Channel evaluation method and related device

Abstract

The invention discloses a channel assessment method and a related device, and relates to the technical field of communication, and the method comprises the steps: initializing an auxiliary manifold natural parameter and statistical information, and setting a damping value; executing the following cycle steps until the target manifold parameter change rate or the statistical information change rate is smaller than a preset threshold; determining m projection of the current round according to the statistical information and the auxiliary manifold natural parameters of the previous round; determining the natural parameters of the auxiliary manifold of the current round according to the m projection of the current round, the natural parameters of the auxiliary manifold of the previous round and the damping value; determining target manifold natural parameters of the round according to the auxiliary manifold natural parameters of the round; according to the statistical information of the previous round and the target manifold natural parameters of the round, determining the statistical information of the round; and finally, according to the target manifold natural parameters of the last round and the statistical information of the last round, a beam domain channel assessment result is determined, and the channel assessment result has higher accuracy.

Description

Technical Field

[0001] The present application relates to the field of communication technology, and in particular to a channel assessment method and related devices. Background Art

[0002] Massive Multiple-Input Multiple-Output (MIMO) technology is regarded as one of the core key technologies of the fifth-generation cellular communication system. In a massive MIMO system, a base station can serve a large number of users at the same time by deploying a large number of antennas, providing high-speed data transmission services under the same time and frequency resources. This design not only greatly improves the system capacity, but also significantly improves energy efficiency. On the other hand, orthogonal frequency division multiplexing (OFDM) technology, as a multi-carrier modulation scheme, effectively mitigates the impact of frequency selective fading in broadband wireless communications. At present, the combination of massive MIMO and OFDM technology, that is, massive MIMO-OFDM technology, has played a vital role in 5G communication systems and has received great attention in the development of future sixth-generation (6G) systems.

[0003] In massive MIMO-OFDM systems, the accuracy of channel estimation is critical to system performance. To achieve efficient channel estimation, a pilot-based method is usually used, that is, the transmitter sends a pilot signal periodically, and the receiver estimates the channel state information CSI based on the received pilot signal, that is, obtains the posterior information of the channel. When the prior distribution of the channel parameters follows a Gaussian distribution, its posterior distribution will also be Gaussian. At this time, the posterior information of the channel parameters can be characterized by the mean and covariance matrix. However, in massive MIMO-OFDM systems, due to the extremely high channel dimension, calculating these posterior statistics becomes extremely challenging. Traditional estimation methods, such as the minimum mean square error (MMSE) estimator, often have unacceptable computational complexity due to the need to process large-scale matrix inversion operations.

[0004] The information geometry IG channel estimation algorithm has the advantage of low complexity. Based on the statistical channel model of space-frequency SF beams, the IG algorithm transforms the channel estimation process into an iterative projection process, and it has been proved that the IG algorithm can accurately estimate the posterior mean at its stable point. In order to further reduce the computational complexity, a simplified information geometry SIG algorithm is proposed based on the IG algorithm. The SIG algorithm utilizes the constant amplitude characteristics of the measurement matrix elements and derives several key properties of natural parameters. These properties simplify the iterative process of the IG algorithm, thus forming the SIG algorithm. It is worth noting that the posterior mean estimated by the SIG algorithm is asymptotically optimal.

[0005] However, both the IG algorithm and the SIG algorithm need to obtain accurate SF beam-domain statistical CSI before instantaneous channel estimation, which is often difficult to achieve in reality. The traditional statistical CSI acquisition method uses the estimated instantaneous channel to calculate the statistical CSI. Therefore, before channel estimation, it is impossible to obtain more accurate statistical CSI, which limits the practicality of the SIG algorithm and reduces the channel estimation performance. Summary of the invention

[0006] In view of this, the main purpose of the present application is to provide a channel assessment method, apparatus, device, program product and readable storage medium to improve channel assessment performance.

[0007] The first aspect of the present application provides a channel assessment method, the method comprising:

[0008] Initialize the auxiliary manifold natural parameters and statistical information, and set the damping value;

[0009] Execute the following loop steps until the target manifold parameter change rate or the statistical information change rate is less than a preset threshold: determine the m-projection of this round according to the statistical information of the previous round and the auxiliary manifold natural parameters of the previous round; determine the auxiliary manifold natural parameters of this round according to the m-projection of this round, the auxiliary manifold natural parameters of the previous round and the damping value; determine the target manifold natural parameters of this round according to the auxiliary manifold natural parameters of this round; determine the statistical information of this round according to the statistical information of the previous round and the target manifold natural parameters of this round; determine the target manifold parameter change rate based on the target manifold natural parameters of this round and the target manifold natural parameters of the previous round; determine the statistical information change rate based on the statistical information of this round and the statistical information of the previous round;

[0010] In response to the target manifold parameter change rate or the statistical information change rate being less than a preset threshold, the mean of the target manifold is determined according to the last round of target manifold natural parameters and the last round of statistical information, and the mean of the target manifold is used as the beam domain channel evaluation result.

[0011] In some implementations of the first aspect of the present application, the method further includes:

[0012] The beam domain channel estimation result is converted into the space-frequency domain channel estimation result.

[0013] In some implementations of the first aspect of the present application, the auxiliary manifold natural parameter includes: an auxiliary manifold first-order average natural parameter and an auxiliary manifold second-order average natural parameter;

[0014] According to the statistical information of the previous round and the natural parameters of the auxiliary manifold of the previous round, the m-projection of this round is determined, including:

[0015] The first m-projection of this round is determined according to the statistical information of the previous round and the first-order average natural parameters of the auxiliary manifold of the previous round; the second m-projection of this round is determined according to the statistical information of the previous round and the second-order average natural parameters of the auxiliary manifold of the previous round.

