Statistical prior assisted minimum mean square error interference suppression combined receiver method

Abstract

The invention provides a statistical prior assisted minimum mean square error interference suppression combined receiver method, which comprises the following steps of: firstly, constructing a special channel knowledge map for each interference cell; then, further performing grid division on the channel knowledge map, and realizing dynamic grid merging and hybrid time scale updating through reliability in space and time dimensions; besides, a frequency domain interference covariance matrix sharing mechanism based on channel correlation in a coherent bandwidth and an eigenvalue decomposition and compression method based on beam domain sparsity are introduced, so that the storage requirement and the calculation complexity of the channel knowledge map are further reduced; and finally, based on an interference cell user scheduling result, obtaining a joint interference covariance matrix in combination with the channel knowledge map, and applying the joint interference covariance matrix to a minimum mean square error interference suppression combination receiver. According to the method, the interference suppression performance of the minimum mean square error interference suppression merging receiver can be improved, the robustness and adaptability of the system are enhanced, and the method has relatively high practical value and popularization potential.

Description

Technical Field

[0001] The present invention belongs to the field of communications, and particularly relates to a method for a statistical prior-assisted minimum mean square error interference suppression combining receiver. Background Art

[0002] With the increasing demand for high data rates and efficient spectrum utilization, multi-cell MIMO systems have received extensive attention in the next-generation wireless networks. These systems utilize advanced technologies such as multiple antennas, spatial multiplexing, and coordinated transmission to serve multiple users simultaneously in different cells, thus providing an effective solution for maximizing spectrum utilization. However, due to co-channel interference, the spectral efficiency of multi-cell MIMO systems drops severely, which means that effective interference management techniques are crucial for realizing the potential of multi-cell MIMO systems.

[0003] The minimum mean square error interference suppression combining receiver is a key algorithm for dealing with interference. The receiver utilizes the correlation of interference between receiving antennas for interference suppression and combining, thereby improving system performance. Since the minimum mean square error interference suppression combining receiver is sensitive to the accuracy of the interference covariance matrix, traditional real-time interference covariance matrix estimation methods will suffer severe performance degradation in a strong interference environment. In contrast, the interference covariance matrix estimation method based on the channel knowledge map shows the potential to mitigate the impact of strong interference. However, the challenge of backhaul overhead makes it difficult to directly extend to multi-cell systems. Summary of the Invention

[0004] Object of the Invention: The core object of the present invention is to propose a method for a statistical prior-assisted minimum mean square error interference suppression combining receiver to solve the problem of high dependence of traditional interference covariance estimation methods on real-time channel state information. By deeply exploring the potential value of statistical prior information, the performance of the minimum mean square error interference suppression combining receiver is improved, and at the same time, the dependence of the system on real-time channel information is reduced, thereby optimizing the performance of the overall communication system.

[0005] Technical Solution: To achieve the above object of the invention, the present invention provides a method for a statistical prior-assisted minimum mean square error interference suppression combining receiver, including the following steps:

[0006] Construct a dedicated channel knowledge map for each interfering cell, and map the user location to the corresponding interference covariance matrix;

[0007] Perform grid division on the channel knowledge map so that users within each grid share the same interference covariance matrix;

[0008] In the spatial dimension, dynamic grid merging is performed based on spatial reliability assessment to merge grids with similar statistical characteristics; in the time dimension, based on time reliability assessment, hybrid time-scale updates are achieved, and different time scales are used for updates according to the speed of channel change in different grids.

[0009] The serving cell base station queries the corresponding interference covariance matrix from the constructed channel knowledge map based on the grid positions of the interfering users in the interfering cells to obtain the joint interference covariance matrix, which is applied to the minimum mean square error interference suppression combining receiver.

[0010] Preferably, on the grid-based channel knowledge map, the positions q, time slots t, and subcarriers k of the interfering users are mapped to a discrete interference covariance set The mapping of the l-th interfering cell is expressed as:

[0011]

[0012] where and K, T, Q l respectively represent the number of subcarriers, the number of time slots, and the number of grids divided by the l-th interfering cell, represents the interference covariance matrix of the j-th grid.

