Uplink multi-user signal detection method and device suitable for XL-MIMO

Abstract

The application discloses an uplink multi-user signal detection method and device suitable for XL-MIMO. The method comprises the following steps: obtaining channel estimation information according to a received uplink pilot signal, and acquiring position information of each user to a base station antenna array; based on the channel estimation information and the position information, energy similarity, position similarity and channel correlation among users are respectively acquired, and a grouping joint metric is constructed; based on the grouping joint metric, each user is grouped, and an active antenna set corresponding to each user group is determined; in one inter-group iteration, each user group is sequentially processed, and based on the detection result of other user groups, an approximate expectation propagation iteration detection is performed in the current user group, and the detection result is updated; if the inter-group iteration satisfies a preset termination condition, the detection result of each user group is output. The joint accurate grouping is realized, the detection dimension is reduced, the inter-group interference is suppressed, the overall calculation complexity is reduced, and the effective balance between the detection performance and the calculation efficiency is realized.

Description

Technical Field

[0001] This application relates to the field of wireless communication technology, specifically to Extra Large Multiple Input Multiple Output (XL-MIMO) systems in Sixth Generation Mobile Communication (6G), and particularly to an uplink multi-user signal detection method and apparatus suitable for XL-MIMO. Background Technology

[0002] As sixth-generation mobile communication systems evolve towards XL-MIMO, the aperture of base station arrays increases significantly, and the base station receiver array may contain hundreds or even more antenna elements. Under such large-aperture arrays, the traditional far-field plane wave assumption no longer holds, and users are more likely to be within the Rayleigh distance of the base station array, resulting in near-field propagation effects.

[0003] In near-field XL-MIMO systems, the propagation distances from the user to different antenna elements in the array are no longer approximately equal. The signal phase exhibits spherical wave variations, and the signal amplitude also changes significantly with the antenna position. More importantly, due to the influence of array aperture, obstruction, and other factors, a user's channel energy is typically concentrated only in a portion of the antenna area, rather than being uniformly distributed across the entire array. This phenomenon is commonly referred to as spatial nonstationarity, or the user having a visible area on the base station array.

[0004] For uplink multi-user detection, the base station needs to recover the transmitted symbols of multiple users from the received signal. Traditional linear detection methods, such as Zero Forcing (ZF) and Minimum Mean Square Error (MMSE), have low complexity, but their bit error rate performance is limited in scenarios with strong multi-user interference, strong channel correlation, or low to medium signal-to-noise ratio. Expectation Propagation (EP) is a class of nonlinear iterative detection methods that outperform traditional linear methods. Based on the Maximum A Posteriori (MAP) criterion, it approximates the posterior distribution of the transmitted signal through a Gaussian distribution, continuously updating parameters in the iterations to gradually approximate the true distribution. The detection algorithm based on Approximate Expectation Propagation (AEP) further reduces the computational complexity of EP by rewriting and simplifying the moment matching process and using Newman series to simplify matrix inversion. However, in the XL-MIMO scenario, if global EP / AEP detection is performed directly on the complete channel matrix, it is necessary to handle large-dimensional matrices and multiple rounds of iterative updates, resulting in high computational complexity and implementation overhead.

[0005] Existing local detection schemes attempt to reduce complexity through grouping, but often group users based on only a single dimension, failing to fully utilize the spatial geometry and energy distribution characteristics of the near-field channel, and lacking effective inter-group interference suppression mechanisms. This results in residual inter-group interference severely impacting detection performance. Therefore, there is an urgent need for an uplink multi-user signal detection method that can fully exploit the non-stationary characteristics of near-field space, while ensuring detection performance while reducing computational complexity. Summary of the Invention

[0006] At least one embodiment of this application provides an uplink multi-user signal detection method and apparatus suitable for XL-MIMO, which solves the problem in the prior art that it is difficult to balance the computational efficiency and detection performance of XL-MIMO uplink signal detection.

[0007] To solve the above-mentioned technical problems, this application is implemented as follows:

[0008] This application provides an uplink multi-user signal detection method suitable for XL-MIMO, including:

[0009] Channel estimation is performed based on the received uplink pilot signal to obtain channel estimation information, and the location information of each user to the base station antenna array is obtained.

[0010] Based on the channel estimation information and the location information, the energy similarity, location similarity and channel correlation between users are obtained respectively, and a group joint metric for characterizing the comprehensive similarity between users is constructed.