[0016] In some implementations of the first aspect of the present application, determining the auxiliary manifold natural parameters of the current round according to the m-projection of the current round, the auxiliary manifold natural parameters of the previous round, and the damping value includes:

[0017] The first-order average natural parameters of the auxiliary manifold of this round are determined according to the first m-projection of this round, the first-order average natural parameters of the auxiliary manifold of the previous round and the damping value; the second-order average natural parameters of the auxiliary manifold of this round are determined according to the second m-projection of this round, the second-order average natural parameters of the auxiliary manifold of the previous round and the damping value.

[0018] In some implementations of the first aspect of the present application, the target manifold natural parameters include: first-order natural parameters of the target manifold and second-order natural parameters of the target manifold;

[0019] According to the natural parameters of the auxiliary manifold of this round, the natural parameters of the target manifold of this round are determined, including:

[0020] The first-order natural parameters of the target manifold of this round are determined according to the first-order average natural parameters of the auxiliary manifold of this round; the second-order natural parameters of the target manifold of this round are determined according to the second-order average natural parameters of the auxiliary manifold of this round.

[0021] In some implementations of the first aspect of the present application, determining the statistical information of this round according to the statistical information of the previous round and the natural parameters of the target manifold of this round includes:

[0022] The statistical information of this round is determined according to the statistical information of the previous round, the first-order natural parameters of the target manifold of this round, and the second-order natural parameters of the target manifold of this round.

[0023] In some implementations of the first aspect of the present application, determining the mean of the target manifold according to the natural parameters of the target manifold in the last round and the statistical information of the Tth round includes:

[0024] The mean of the target manifold is determined according to the first-order natural parameters of the target manifold in the last round, the second-order natural parameters of the target manifold in the last round, and the statistical information of the Tth round.

[0025] A second aspect of the present application provides a channel assessment device, the device comprising:

[0026] Initialization module, used to initialize the natural parameters and statistical information of the auxiliary manifold and set the damping value;

[0027] A loop execution module is used to execute the following loop steps until the target manifold parameter change rate or the statistical information change rate is less than a preset threshold: determine the m-projection of this round according to the statistical information of the previous round and the auxiliary manifold natural parameters of the previous round; determine the auxiliary manifold natural parameters of this round according to the m-projection of this round, the auxiliary manifold natural parameters of the previous round and the damping value; determine the target manifold natural parameters of this round according to the auxiliary manifold natural parameters of this round; determine the statistical information of this round according to the statistical information of the previous round and the target manifold natural parameters of this round; determine the target manifold parameter change rate based on the target manifold natural parameters of this round and the target manifold natural parameters of the previous round; determine the statistical information change rate based on the statistical information of this round and the statistical information of the previous round;

[0028] The channel evaluation module is used to determine the mean of the target manifold according to the natural parameters of the target manifold of the last round and the statistical information of the last round in response to the target manifold parameter change rate or the statistical information change rate being less than a preset threshold, and use the mean of the target manifold as the beam domain channel evaluation result.

[0029] A third aspect of the present application provides a channel assessment device, which includes a memory and a processor, and the processor is used to execute a program stored in the memory to run a channel assessment method as described in any one of the first aspects.

[0030] A fourth aspect of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of any one of the channel assessment methods in the first aspect.

[0031] A fifth aspect of the present application provides a readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the various steps of any channel assessment method in the first aspect are implemented.

[0032] The technical solution provided by this application has the following beneficial effects:

[0033] In the technical solution provided in the present application, first, initialization is performed to provide initial conditions for subsequent iterations; then a loop step is entered to gradually update the natural parameters of the auxiliary manifold and the natural parameters of the target manifold until the change rate of the target manifold parameters or the change rate of the statistical information is less than a preset threshold.

[0034] Specifically, in the loop step: first, the m-projection of this round is determined according to the statistical information of the previous round and the natural parameters of the auxiliary manifold of the previous round; the complexity of parameter calculation is reduced by calculating the projection. Then, the natural parameters of the auxiliary manifold of this round are determined according to the m-projection of this round, the natural parameters of the auxiliary manifold of the previous round and the damping value; in each round of iteration, the gradual update under the control of the damping value can avoid excessive fluctuations, and at the same time, the stability and convergence of the estimation are ensured by the parameter update after the projection, which helps to improve the accuracy of channel estimation. After that, the natural parameters of the target manifold of this round are determined according to the natural parameters of the auxiliary manifold of this round; by gradually determining the natural parameters of the target manifold in the loop process, not only the calculation is simplified, but also the channel estimation is made more accurate by correcting the natural parameters in each iteration. Finally, new statistical information is determined according to the statistical information of the previous round and the natural parameters of the target manifold of this round. The updated natural parameters of the target manifold are used to update and correct the statistical information, so that the update of each round of statistical information further corrects the estimation of the channel, ensuring that the algorithm can converge to the real channel state in the step-by-step iteration. At the same time, the updated statistical information continues to be used in the next round of iteration, avoiding the reliance on previously acquired statistical CSI.

[0035] After the target manifold parameter change rate or the statistical information change rate is less than a preset threshold, the mean of the target manifold is determined as the final channel evaluation result according to the final target manifold natural parameters and statistical information.