[0013] Preferably, the spatial reliability is the degree of statistical prior change at different positions within the same grid within a given time. By comparing the difference between the statistical interference covariance matrix of the grid and the interference covariance matrix at random positions within the grid, the spatial reliability of the channel within the grid is measured.

[0014] Preferably, the dynamic grid merging includes: setting a spatial reliability threshold. If the reliability metric of a certain grid is lower than the preset threshold, it is considered that the grid has spatial reliability. In this case, the grid can be merged with adjacent grids that are also determined to be spatially reliable; on the contrary, if the reliability metric exceeds the threshold, it is considered that the grid is spatially unreliable and the grid division needs to be further refined.

[0015] Preferably, the time reliability is used to evaluate the stability of the channel within the same grid over time. By comparing the difference between the statistical interference covariance matrices at different times within the same grid, the stability of the channel change within the grid over time is measured.

[0016] Preferably, the hybrid time-scale update includes: setting a time reliability threshold. If the reliability metric of a certain time period is lower than the threshold, the time period is considered to be time-reliable, indicating stable channel conditions. In this case, the update frequency is reduced; on the contrary, if the metric value exceeds the threshold, the update frequency is increased.

[0017] Further, the method further includes sharing the interference covariance matrix in the frequency domain among different subcarriers based on the correlation of the channel within the coherence bandwidth; let be the set of subcarriers within the coherence bandwidth, and rewrite the channel knowledge map according to the interference covariance matrix of the q-th grid as:

[0018]

[0019] where represents the interference covariance matrix of the k-th subcarrier in the q-th grid at the t-th time slot, and represents the number of elements in the subcarrier set .

[0020] Further, the method further includes, based on the channel sparsity characteristic in the beam domain, adopting an eigenvalue decomposition compression method to only retain the main eigenvalues and the corresponding eigenvectors to achieve the compression of the interference covariance matrix; the compression ratio is determined based on an energy threshold, and the number of selected eigenvalues satisfies:

[0021]

[0022] where p represents the cumulative energy ratio threshold, r represents the rank of the interference covariance matrix, σ i represents the non-zero eigenvalue, and n represents the number of selected eigenvalues.

[0023] Preferably, the joint interference covariance matrix is obtained by querying the channel knowledge map based on the user scheduling result of the interfering cell, including: given the scheduling result of the interfering cell, where represents the index set of the users scheduled in the l-th interfering cell, L is the number of interfering cells, N is the number of interfering users in each interfering cell, map each user to a location index set where represents the location of the user scheduled in the l-th interfering cell; by using the location set to query the corresponding channel knowledge map to obtain the interference covariance matrix R t′,l,n,k of each scheduled user, and finally obtain the joint interference covariance matrix where t′ represents the time when the channel knowledge map is established, and k represents the subcarrier index.

[0024] The present invention also provides a computer program product, including a computer program, and when the computer program is executed by a processor, it implements the steps of the method of a statistical prior-assisted minimum mean square error interference suppression combining receiver.

[0025] Beneficial effects: The present invention provides a method for a statistical prior-assisted minimum mean square error interference suppression combining receiver. This method constructs a channel knowledge map and estimates the joint interference covariance matrix based on this channel knowledge map, which is applied to the minimum mean square error interference suppression combining receiver. The present invention can be divided into two stages. In the first stage, a dedicated channel knowledge map is constructed for each interfering cell, and then the channel knowledge map is further divided into grids. In addition, a frequency-domain interference covariance matrix sharing mechanism based on the channel correlation within the coherence bandwidth and an eigenvalue decomposition compression method based on beam-domain sparsity are introduced to further reduce the storage requirements and computational complexity of the channel knowledge map. In the second stage, based on the user scheduling results of the interfering cells, the joint interference covariance matrix is obtained by combining the channel knowledge map and applied to the minimum mean square error interference suppression combining receiver. The present invention can significantly improve the interference suppression performance of the minimum mean square error interference suppression combining receiver, effectively improve the spectral efficiency of the multi-cell MIMO system under the condition of low backhaul overhead, and enhance the robustness and adaptability of the system, having high practical value and promotion potential. Description of the Drawings

[0026] Figure 1 This is the overall flowchart of the method of the present invention.