[0011] Based on the grouping joint metric, each user is grouped to obtain multiple user groups, and the set of active antennas corresponding to each user group is determined.

[0012] In one inter-group iteration, each user group is processed sequentially. For the current user group, based on the detection results of the user groups that have been processed in this iteration and the detection results of the unprocessed user groups in the previous iteration, an approximate expectation propagation iterative detection is performed in the current user group, and the detection results corresponding to the current user group are updated.

[0013] If the inter-group iterations meet the preset termination condition, the detection results for each user group are output.

[0014] Preferably, in the uplink multi-user signal detection method for XL-MIMO as described above, the step of obtaining energy similarity, location similarity, and channel correlation among users based on the channel estimation information and the location information, and constructing a grouped joint metric to characterize the comprehensive similarity among users, includes:

[0015] Based on the channel estimation information matrix, the normalized energy vector of each user is obtained, and the energy similarity between any two users is calculated.

[0016] Based on the location information, calculate the location similarity between any two users;

[0017] Based on the channel estimation information, calculate the channel correlation between any two users;

[0018] The group joint metric is constructed based on the energy similarity, the location similarity, and the channel correlation.

[0019] Specifically, in the uplink multi-user signal detection method for XL-MIMO described above, the step of calculating the location similarity between any two users based on the location information includes:

[0020] Based on the location information, the distance difference, azimuth difference, and pitch difference between any two users are normalized to obtain a normalized location difference quantity. The location information includes: the user's estimated distance, estimated azimuth, and estimated pitch.

[0021] The kernel function is used to map the normalized positional differences to obtain the bounded positional similarity.

[0022] Preferably, in the uplink multi-user signal detection method for XL-MIMO as described above, the step of grouping the users based on the grouped joint metric to obtain multiple user groups includes:

[0023] Based on the grouped joint metric, calculate the local density and global coupling degree of each user;

[0024] Based on the difference between the local density and the global coupling degree, a representative metric for each user is constructed;

[0025] Based on the representativeness metric and the grouped joint metric, a preset number of seed users are determined;

[0026] Each seed user is used as the group center of a user group, and the intra-group affinity of the candidate user with each user group and the inter-group coupling of the candidate user within one user group to other user groups are calculated.

[0027] Based on the allocation metric constructed from the difference between the intra-group affinity and the inter-group coupling, the candidate user is assigned to the target user group that maximizes the allocation metric.

[0028] Preferably, in the uplink multi-user signal detection method for XL-MIMO as described above, determining a preset number of seed users based on the representativeness metric and the grouped joint metric includes:

[0029] The user with the highest representative metric is selected as the first seed user in the seed set;

[0030] The representativeness metric of the candidate user is subtracted from the difference in similarity between the candidate user and the seed set;

[0031] The candidate user with the largest difference is identified as the new seed user and stored in the seed set until a preset number of seed users are determined.

[0032] Preferably, in the uplink multi-user signal detection method for XL-MIMO as described above, determining the set of active antennas corresponding to each user group includes:

[0033] The total energy of the target user group on each antenna subarray is obtained, and the total energy corresponding to each antenna subarray is sorted from largest to smallest to obtain a subarray sequence, wherein the target user group is any one of the user groups;

[0034] The number of subarrays required for the target user group is determined based on the cumulative energy threshold criterion.

[0035] Target subarrays corresponding to the number of subarrays are selected sequentially from the subarray sequence, and the union of the antennas contained in the target subarrays is taken as the active antenna set of the target user group.

[0036] Preferably, in the uplink multi-user signal detection method for XL-MIMO as described above, the step of performing approximate expected propagation iterative detection within the current user group based on the detection results of the user group processed in the current iteration and the detection results of the unprocessed user group in the previous iteration, and updating the detection results corresponding to the current user group, includes:

[0037] The residual signal is obtained by subtracting the interference generated on the active antenna set corresponding to the current user group by the detection results of the user group that have been processed in the current iteration and the detection results of the user group that have not been processed in the previous iteration from the received signal received by the base station.

[0038] Based on the residual signal and the channel of the current user group on the corresponding active antenna set, a corresponding local detection model is constructed, and the approximate expected propagation iterative detection is performed within the local detection model to obtain the symbol soft estimate of the current user group.