[0036] It can be seen that in the above technical solution, the present application does not need to know the beam domain statistics in advance, but updates the statistical information by updating the natural parameters of the target manifold during the cyclic iteration process, and updating the natural parameters of the target manifold only involves multiplication operations of vectors and scalars, so the computational complexity is relatively low. It can be seen that the complexity of the technical solution provided by the present application is not significantly improved compared with the prior art, and it solves the problem that the SIG algorithm needs to know the statistical information in advance without increasing the complexity of the algorithm, which significantly improves the practicality of the SIG algorithm. At the same time, the final channel evaluation result has higher accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A schematic diagram of a flow chart of a channel assessment method provided in an embodiment of the present application;

[0038] Figure 2 A schematic diagram of a flow chart of another channel assessment method provided in an embodiment of the present application;

[0039] Figure 3 A performance comparison chart of the EM-SIG algorithm and the SIG algorithm provided in the embodiments of the present application;

[0040] Figure 4A statistical information diagram estimated by the EM-SIG algorithm provided in an embodiment of the present application;

[0041] Figure 5 A comparison chart of the accuracy of the EM-SIG algorithm statistical information estimation provided in the embodiment of the present application;

[0042] Figure 6 A schematic diagram of the structure of a channel assessment device provided in an embodiment of the present application;

[0043] Figure 7 A schematic diagram of the structure of a channel assessment device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0045] Embodiments of the present application provide a channel assessment method, apparatus, device, program product, and readable storage medium.

[0046] See also Figure 1 As shown, an embodiment of the present application provides a channel assessment method, which may specifically include the following steps:

[0047] S101: Initialize the auxiliary manifold natural parameters and statistical information, and set the damping value.

[0048] In an embodiment of the present application, the auxiliary manifold is used to approximate some nonlinear terms in the posterior probability density function, and its characteristics are characterized by the auxiliary manifold natural parameters. The statistical information is the prior statistical information that characterizes the channel state during the channel estimation process. The damping value is used to control the update step size of each subsequent iteration to prevent data divergence during the iteration process.

[0049] The embodiment of the present application provides reasonable initial conditions for subsequent iterations by initializing natural parameters, ensuring that the subsequent cyclic iteration process will not deviate from the target. The setting of the damping value ensures that the subsequent cyclic iteration process can gradually converge to a stable channel estimation result.

[0050] S102: Determine whether the target manifold parameter change rate or the statistical information change rate in the current iteration is less than a preset threshold. If so, execute S108; if not, execute the loop steps of S103-S107.

[0051] In an embodiment of the present application, in each round of iteration, it will be judged whether the target manifold parameter rate of change or the statistical information rate of change is less than a preset threshold. If the target manifold parameter rate of change or the statistical information rate of change is less than the preset threshold, the final channel evaluation stage is entered, i.e., S108 is executed, otherwise S103-S107 are continued to be iteratively updated. By presetting a reasonable threshold, it is possible to ensure that the iteration terminates reasonably, avoid wasting computer resources, and also provide enough iteration times to converge to the optimal channel estimation result.

[0052] S103: Determine the m-projection of this round according to the statistical information of the previous round and the natural parameters of the auxiliary manifold of the previous round;

[0053] In an embodiment of the present application, m is projected as a distribution on the target manifold, and the distribution satisfies the minimum Kullback-Leibler (KL) divergence to the auxiliary manifold, thereby reducing the computational complexity and facilitating improving the operation rate in subsequent steps.

[0054] In addition, it can be understood that when this round is the first round, in S103, the statistical information of the previous round is the statistical information initialized in S101, and the auxiliary manifold natural parameters of the previous round are the auxiliary manifold natural parameters initialized in S101.

[0055] The embodiment of the present application calculates the m projections of the current round through the natural parameters and statistical information of the previous round, thereby introducing the projection operation, thereby reducing the complexity of the calculation and avoiding the computational overhead caused by directly processing high-dimensional matrices. At the same time, the average natural parameters of the auxiliary manifold are closer to the natural parameters of the target manifold, so that the estimation of the channel state gradually approaches the actual channel conditions.

[0056] S104: Determine the auxiliary manifold natural parameters of this round according to the m-projection of this round, the auxiliary manifold natural parameters of the previous round, and the damping value.

[0057] In an embodiment of the present application, the natural parameters of the auxiliary manifold of the current round are updated according to the m-projection, combined with the natural parameters of the auxiliary manifold of the previous round and the preset damping value. By combining the projection result and the estimation of the previous round, the natural parameters of the auxiliary manifold of the current round are optimized, so that the channel state estimation is more accurate. At the same time, the use of the damping value also ensures the smoothness and stability of the update process.

[0058] S105: Determine the target manifold natural parameters of this round according to the auxiliary manifold natural parameters of this round.

[0059] In the embodiments of the present application, the target manifold refers to the space that is ultimately desired to be optimized, and the natural parameters of the target manifold are used to describe the state information of the channel on the target manifold.

[0060] S106: Determine the statistical information of this round according to the statistical information of the previous round and the natural parameters of the target manifold of this round.

[0061] In an embodiment of the present application, the current statistical information is updated using the natural parameters of the target manifold of this round and the statistical information of the previous round. The statistical information can be gradually corrected in each round of iteration, thereby ensuring that the channel information is more accurate and making the final channel estimate closer to the channel characteristics in the real environment.

[0062] S107: Determine the target manifold parameter change rate based on the target manifold natural parameters of this round and the target manifold natural parameters of the previous round; determine the statistical information change rate based on the statistical information of this round and the statistical information of the previous round, and execute S102.

[0063] In an embodiment of the present application, after the natural parameters and statistical information of the target manifold of this round are determined in each iteration process, they will be compared with the natural parameters and statistical information of the target manifold of the previous round to determine the rate of change of the natural parameters of the target manifold between the two rounds and the rate of change of the statistical information between the two rounds.

[0064] S108: Determine the mean of the target manifold according to the natural parameters of the target manifold of the last round and the statistical information of the last round, and use the mean of the target manifold as the beam domain channel evaluation result.