[0027] Figure 2 This is a schematic diagram of the base station cooperation strategy with low backhaul overhead in a multi-cell system.

[0028] Figure 3 This is a schematic diagram of the grid division of the interfering cell. Detailed Embodiments

[0029] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0030] As Figure 1 shown, a method for a statistical prior-assisted minimum mean square error interference suppression combining receiver disclosed in an embodiment of the present invention mainly includes: constructing a dedicated channel knowledge map for each interfering cell and mapping the user location to the corresponding interference covariance matrix; dividing the channel knowledge map into grids so that the users within each grid share the same interference covariance matrix; in the spatial dimension, performing dynamic grid merging based on spatial reliability assessment and merging the grids with similar statistical characteristics; in the time dimension, realizing hybrid time-scale update based on time reliability assessment and using different time scales for update according to the channel change speed of different grids; the serving cell base station queries the corresponding interference covariance matrix from the constructed channel knowledge map based on the grid positions of the interfering users in the interfering cells to obtain the joint interference covariance matrix, which is applied to the minimum mean square error interference suppression combining receiver.

[0031] Figure 2 Schematic diagram of a base station cooperation strategy for low backhaul overhead in a multi-cell system, including the following key sub-steps:

[0032] Sub-step 1 - Interfering cell scheduling: During the idle period, the interfering cell schedules users at different locations to obtain channels, enabling the serving cell to estimate interference characteristics at multiple spatial positions;

[0033] Sub-step 2 - Data exchange: In addition to user location information, the interfering cell also shares control signaling with the serving cell through the backhaul link, including pilot sequences, resource allocation information, and transmit power, etc.;

[0034] Sub-step 3 - Interference channel estimation: The serving cell estimates the interference channel h using various channel estimation algorithms based on the received signals and shared data;

[0035] Sub-step 4 - Channel knowledge map construction: The serving cell constructs a dedicated channel knowledge map for the interfering cell, mapping the user location to the corresponding interference covariance matrix R, where R = hh H 。

[0036] Figure 3 Schematic diagram of the grid division of the interfering cell. Before establishing the channel knowledge map, the coverage area of the interfering cell is first divided into several grids. Subsequently, the system collects historical channel state information within each grid through the base station cooperation strategy with low backhaul overhead. After processing these historical channel state information, a channel knowledge map is formed. This channel knowledge map is stored in the network in the form of a database for subsequent use by the interference suppression receiver.

[0037] Specifically, the embodiment of the present invention can be divided into two stages. In the first stage, a channel knowledge map is constructed. Using the base station cooperation strategy with low backhaul overhead, a dedicated channel knowledge map is constructed for each interfering cell, and then the channel knowledge map is further divided into grids. In addition, more preferably, a frequency-domain interference covariance matrix sharing mechanism based on channel correlation within the coherent bandwidth and an eigenvalue decomposition compression method based on beam domain sparsity are introduced to further reduce the storage requirement and computational complexity of the channel knowledge map. In the second stage, based on the user scheduling result of the interfering cell, the joint interference covariance matrix is obtained by combining the channel knowledge map and applied to the minimum mean square error interference suppression combining receiver. Specifically as follows:

[0038] Step 1, using the base station cooperation strategy with low backhaul overhead, construct a dedicated channel knowledge map for each interfering cell, mapping the user location to the corresponding interference covariance matrix. Specifically, assume there are L interfering cells, and the channel knowledge map of the l-th interfering cell can be expressed as:

[0039]

[0040] Among them, R l [t] = h l [t] h l [t] H represents the interference covariance matrix of the l-th interfering cell, and h l [t] represents the channel matrix at the user location q l [t] of the l-th interfering cell.

