2. With X(t) m (m=1,2,…,M-p'+1) as input, with x n (t m+p’-1 Using (m=1,2,…,M-p'+1) as the output, fit it to GPR to obtain Ψ. p’ Where p' = p, p+1, p+2, ..., L; S3. X(t) m (m=M-p+2,M-p+3,…,M) Input Ψ respectively p ,Ψ p+1 ,…,Ψ L We get L-p+1 x n (t M+p-1 ); S4. Calculate L-p+1 x n (t M+p-1 The average of ) is used as x n (t M+p-1 The predicted value of ); S5. If p is not equal to L, then p = p + 1 and execute step S2; otherwise, the execution is complete.

6. A channel time series prediction device, characterized in that, include: The transceiver module is used to acquire the channel time series at M time points to obtain X(t) m )={x n (t m )|m=1,2,…,M}(n=1,2,…,N), where N is the number of channel time series within a given time moment, and M and N are both positive integers; Processing module, used for processing based on X(t) m )={x n (t m Construct a mapping Ψ for each m=1,2,…,M (n=1,2,…,N). p , among which, Ψ p (X(t m ))=x n (t m+p-1 (p=2,3,…,L), where L-1 is the number of time steps to be predicted; The processing module is also used to use X(t) m (m=1,2,…,M-p+1) as input, with x n (t m+p-1 Using (m=1,2,…,M-p+1) as the output, we perform a Gaussian process regression (GPR) fitting to obtain Ψ. p (p=2,3,…,L); The processing module is also used to process X(t) m (m=M-p+2,M-p+3,…,M) Input the Ψ respectively p (p=2,3,…,L), obtain x n (t m+p-1 Predicted values ​​for (n=1,2,…,N;m=1,2,…,M;p=2,3,…,L).

7. The prediction device according to claim 6, characterized in that, The processing module is specifically used for: Construct the Space-Time Information (STI) equation, which is expressed as: 。 8. The prediction device according to claim 6, characterized in that, The processing module is further configured to: Obtain the Ψ p Length dimension parameter l p , where l p ={l p1 ,l p2 ,…,l pN }, l pn With x n (t m )correspond; Take the l p ={l p1 ,l p2 ,…,l pN Find the largest P values ​​in} to obtain l p '={l p '1, l p '2, …, l p ' P }; Determine the l p '={l p '1, l p '2, …, l p ' P x corresponding to each value in} n (t m ), thus obtaining the characteristic variable X p '(t m )={x p '1(t m ), x p '2(t m ), …, x p ' P (t m )}; With Ψ p (X p '(t m (n=1,2,…,N) is the output, and the characteristic variable X is the output. p '(t m Using (n=1,2,…,N) as input, fit the GPR to obtain the optimized Ψ. p (p=1,2,…,L).

9. The prediction device according to any one of claims 6-8, characterized in that, The processing module is specifically used for: X(t) m (m=M-p+2,M-p+3,…,M) respectively input the Ψ p (p=2,3,…,L), respectively obtain Ψ p x corresponding to (p=2,3,…,L) n (t M+p-1 ); Calculate the Ψ p x corresponding to (p=2,3,…,L) n (t M+p-1 The average value of ) is used as the x n (t M+p-1 The predicted value of ).

10. The prediction device according to any one of claims 6-8, characterized in that, The processing module is specifically used for: S1. Let p=2; S2. With X(t) m (m=1,2,…,M-p'+1) as input, with x n (t m+p’-1 Using (m=1,2,…,M-p'+1) as the output, fit it to GPR to obtain Ψ. p’ Where p' = p, p+1, p+2, ..., L; S3. X(t) m (m=M-p+2,M-p+3,…,M) Input Ψ respectively p ,Ψ p+1 ,…,Ψ L We get L-p+1 x n (t M+p-1 ); S4. Calculate L-p+1 x n (t M+p-1 The average of ) is used as x n (t M+p-1 The predicted value of ); S5. If p is not equal to L, then p = p + 1 and execute step S2; otherwise, the execution is complete.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that causes a computer device to perform the method as described in any one of claims 1-5.

12. A computer program product, characterized in that, The computer program product includes computer-executable instructions stored in a computer-readable storage medium; at least one processor of the device reads the computer-executable instructions from the computer-readable storage medium, and the at least one processor executes the computer-executable instructions to cause the device to perform the method as described in any one of claims 1-5.

13. A communication device, characterized in that, The communication device includes at least one processor, memory, and communication interface; The at least one processor is coupled to the memory and the communication interface; The memory is used to store instructions, the processor is used to execute the instructions, and the communication interface is used to communicate with other communication devices under the control of the at least one processor; When the instruction is executed by the at least one processor, it causes the at least one processor to perform the method as described in any one of claims 1-5.

14. A chip system, characterized in that, The chip system includes a processor and a memory, the memory and the processor being interconnected via a circuit, the memory storing instructions, and the processor being used to execute the method as described in any one of claims 1-5.

Images