[0065] In an embodiment of the present application, in response to the target manifold parameter change rate or the statistical information change rate being less than a preset threshold, the target manifold natural parameters and statistical information obtained in the last round of iteration are taken, and the mean of the target manifold is calculated as the final beam domain channel evaluation result. That is, the present application estimates the channel state in the beam domain through the mean of the target manifold.

[0066] In some implementations of the embodiments of the present application, S108 further includes the following steps:

[0067] The beam domain channel estimation result is converted into the space-frequency domain channel estimation result.

[0068] Among them, the beam domain channel evaluation result is the channel state information estimated by the channel in the beam domain, which is described by the mean of the target manifold in this application; and the space-frequency domain further combines the frequency dimension compared to the beam domain, so the space-frequency domain channel estimation result refers to the channel state information estimated by the channel in the spatial dimension and the frequency dimension. By mapping the channel state in the beam space back to the combined space of space and frequency, this application can better adapt to large-scale MIMO-OFDM systems and achieve more accurate channel state feedback.

[0069] exist Figure 1In the process shown, first, initialization is performed to provide initial conditions for subsequent iterations; then a loop step is entered to gradually update the auxiliary manifold natural parameters and the target manifold natural parameters until the target manifold parameter change rate or the statistical information change rate is less than a preset threshold.

[0070] Specifically, in the loop step: first, the m-projection of this round is determined according to the statistical information of the previous round and the natural parameters of the auxiliary manifold of the previous round; the complexity of parameter calculation is reduced by calculating the projection. Then, the natural parameters of the auxiliary manifold of this round are determined according to the m-projection of this round, the natural parameters of the auxiliary manifold of the previous round and the damping value; in each round of iteration, the gradual update under the control of the damping value can avoid excessive fluctuations, and at the same time, the stability and convergence of the estimation are ensured by the parameter update after the projection, which helps to improve the accuracy of channel estimation. After that, the natural parameters of the target manifold of this round are determined according to the natural parameters of the auxiliary manifold of this round; by gradually determining the natural parameters of the target manifold in the loop process, not only the calculation is simplified, but also the channel estimation is made more accurate by correcting the natural parameters in each iteration. Finally, new statistical information is determined according to the statistical information of the previous round and the natural parameters of the target manifold of this round. The updated natural parameters of the target manifold are used to update and correct the statistical information, so that the update of each round of statistical information further corrects the estimation of the channel, ensuring that the algorithm can converge to the real channel state in the step-by-step iteration. At the same time, the updated statistical information continues to be used in the next round of iteration, avoiding the reliance on previously acquired statistical CSI.

[0071] After the target manifold parameter change rate or the statistical information change rate is less than a preset threshold, the mean of the target manifold is determined as the final channel evaluation result according to the final target manifold natural parameters and statistical information.

[0072] It can be seen that in the above technical solution, the present application does not need to know the beam domain statistics in advance, but updates the statistical information by updating the natural parameters of the target manifold during the cyclic iteration process, and updating the natural parameters of the target manifold only involves multiplication operations of vectors and scalars, so the computational complexity is relatively low. It can be seen that the complexity of the technical solution provided by the present application is not significantly improved compared with the prior art, and it solves the problem that the SIG algorithm needs to know the statistical information in advance without increasing the complexity of the algorithm, which significantly improves the practicality of the SIG algorithm. At the same time, the final channel evaluation result has higher accuracy.

[0073] In terms of system model design, the present embodiment considers a MIMO-OFDM system, where the base station is equipped with N r =N r,v ×N r,h The uniform array of antennas has N antennas in the vertical and horizontal directions. r,v and Nr,h , the number of users is K, and each user is equipped with a single antenna. For OFDM modulation, the number of subcarriers is N c , the system sampling interval is T s , the cyclic prefix length is N g The subcarrier index used for uplink channel estimation is N p is the number of subcarriers used for channel estimation, N1 and N p are the starting and last subcarrier index numbers respectively, and i is the subcarrier sequence number.

[0074] During uplink transmission, the received signal in the frequency domain is expressed as:

[0075]

[0076] Where Y is the received signal, x k is the pilot signal sent by user k, Z is the noise, is the noise; the noise power is σ.

[0077] Introducing the beam domain channel model:

[0078] G k =VH k F T (2)

[0079] in is the statistical beam domain channel matrix. and is the delay domain sampling matrix, F τ is the delay domain oversampling factor. is the angle domain sampling matrix, where is the angular domain sampling matrix in the vertical direction,

[0080] N v =F v N r,v , F v is the oversampling factor in the vertical direction. is the angle domain sampling matrix in the horizontal direction, N h =F h N r，h , F h is the oversampling factor in the horizontal direction.

[0081] Substituting formula (2) into the received signal model yields:

[0082] Y=VH a M+Z (3)

[0083] The angle delay domain channel Delay Domain Transformation Matrix

[0084] The beam domain energy matrix is ​​defined as:

[0085]

[0086] Due to the sparseness of the channel, the values ​​of many elements in the beam domain energy matrix are close to zero. In the algorithm, these elements with small energy can be set to zero.

[0087] Vectorizing formula (3), we get:

[0088] y=Ah+z (5)

[0089] Where, the beam domain channel vector h = vec(H a ), sampling matrix Beam domain channel dimension N = N r N p , M=N v N h KN τ y is obtained by vectorizing Y, and z is obtained by vectorizing Z. Among them, the prior variance matrix D=Diag(d), d=vec{Ω}.