[0041] Step 2: Further divide the channel knowledge map into grids so that users within each grid share the same interference covariance matrix. Consider the physical area covered by the l-th interfering cell, which is divided into Q l grids, and the size of each grid is d × d m 2 . In addition to the above simple division, the sub-region division interval can be dynamically adjusted according to the type of propagation environment and the historical user distribution density, etc. For example, the division interval in the urban scenario is smaller than that in the open scenario.

[0042] After the channel knowledge map is divided into grids, it allows users within the same grid to share the same covariance matrix. Assume that the channel model of the n-th interfering user of the l-th interfering cell at the k-th subcarrier in the t-th time slot is:

[0043]

[0044] Among them, represents the Kronecker product, and S′ n represents the number of multipaths experienced by the n-th interfering user of the l-th interfering cell. α t,l,n,k,s is a complex gain, and respectively represent the angle of arrival and the zenith angle of the s-th path, and a x (η), η ∈ {θ l,n,s , φ l,n,s} represents the steering vector, which is defined as:

[0045]

[0046] Among them, λ c represents the wavelength, Δ ∈ {Δ rv , Δ rh}, where Δ rv and Δ rh respectively represent the antenna spacings of the vertical receiving antenna and the horizontal receiving antenna.

[0047] We assume that in a non-correlated scattering environment, different paths are independent of each other, so the complex path gain satisfies:

[0048]

[0049] where δ(.) represents the Dirac function, and ρ t,l,n,k,s represents the channel power of the nth user in the lth interfering cell at the kth subcarrier and the tth time slot.

[0050] Therefore, the statistical interference covariance of the interference channel can be expressed as:

[0051]

[0052] where E{·} represents the expectation of environmental perturbations.

[0053] Based on the spatial consistency assumption, the interference covariance can be further expressed as:

[0054]

[0055] On the grid-based channel knowledge map, we replace the angle parameters related to scatterers with a location-based parameter set, and map the location q, time slot t, and subcarrier k of the interfering users to a discrete set of interference covariances :

[0056]

[0057] where and

[0058] Step 3: For the spatio-temporal dynamic characteristics of the wireless channel, a reliability evaluation mechanism is introduced in the spatial and temporal dimensions to dynamically update the statistical prior information of the grid. In the spatial dimension, based on the spatial reliability evaluation mechanism, a dynamic grid merging strategy is implemented to merge grids with highly similar statistical characteristics. In the temporal dimension, based on the temporal reliability evaluation mechanism, a hybrid time-scale update strategy is proposed, and different time scales are used for updating according to the speed of channel change in different grids.

[0059] The reliability in the spatial dimension refers to the degree of change in the statistical prior at different positions within the same grid within a given time. For the interference covariance matrix of the kth subcarrier in the qth grid at a certain moment its spatial reliability is defined as:

[0060]

[0061] where ||.|| represents the l2 norm, Denote the interference covariance matrix of the \(k\)-th subcarrier in the \(q\)-th grid, which depends on the user location \(Q\) modeled as a random variable. The above formula provides a measure of the spatial reliability of the channel within the grid by comparing the difference between the statistical interference covariance matrix of the grid and the interference covariance matrix at a random location within the grid.

[0062] For practical evaluation convenience, we use the sample mean to approximate the expected value. Specifically, we assume that \(P\) spatial sampling points are randomly selected from each grid for spatial reliability evaluation, and the above formula can be approximated as:

[0063]

[0064] where, denotes the interference covariance matrix of the \(p\)-th spatial sampling point within the \(q\)-th grid.

[0065] The dynamic grid merging strategy is based on the spatial reliability evaluation mechanism to determine whether to merge adjacent grids. Specifically, first, a spatial reliability threshold is set. If the reliability measure of a certain grid is lower than the preset threshold, it is considered that the grid has spatial reliability, that is, the interference covariance matrix within the grid changes little. In this case, this grid can be merged with adjacent grids that are also determined to be spatially reliable, so as to share the same interference covariance matrix in a larger grid area and reduce the complexity of constructing the channel knowledge map. On the contrary, if the reliability measure exceeds the threshold, it is considered that the grid is spatially unreliable, indicating that the environment within the grid changes greatly, and the grid division needs to be further refined to ensure the accuracy of the channel knowledge map. When the reliability measure is still high even at an extremely small grid scale, it indicates that the channel knowledge map cannot be effectively applied in this area, and real-time channel estimation must be used at this time to obtain an accurate joint interference covariance matrix.