[0039] Perform constellation hard decision on the symbol soft estimate, update the detection result corresponding to the current user group based on the symbol estimate after hard decision, and mark the current user group as the processed user group.

[0040] Preferably, the uplink multi-user signal detection method for XL-MIMO described above further includes, before sequentially processing each user group within one inter-group iteration:

[0041] Based on the received signal and the channel estimation information, the initial detection results for each user are obtained through low-complexity linear minimum mean square error detection.

[0042] Another embodiment of this application provides a control device for uplink multi-user signal detection suitable for XL-MIMO, comprising:

[0043] The first processing module is used to perform channel estimation based on the received uplink pilot signal, obtain channel estimation information, and acquire the location information of each user to the base station antenna array.

[0044] The second processing module is used to obtain the energy similarity, location similarity and channel correlation between users based on the channel estimation information and the location information, and to construct a grouped joint metric to characterize the comprehensive similarity between users.

[0045] The third processing module is used to group each user based on the grouping joint metric to obtain multiple user groups, and to determine the set of active antennas corresponding to each user group.

[0046] The fourth processing module is used to process each user group sequentially in one inter-group iteration, wherein, for the current user group, based on the detection results of the user groups that have been processed in the current iteration and the detection results of the unprocessed user groups in the previous iteration, an approximate expectation propagation iterative detection is performed in the current user group, and the detection results corresponding to the current user group are updated.

[0047] The fifth processing module is used to output the detection results of each user group if the inter-group iteration meets the preset termination condition.

[0048] Another embodiment of this application provides an electronic device, including: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the uplink multi-user signal detection method for XL-MIMO as described above.

[0049] Another embodiment of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the uplink multi-user signal detection method for XL-MIMO as described above.

[0050] Another embodiment of this application provides a computer program product including computer instructions that, when executed by a processor, implement the steps of the uplink multi-user signal detection method applicable to XL-MIMO as described above.

[0051] Compared with existing technologies, the uplink multi-user signal detection method and apparatus for XL-MIMO provided in this application constructs a multi-dimensional joint metric by fusing energy similarity, location similarity and channel correlation among users. This enables joint and accurate grouping of users and antennas, decomposing the high-dimensional detection task that originally operated on a global large-scale channel matrix into multiple low-dimensional local detection sub-problems, thereby reducing the detection dimensionality from the source. On this basis, a hierarchical architecture combining inter-group serial interference cancellation and intra-group approximate expectation propagation iterative detection is adopted. On the one hand, inter-group serial interference cancellation is used to gradually suppress residual interference between different user groups. On the other hand, the approximate expectation propagation algorithm is used within each group to complete fine nonlinear detection. Thus, while maintaining near-global nonlinear detection performance, the overall computational complexity is significantly reduced, achieving an effective balance between detection performance and computational efficiency. This method is particularly suitable for near-field XL-MIMO uplink multi-user communication scenarios. Attached Figure Description

[0052] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0053] Figure 1 This is one of the flowcharts illustrating the uplink multi-user signal detection method applicable to XL-MIMO in this application;

[0054] Figure 2 This is the second flowchart illustrating the uplink multi-user signal detection method applicable to XL-MIMO in this application;

[0055] Figure 3 This is the third flowchart illustrating the uplink multi-user signal detection method applicable to XL-MIMO in this application;

[0056] Figure 4 This is the fourth flowchart illustrating the uplink multi-user signal detection method applicable to XL-MIMO in this application;

[0057] Figure 5 This is the fifth flowchart illustrating the uplink multi-user signal detection method applicable to XL-MIMO in this application;

[0058] Figure 6 This is the sixth flowchart illustrating the uplink multi-user signal detection method applicable to XL-MIMO in this application;

[0059] Figure 7 A schematic diagram illustrating the bit error rate of different signal detection algorithms in an XL-MIMO multi-user uplink scenario;

[0060] Figure 8 A diagram illustrating the normalized computational complexity of different detection algorithms;

[0061] Figure 9 This is a schematic diagram of the control device for uplink multi-user signal detection applicable to XL-MIMO according to this application. Detailed Implementation

[0062] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0063] The terms “first,” “second,” etc., used in the specification and claims of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus. The terms “and / or” in the specification and claims indicate at least one of the connected objects.