[0090] The SIG algorithm mentioned in the background art extracts paths with non-zero energy based on statistical information. The index set of non-zero elements in d is The extraction matrix is ​​defined as:

[0091]

[0092] Among them, s p is the pth column of the identity matrix, then Simplified sampling matrix in

[0093] This application further introduces the idea of ​​the expectation maximization EM algorithm on the SIG algorithm and proposes a simplified information geometry EM-SIG algorithm based on expectation maximization. The EM-SIG algorithm does not need to calculate statistical information in advance, so there is no need to extract the path with non-zero energy first. That is, the matrix extracted in formula (6) is E=I M×M .

[0094] In terms of design, the embodiment of the present application introduces the idea of ​​maximum expectation and regards h as a hidden variable. In the algorithm iteration process, the statistical information D is updated by using the parameters calculated in this round of SIG process combined with the idea of ​​expectation maximization algorithm.

[0095] The statistical information is updated by:

[0096]

[0097] The Q function can be written as

[0098]

[0099] According to the Bayesian formula:

[0100] p(y,h|D)=p(y|h)p(h|D) (9)

[0101] Substituting into formula (8), the Q function can be written as:

[0102]

[0103] In formula (10), the first term is independent of D, and only the second term is related to D, so we can get:

[0104]

[0105] In the SIG algorithm, the elements of h are assumed to be independent complex Gaussian distributions, so lnp(h|D) in equation (11) is:

[0106]

[0107] Among them, h m =[h] m , d m =[D] m,m , then its partial differential can be calculated as:

[0108]

[0109] In practical applications, the posterior probability is usually not available. However, in the SIG algorithm, the posterior probability can be approximated by the target manifold, so in formula (11), P(h|y,D t ) can be expressed by the target manifold parameters as:

[0110]

[0111] in and It can be obtained in the SIG algorithm, that is

[0112]

[0113]

[0114] Therefore, the partial derivative of the function in formula (11) with respect to D can be calculated as:

[0115]

[0116] The calculation formula (17) can be written as:

[0117]

[0118] in

[0119]

[0120] make We can get:

[0121]

[0122] Rewriting equation (21) into matrix form, we get:

[0123]

[0124] Where I is the identity matrix, D t is the statistical information of the previous iteration, D t+1 This is the updated statistics for this round. is the second-order natural parameter of the target manifold after this round of iteration, is the first-order natural parameter of the target manifold after this round of iteration.

[0125] In formula (18), when When , we can get the following formula, where is less than A real number.

[0126]

[0127] Right now The same reason Sometimes, there is Therefore, formula (22) is the maximum point, which satisfies formula (11).

[0128] See also Figure 2 As shown, an embodiment of the present application provides a channel assessment method. In this embodiment, the auxiliary manifold natural parameters specifically include: the auxiliary manifold first-order average natural parameters and the auxiliary manifold second-order average natural parameters; the m-projection specifically includes: the first m-projection and the second m-projection; the target manifold natural parameters include: the target manifold first-order natural parameters and the target manifold second-order natural parameters; the method specifically may include the following steps:

[0129] S201: Initialize the auxiliary manifold first-order average natural parameters, the auxiliary manifold second-order average natural parameters and statistical information, and set the damping value.

[0130] The first-order average natural parameter is specifically the mean of the channel parameter, and the second-order average natural parameter is the variance of the channel parameter.

[0131] S202: Determine whether the target manifold parameter change rate or the statistical information change rate in the current iteration is less than a preset threshold. If so, execute S208; if not, execute the loop steps of S203-S207.

[0132] Right now:

[0133] Among them, δ1, δ2 and δ3 are preset thresholds.

[0134] S203: Determine the first m-projection of this round according to the statistical information of the previous round and the first-order average natural parameters of the auxiliary manifold of the previous round; determine the second m-projection of this round according to the statistical information of the previous round and the second-order average natural parameters of the auxiliary manifold of the previous round.

[0135] In some implementations of the embodiments of the present application, the first m projection of this round is calculated by the following first formula (24):

[0136]

[0137] In the first formula (24), I represents the identity matrix, N represents the beam domain channel dimension, A represents the sampling matrix, and D t represents the statistical information of the previous round, σ represents the noise variance, represents the first-order average natural parameter of the auxiliary manifold in the previous round, represents the second-order average natural parameter of the auxiliary manifold in the previous round, tr represents the matrix trace operation, A represents the sampling matrix, y represents the received signal, and A H represents the conjugate transpose of the sampling matrix, represents the first m-th projection of this round;

[0138] In some implementations of the embodiments of the present application, the second m-projection of this round is calculated by the following second formula (25):

[0139]

[0140] In the second formula (25), N represents the beam domain channel dimension, D t Indicates the statistics of the previous round. represents the second-order average natural parameter of the auxiliary manifold in the previous round, σ represents the noise variance, tr represents the matrix trace operation, represents the second m-projection of this round.

[0141] In addition, it can be understood that when this round is the first round, in S103, the statistical information of the previous round is the statistical information initialized in S101, the first-order average natural parameters of the auxiliary manifold of the previous round are the first-order average natural parameters of the auxiliary manifold initialized in S101, and the second-order average natural parameters of the auxiliary manifold of the previous round are the second-order average natural parameters of the auxiliary manifold initialized in S101.

[0142] S204: Determine the first-order average natural parameters of the auxiliary manifold of this round according to the first m projections of this round, the first-order average natural parameters of the auxiliary manifold of the previous round and the damping value; determine the second-order average natural parameters of the auxiliary manifold of this round according to the second m projections of this round, the second-order average natural parameters of the auxiliary manifold of the previous round and the damping value.