[0066] The reliability evaluation in the time dimension concerns the stability of the channel over time within the same grid. This is particularly important in dynamic scenarios because the movement of users or environmental changes can cause changes in channel conditions. For the interference covariance matrix of the \(k\)-th subcarrier in the \(t\)-th time slot The time reliability can be defined as:

[0067]

[0068] where, denotes the interference covariance matrix of the \(k\)-th subcarrier in the \(t\)-th time slot, and \(T\) represents a random variable among \(T\) time slots. This measure is used to evaluate the stability of the interference covariance matrix over time within the grid.

[0069] To approximate the expected value, we select \(J\) random time sampling points from \(T\) time slots to obtain the following expression:

[0070]

[0071] Among them, represents the interference covariance matrix at the j-th time sampling point.

[0072] The hybrid time-scale update strategy determines the update frequency of the channel knowledge map based on the time reliability evaluation mechanism. Specifically, first, a time reliability threshold is set. If the reliability metric for a certain time period is lower than this threshold, then this time period is considered time-reliable, indicating that the channel conditions are stable. In this case, updates can be performed at a lower frequency, thereby reducing the complexity of updating the interference covariance matrix. On the contrary, if the metric value exceeds the threshold, more frequent updates are required to capture the rapid changes in the channel. For regions where the time reliability is still low even at the fine time scale, real-time channel estimation must be used to replace the statistical prior to ensure the generation of an accurate joint interference covariance matrix.

[0073] Step 4: Based on the correlation of the channel within the coherence bandwidth, a frequency-domain interference covariance matrix sharing mechanism is proposed to achieve the sharing of the interference covariance matrix among different subcarriers. The frequency-domain interference covariance matrix sharing mechanism for channel correlation within the coherence bandwidth utilizes the significant correlation among different subcarriers within the coherence bandwidth, enabling the interference covariance matrix to be shared among subcarriers, thereby avoiding storing the interference covariance matrix separately for each subcarrier and significantly reducing the storage overhead. Specifically, let be the set of subcarriers within the coherence bandwidth, and the interference covariance matrix of the q-th grid can be expressed as:

[0074]

[0075] According to the above formula, the channel knowledge map can be rewritten as:

[0076]

[0077] Among them, represents the interference covariance matrix of the k-th subcarrier of the q-th grid in the t-th time slot, represents the subcarrier set the number of elements of.

[0078] In addition, based on the channel sparsity characteristics in the beam domain, an eigenvalue decomposition compression method is adopted to retain only the main eigenvalues and the corresponding eigenvectors to achieve effective compression of the interference covariance matrix. The eigenvalue decomposition compression method based on beam-domain sparsity utilizes the low-rank property of the interference covariance matrix and can be well approximated by a few dominant eigenvalues and their corresponding eigenvectors, thereby significantly reducing the storage requirements while retaining the main features of the original matrix.

[0079] Specifically, given the interference covariance matrix The eigenvalue decomposition can be expressed as:

[0080]

[0081] where is the identity matrix, whose columns contain the eigenvectors of r forming an orthogonal basis for the column space of this matrix, and N is a diagonal matrix, whose diagonal elements are the eigenvalues, expressed as:

[0082]

[0083] where are the non-zero eigenvalues arranged in descending order, and r represents the rank of

[0084] To balance the trade-off between compression accuracy and storage overhead, the compression ratio is determined based on an energy threshold. Specifically, the number of selected eigenvalues satisfies:

[0085]

[0086] where p represents the cumulative energy ratio threshold. By setting different cumulative energy ratio thresholds p, different compression ratios for the statistical interference covariance matrix of each grid can be achieved. In this case, only the first n eigenvalues and their corresponding eigenvectors need to be stored.

[0087] Therefore, the low-rank approximation of the statistical interference covariance matrix can be expressed as:

[0088]

[0089] where is a diagonal matrix containing the first n eigenvalues, and is the matrix composed of the corresponding eigenvectors.

[0090] Step 5, determine the index set of interfering users for the current scheduling according to the real-time scheduling information of the interfering cell base station, and map these users to the corresponding grid position index set. The serving cell base station queries the corresponding interference covariance matrix from the constructed channel knowledge map based on the grid position of the user, and further combines the interference covariance information of all scheduled users to obtain the joint interference covariance matrix.

[0091] Specifically, given the scheduling result of the interfering cell where Denote the index set of users scheduled in the \(l\)-th interfering cell. \(N\) is the number of interfering users in each interfering cell. We can map each user to a location index set. where represents the location of the scheduled user in the \(l\)-th interfering cell. By using the location set the corresponding channel knowledge map can be queried so as to obtain the interference covariance matrix \(R\) of each scheduled user t′,l,n,k , and finally the joint interference covariance matrix

[0092] Finally, under the quasi-static assumption, the minimum mean square error interference suppression combining receiver utilizes this joint interference covariance matrix to replace the instantaneous joint interference covariance matrix that needs to be estimated in real time in the traditional method, and realizes interference suppression and signal detection.

[0093] The quasi-static assumption means that the propagation environment remains quasi-static on a relatively long time scale. When the base station location is fixed, the change of the channel matrix \(h[t]\) is mainly caused by the change of the user location (denoted as \(q[t]\)) and the change of the propagation environment (denoted as \(E[t]\)). Therefore, the channel matrix \(h[t]\) can be expressed as:

[0094] \(h[t]=g(q[t],E[t])\),

[0095] where \(g(\cdot)\) is a mapping function. The propagation environment remains quasi-static on a relatively long time scale \(T\) relative to the signal transmission period, that is, when \(|t'-t|\leq T\), \(E[t]\approx E[t']\).

[0096] The minimum mean square error interference suppression combining receiver is expressed as:

[0097]

[0098] where represents the serving cell channel, represents the instantaneous joint interference covariance matrix. Using the quasi-static assumption, the joint interference covariance matrix \(R\) obtained from the channel knowledge map t′,k replaces the joint interference covariance matrix \(R\) that needs to be estimated in real time t′,k and is applied to the minimum mean square error interference suppression combining receiver.

[0099] An embodiment of the present invention also discloses a computer program product, including a computer program, and when the computer program is executed by a processor, the steps of the method for a statistical prior-assisted minimum mean square error interference suppression combining receiver are implemented.

[0100] The program code for implementing the method of the present invention can be written in any combination of one or more programming languages. These program codes can be provided to a processor or a controller of a general-purpose computer, a special-purpose computer, or other programmable data processing devices, such that when the program codes are executed by the processor or the controller, the steps of the method of the present invention are implemented. The program codes can be executed entirely on a machine, partially on a machine, executed partially on a machine and partially on a remote machine as an independent software package, or executed entirely on a remote machine or a server. Where the present invention is not described in detail, it is common knowledge to those skilled in the art.

Claims

1. A method for a statistical prior-assisted minimum mean square error interference suppression combining receiver, characterized in that Including the following steps: Construct a dedicated channel knowledge map for each interfering cell, mapping the user location to the corresponding interference covariance matrix; Perform grid division on the channel knowledge map so that users within each grid share the same interference covariance matrix; In the spatial dimension, perform dynamic grid merging based on spatial reliability assessment, merging grids with similar statistical characteristics; in the time dimension, implement hybrid time-scale update based on time reliability assessment, and use different time scales for update according to the speed of channel change in different grids; The serving cell base station queries the corresponding interference covariance matrix from the constructed channel knowledge map based on the grid location of the interfering cell interfering with the user, obtains the joint interference covariance matrix, and applies it to the minimum mean square error interference suppression combining receiver.