[0143] In some implementations of the embodiments of the present application, the first-order average natural parameter of the auxiliary manifold of this round is calculated by the following third formula (26):

[0144]

[0145] In the third formula, represents the first-order average natural parameter of the auxiliary manifold of this round, α represents the damping value, N represents the channel dimension in the beam domain, represents the first m-th projection of this round, represents the first-order average natural parameter of the auxiliary manifold of the previous round;

[0146] In some implementations of the embodiments of the present application, the second-order average natural parameter of this round is calculated by the following fourth formula:

[0147]

[0148] In the fourth formula (27), represents the second-order average natural parameter of this round, α represents the damping value, N represents the channel dimension in the beam domain, represents the first m-th projection of this round, represents the second-order mean natural parameter of the auxiliary manifold in the previous round.

[0149] S205: Determine the first-order natural parameters of the target manifold of this round according to the first-order average natural parameters of the auxiliary manifold of this round; determine the second-order natural parameters of the target manifold of this round according to the second-order average natural parameters of the auxiliary manifold of this round.

[0150] Among them, the first-order natural parameters of the target manifold describe the mean information of the channel state on the target manifold, that is, the global description of the channel parameters; the second-order natural parameters of the target manifold describe the variance information of the channel state on the target manifold, that is, the volatility of the channel state.

[0151] In some implementations of the embodiments of the present application, the first-order natural parameters of the target manifold of this round are calculated by the following fifth formula (28):

[0152]

[0153] In the fifth formula (28), represents the first-order natural parameter of the target manifold of this round, N represents the channel dimension in the beam domain, represents the first-order average natural parameter of the auxiliary manifold of the previous round;

[0154] In some implementations of the embodiments of the present application, the second-order natural parameters of the target manifold of this round are calculated by the following sixth formula (29):

[0155]

[0156] In the sixth formula (29), represents the second-order natural parameter of the target manifold of this round, N represents the channel dimension in the beam domain, represents the second-order mean natural parameter of the auxiliary manifold in the previous round.

[0157] S206: Determine the statistical information of this round according to the statistical information of the previous round, the first-order natural parameters of the target manifold of this round, and the second-order natural parameters of the target manifold of this round, and execute S202.

[0158] In some implementations of the embodiments of the present application, the statistical information of this round is calculated by the following seventh formula (30):

[0159]

[0160] In the seventh formula (30), D t+1 represents the statistical information of this round, I represents the unit matrix, D t Indicates the statistics of the previous round. represents the second-order natural parameter of the target manifold of this round, Represents the first-order natural parameter of the target manifold of this round.

[0161] S207: Determine the change rate of the first-order parameters of the target manifold based on the first-order natural parameters of the target manifold in this round and the first-order natural parameters of the target manifold in the previous round; determine the change rate of the second-order parameters of the target manifold based on the second-order natural parameters of the target manifold in this round and the second-order natural parameters of the target manifold in the previous round; determine the change rate of the statistical information based on the statistical information of this round and the statistical information of the previous round.

[0162] In an embodiment of the present application, the target manifold natural parameters specifically include: the first-order natural parameters of the target manifold and the second-order natural parameters of the target manifold. Therefore, determining the rate of change of the target manifold parameters specifically includes determining the rate of change of the first-order natural parameters of the target manifold and the rate of change of the second-order natural parameters of the target manifold.

[0163] S208: Determine the mean of the target manifold according to the first-order natural parameters of the target manifold in the last round, the second-order natural parameters of the target manifold in the last round, and the statistical information in the last round, and use the mean of the target manifold as the beam domain channel evaluation result.

[0164] In some implementations of the embodiments of the present application, the mean of the target manifold is calculated by the following eighth formula (31):

[0165]

[0166] In the eighth formula (31), represents the mean of the target manifold, D T Indicates the statistics of the last round, represents the second-order natural parameter of the target manifold of this round, Represents the first-order natural parameter of the target manifold of this round.

[0167] Among them, the beam domain channel evaluation result is the channel state information estimated by the channel in the beam domain, which is described by the mean of the target manifold in this application; and the space-frequency domain further combines the frequency dimension compared to the beam domain, so the space-frequency domain channel estimation result refers to the channel state information estimated by the channel in the spatial dimension and the frequency dimension. By mapping the channel state in the beam space back to the combined space of space and frequency, this application can better adapt to large-scale MIMO-OFDM systems and achieve more accurate channel state feedback.

[0168] exist Figure 2 In the process shown, there is no need to know the beam domain statistics in advance. Instead, the statistical information is updated by updating the natural parameters of the target manifold during the loop iteration. Updating the natural parameters of the target manifold only involves multiplication operations of vectors and scalars, so the computational complexity is low. It can be seen that the complexity of the technical solution provided by the present application is not significantly improved compared with the prior art. It solves the problem that the SIG algorithm needs to know the statistical information in advance without increasing the complexity of the algorithm, and significantly improves the practicality of the SIG algorithm. At the same time, the final channel evaluation result has higher accuracy.

[0169] Below, the performance of the embodiments of the present application is described based on actual application scenarios.

[0170] In actual application scenarios, the number of subcarriers N is taken p=300, cyclic prefix N g =72,N c =1024, the base station is equipped with a 64×1 linear array, and the number of users K = 48. The test channel is the CDL-E channel.

[0171] When performing channel simulation, the normalized mean square error is defined as:

[0172]

[0173] Among them, G k is the spatial frequency domain channel of user k, is the estimated spatial-frequency domain channel of user k.