2. A method for a statistical prior-assisted minimum mean square error interference suppression combining receiver according to claim 1, characterized in that: On the grid-based channel knowledge map, map the location q, time slot t, and subcarrier k of the interfering user to the discrete interference covariance set The mapping of the l-th interfering cell is expressed as: Among them, and K, T, Q l respectively represent the number of subcarriers, the number of time slots, and the number of grids divided by the l-th interfering cell. j ∈ {1,..., Q l} represents the interference covariance matrix of the j-th grid.

3. A method for a statistical prior-assisted minimum mean square error interference suppression combining receiver according to claim 1, characterized in that: The spatial reliability is the degree of statistical prior change at different positions within the same grid within a given time, and the spatial reliability of the channel within the grid is measured by comparing the difference between the statistical interference covariance matrix of the grid and the interference covariance matrix at a random position within the grid.

4. A method for a statistical prior-assisted minimum mean square error interference suppression combining receiver according to claim 1, wherein: The dynamic grid merging includes: setting a spatial reliability threshold. If the reliability metric of a certain grid is lower than the preset threshold, it is considered that the grid has spatial reliability. In this case, the grid can be merged with adjacent grids that are also determined to be spatially reliable; on the contrary, if the reliability metric exceeds the threshold, it is considered that the grid is not reliable in space and the grid division needs to be further refined.

5. A method for a statistical prior-assisted minimum mean square error interference suppression combining receiver according to claim 1, characterized in that: The time reliability is used to evaluate the stability of the channel within the same grid over time, and the stability of the channel change within the grid over time is measured by comparing the difference between the statistical interference covariance matrices at different times in the same grid.

6. A method for a statistical prior-assisted minimum mean square error interference suppression combining receiver according to claim 1, characterized in that: The hybrid time-scale update includes: setting a time reliability threshold. If the reliability metric of a certain time period is lower than the threshold, the time period is considered to be time reliable, indicating stable channel conditions. In this case, the update frequency is reduced; on the contrary, if the metric value exceeds the threshold, the update frequency is increased.

7. A method for a statistical prior-assisted minimum mean square error interference suppression combining receiver according to claim 2, characterized in that: It also includes sharing the frequency-domain interference covariance matrix among different subcarriers based on the correlation of the channel within the coherent bandwidth; let be the set of subcarriers within the coherent bandwidth, and rewrite the channel knowledge map according to the interference covariance matrix of the q-th grid as: Among them, represents the interference covariance matrix of the k-th subcarrier in the q-th grid at the t-th time slot, represents the subcarrier set the number of elements of.

8. A method for a statistical prior-assisted minimum mean square error interference suppression combining receiver according to claim 1, characterized in that: It also includes adopting an eigenvalue decomposition compression method based on the channel sparsity characteristics in the beam domain, only retaining the main eigenvalues and the corresponding eigenvectors to achieve compression of the interference covariance matrix; the compression ratio is determined based on an energy threshold, and the number of selected eigenvalues satisfies: Among them, p represents the cumulative energy ratio threshold, r represents the rank of the interference covariance matrix, σ i represents the non-zero eigenvalue, and n represents the number of selected eigenvalues.

9. A method for a statistical prior-assisted minimum mean square error interference suppression combining receiver according to claim 1, characterized in that: The combined interference covariance matrix is obtained by querying the channel knowledge map based on the interference cell user scheduling results, including: given the interference cell scheduling results where represents the index set of the users scheduled in the \(l\)-th interference cell, \(L\) is the number of interference cells, \(N\) is the number of interfering users in each interference cell, and each user is mapped to a location index set where represents the location of the scheduled users in the \(l\)-th interference cell; by using the location set query the corresponding channel knowledge map to obtain the interference covariance matrix \(R\) of each scheduled user t′,l,n,k , and finally obtain the combined interference covariance matrix where \(t'\) represents the time when the channel knowledge map is established, and \(k\) represents the subcarrier index.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of a statistical prior-assisted minimum mean square error interference suppression combining receiver method according to any one of claims 1-9.

Images