[0174] See also Figure 3 , which shows the comparison between the performance of the method provided by the present invention (abbreviated as EM-SIG in the figure) and the original SIG performance, Figure 3 The PMKL-SIG statistical information in is calculated using a statistical information estimation algorithm based on KL divergence, and then the SIG algorithm is used for channel estimation. For PMKL-12 and PMKL-48 in the legend, both set the condition of SNR=30dB when estimating statistical information, but the user grouping and estimation methods are different. Specifically, PMKL-12 represents a grouping strategy in which 12 users use orthogonal pilots, and these users are divided into a group for joint estimation of statistical information. PMKL-48 means directly estimating the statistical information of 48 users. It can be seen from the figure that the performance of the method of the present invention is much better than the scheme of estimating statistical information of 48 users at the same time. In addition, the performance of the method of the present invention is only lower than that of the PMKL-12 scheme at low signal-to-noise ratio. However, since the PMKL-12 scheme requires 12 users as a group for statistical information estimation, it will occupy more resources, and the complexity of the scheme will be much higher than the EM-SIG algorithm.

[0175] See also Figure 4 A graph of statistical information estimated by the method of the present invention is shown. Figure 4 This is the statistical information of one of the users estimated by the method of the present invention. It can be seen from the figure that the statistical information estimated by the method of the present invention is sparse. There are 104 paths with energy above -40dB, 54 paths with energy above -30dB, 15 paths with energy above -20dB, and 4 paths with energy above -10dB.

[0176] See also Figure 5 The accuracy comparison chart of the statistical information estimation of the method of the present invention is shown in FIG. Figure 5As shown, the EM-SIG line represents the statistical information estimated using the EM-SIG algorithm, and the PMKL-12 line represents the statistical information estimated using the SIG algorithm. As can be seen from the figure, the performance of the statistical information estimated by the method of the present invention is higher than the statistical information estimated by PMKL-12, that is, the present application has a higher evaluation performance than the prior art mentioned in the background technology.

[0177] See also Figure 6 As shown, the embodiment of the present application also provides a channel assessment device, which includes:

[0178] Initialization module 601, used to initialize the natural parameters and statistical information of the auxiliary manifold and set the damping value;

[0179] The loop execution module 602 is used to execute the following loop steps until the target manifold parameter change rate or the statistical information change rate is less than a preset threshold: determine the m-projection of this round according to the statistical information of the previous round and the auxiliary manifold natural parameters of the previous round; determine the auxiliary manifold natural parameters of this round according to the m-projection of this round, the auxiliary manifold natural parameters of the previous round and the damping value; determine the target manifold natural parameters of this round according to the auxiliary manifold natural parameters of this round; determine the statistical information of this round according to the statistical information of the previous round and the target manifold natural parameters of this round; determine the target manifold parameter change rate based on the target manifold natural parameters of this round and the target manifold natural parameters of the previous round; determine the statistical information change rate based on the statistical information of this round and the statistical information of the previous round;

[0180] The channel evaluation module 603 is used to determine the mean of the target manifold according to the natural parameters of the target manifold of the last round and the statistical information of the last round in response to the target manifold parameter change rate or the statistical information change rate being less than a preset threshold, and use the mean of the target manifold as the beam domain channel evaluation result.

[0181] In some implementations of the embodiments of the present application, the device further includes:

[0182] The result conversion module is used to convert the beam domain channel estimation result into the space-frequency domain channel estimation result.

[0183] In some implementations of the embodiments of the present application, the auxiliary manifold natural parameter includes: the auxiliary manifold first-order average natural parameter and the auxiliary manifold second-order average natural parameter;

[0184] According to the statistical information of the previous round and the natural parameters of the auxiliary manifold of the previous round, the m-projection of this round is determined, including:

[0185] The first m-projection of this round is determined according to the statistical information of the previous round and the first-order average natural parameters of the auxiliary manifold of the previous round; the second m-projection of this round is determined according to the statistical information of the previous round and the second-order average natural parameters of the auxiliary manifold of the previous round.

[0186] In some implementations of the embodiments of the present application, the auxiliary manifold natural parameters of the current round are determined according to the m-projection of the current round, the auxiliary manifold natural parameters of the previous round, and the damping value, including:

[0187] The first-order average natural parameters of the auxiliary manifold of this round are determined according to the first m-projection of this round, the first-order average natural parameters of the auxiliary manifold of the previous round and the damping value; the second-order average natural parameters of the auxiliary manifold of this round are determined according to the second m-projection of this round, the second-order average natural parameters of the auxiliary manifold of the previous round and the damping value.

[0188] In some implementations of the embodiments of the present application, the target manifold natural parameters include: first-order natural parameters of the target manifold and second-order natural parameters of the target manifold;

[0189] According to the natural parameters of the auxiliary manifold of this round, the natural parameters of the target manifold of this round are determined, including:

[0190] According to the first-order average natural parameters of the auxiliary manifold of this round, the first-order natural parameters of the target manifold of this round are determined; according to the second-order average natural parameters of the auxiliary manifold of this round, the second-order natural parameters of the target manifold of this round are determined;

[0191] In some implementations of the embodiments of the present application, the statistical information of this round is determined according to the statistical information of the previous round and the natural parameters of the target manifold of this round, including:

[0192] The statistical information of this round is determined according to the statistical information of the previous round, the first-order natural parameters of the target manifold of this round, and the second-order natural parameters of the target manifold of this round.

[0193] In some implementations of the embodiments of the present application, determining the mean of the target manifold according to the natural parameters of the target manifold in the last round and the statistical information in the last round includes:

[0194] The mean of the target manifold is determined according to the first-order natural parameters of the target manifold in the last round, the second-order natural parameters of the target manifold in the last round, and the statistical information in the last round.

[0195] like Figure 7 As shown, the embodiment of the present application further provides a channel assessment device, including: a memory 701, a processor 702;

[0196] Among them, the memory 701 is used to store programs; the processor 702 is used to execute the programs in the memory to implement the various steps of the channel assessment method provided in the embodiment of the present application.

[0197] Furthermore, the fourth aspect of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the channel assessment method of any one of the first aspects.

[0198] The embodiment of the present application further provides a readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the various steps of the channel assessment method provided in the embodiment of the present application are implemented.

[0199] Finally, it should be noted that, in the embodiments of the present application, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements includes not only those elements, but also includes other elements that are not clearly listed, or also includes elements inherent to such process, method, article or equipment. In the absence of more restrictions, the elements limited by the statement "comprise one..." do not exclude the presence of other identical elements in the process, method, article or equipment including the elements.

[0200] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A channel assessment method, characterized in that: The method comprises: Initialize the auxiliary manifold natural parameters and statistical information, and set the damping value; Execute the following loop steps until the target manifold parameter change rate or the statistical information change rate is less than a preset threshold; Determine the m-projection of this round based on the statistical information of the previous round and the natural parameters of the auxiliary manifold of the previous round; Determine the auxiliary manifold natural parameters of this round according to the m-projection of this round, the auxiliary manifold natural parameters of the previous round and the damping value; Determining the target manifold natural parameters of this round according to the auxiliary manifold natural parameters of this round; Determine the statistical information of this round according to the statistical information of the previous round and the natural parameters of the target manifold of this round; Determine the target manifold parameter change rate based on the target manifold natural parameters of the current round and the target manifold natural parameters of the previous round; determine the statistical information change rate based on the statistical information of the current round and the statistical information of the previous round; In response to the target manifold parameter change rate or the statistical information change rate being less than the preset threshold, the mean of the target manifold is determined according to the last round of target manifold natural parameters and the last round of statistical information, and the mean of the target manifold is used as the beam domain channel evaluation result.

2. The method according to claim 1, characterized in that The method further comprises: The beam-domain channel estimation result is converted into a space-frequency-domain channel estimation result.

3. The method according to claim 1, characterized in that The auxiliary manifold natural parameters include: the auxiliary manifold first-order average natural parameters and the auxiliary manifold second-order average natural parameters; Determining the m-projection of this round according to the statistical information of the previous round and the natural parameters of the auxiliary manifold of the previous round includes: The first m-projection of this round is determined according to the statistical information of the previous round and the first-order average natural parameters of the auxiliary manifold of the previous round; the second m-projection of this round is determined according to the statistical information of the previous round and the second-order average natural parameters of the auxiliary manifold of the previous round.

4. The method according to claim 3, characterized in that Determining the auxiliary manifold natural parameters of this round according to the m-projection of this round, the auxiliary manifold natural parameters of the previous round and the damping value includes: Determine the first-order average natural parameter of the auxiliary manifold of this round according to the first m-projection of the current round, the first-order average natural parameter of the auxiliary manifold of the previous round and the damping value; The second-order average natural parameters of the auxiliary manifold of this round are determined according to the second m-projection of this round, the second-order average natural parameters of the auxiliary manifold of the previous round and the damping value.

5. The method according to claim 4, characterized in that The target manifold natural parameters include: first-order natural parameters of the target manifold and second-order natural parameters of the target manifold; Determining the target manifold natural parameters of this round according to the auxiliary manifold natural parameters of this round includes: The first-order natural parameters of the target manifold of this round are determined according to the first-order average natural parameters of the auxiliary manifold of this round; and the second-order natural parameters of the target manifold of this round are determined according to the second-order average natural parameters of the auxiliary manifold of this round.

6. The method according to claim 5, characterized in that Determining the statistical information of this round according to the statistical information of the previous round and the natural parameters of the target manifold of this round includes: The statistical information of this round is determined according to the statistical information of the previous round, the first-order natural parameters of the target manifold of this round, and the second-order natural parameters of the target manifold of this round.

7. The method according to claim 6, characterized in that Determining the mean of the target manifold according to the natural parameters of the target manifold of the last round and the statistical information of the last round includes: The mean of the target manifold is determined according to the first-order natural parameters of the target manifold in the last round, the second-order natural parameters of the target manifold in the last round, and the statistical information in the last round.

8. A channel assessment device, characterized in that: The device comprises: Initialization module, used to initialize the natural parameters and statistical information of the auxiliary manifold and set the damping value; A loop execution module, used to execute the following loop steps until the target manifold parameter change rate or the statistical information change rate is less than a preset threshold: determine the m-projection of this round according to the statistical information of the previous round and the auxiliary manifold natural parameters of the previous round; determine the auxiliary manifold natural parameters of this round according to the m-projection of this round, the auxiliary manifold natural parameters of the previous round and the damping value; determine the target manifold natural parameters of this round according to the auxiliary manifold natural parameters of this round; determine the statistical information of this round according to the statistical information of the previous round and the target manifold natural parameters of this round; determine the target manifold parameter change rate based on the target manifold natural parameters of this round and the target manifold natural parameters of the previous round; determine the statistical information change rate based on the statistical information of this round and the statistical information of the previous round; A channel assessment module is used to determine the mean of the target manifold in response to the target manifold parameter change rate or the statistical information change rate being less than the preset threshold value, based on the last round of target manifold natural parameters and the last round of statistical information, and use the mean of the target manifold as the beam domain channel assessment result.

9. A channel assessment device, characterized in that: The device includes a memory and a processor, and the processor is used to execute a program stored in the memory to run the channel assessment method according to any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the channel estimation method according to any one of claims 1 to 7 are implemented.

11. A readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each step of the channel assessment method according to any one of claims 1 to 7 is implemented.

Images