[0064] The technologies described in this document are not limited to NR systems and Long Time Evolution (LTE) / LTE-Advanced (LTE-A) systems, but can also be used in various wireless communication systems such as Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal Frequency Division Multiple Access (OFDMA), Single-carrier Frequency-Division Multiple Access (SC-FDMA), and other systems. The terms "system" and "network" are often used interchangeably. CDMA systems can implement radio technologies such as CDMA2000 and Universal Terrestrial Radio Access (UTRA). UTRA includes Wideband Code Division Multiple Access (WCDMA) and other CDMA variants. TDMA systems can implement radio technologies such as the Global System for Mobile Communication (GSM). OFDMA systems can implement radio technologies such as Ultra Mobile Broadband (UMB), Evolution-UTRA (E-UTRA), IEEE 802.21 (Wi-Fi), IEEE 802.16 (WiMAX), IEEE 802.20, and Flash-OFDM. UTRA and E-UTRA are part of the Universal Mobile Telecommunications System (UMTS). LTE and more advanced LTE (such as LTE-A) are newer versions of UMTS that use E-UTRA. UTRA, E-UTRA, UMTS, LTE, LTE-A, and GSM are described in documents from an organization called the 3rd Generation Partnership Project (3GPP). CDMA2000 and UMB are described in documents from an organization called 3rd Generation Partnership Project 2 (3GPP2).The techniques described herein can be used with the systems and radio technologies mentioned above, as well as with other systems and radio technologies. However, the following description describes NR systems for illustrative purposes, and NR terminology is used in most of the following description, although these techniques can also be applied to applications beyond NR systems.

[0065] The following description provides examples and is not intended to limit the scope, applicability, or configuration set forth in the claims. Changes may be made to the function and arrangement of the elements discussed without departing from the spirit and scope of this disclosure. Various procedures or components may be appropriately omitted, substituted, or added to the examples. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Furthermore, features described with reference to certain examples may be combined in other examples.

[0066] Wireless communication systems include terminals and network equipment. A terminal, also known as a user terminal or user equipment (UE), can be a mobile phone, tablet computer, laptop computer, personal digital assistant (PDA), mobile internet device (MID), wearable device, or in-vehicle device, etc. It should be noted that the specific type of terminal is not limited in the embodiments of this application. Network equipment can be base stations and / or core network elements. The aforementioned base stations can be 5G and later versions of base stations (e.g., gNB, 5G NR NB, etc.), or base stations in other communication systems (e.g., eNB, WLAN access point, or other access point, etc.). The base station can be referred to as Node B, Evolved Node B, Access Point, Base Transceiver Station (BTS), Radio Base Station, Radio Transceiver, Basic Service Set (BSS), Extended Service Set (ESS), B Node, Evolved B Node (eNB), Home B Node, Home Evolved B Node, WLAN Access Point, WiFi Node, or any other suitable term in the field, as long as the same technical effect is achieved. The base station is not limited to specific technical terms. It should be noted that in the embodiments of this application, only the base station in the NR system is used as an example, but the specific type of base station is not limited.

[0067] Base stations can communicate with terminals under the control of a base station controller, which in various examples may be part of the core network or some base stations. Some base stations may communicate control information or user data with the core network via backhaul. In some examples, some of these base stations may communicate with each other directly or indirectly via backhaul links, which may be wired or wireless communication links. Wireless communication systems can support operation on multiple carriers (waveform signals of different frequencies). Multi-carrier transmitters can transmit modulated signals on multiple carriers simultaneously. For example, each communication link may be a multi-carrier signal modulated according to various radio technologies. Each modulated signal may be transmitted on a different carrier and may carry control information (e.g., reference signals, control channels, etc.), overhead information, data, etc.

[0068] Base stations can wirelessly communicate with terminals via one or more access point antennas. Each base station can provide communication coverage for its respective coverage area. The coverage area of ​​an access point can be divided into sectors that constitute only a part of that coverage area. Wireless communication systems may include different types of base stations (e.g., macro base stations, micro base stations, or pico base stations). Base stations may also utilize different radio technologies, such as cellular or WLAN radio access technologies. Base stations may be associated with the same or different access networks or operator deployments. The coverage areas of different base stations (including coverage areas of the same or different types of base stations, coverage areas using the same or different radio technologies, or coverage areas belonging to the same or different access networks) may overlap.

[0069] Communication links in a wireless communication system may include an uplink for carrying uplink (UL) transmissions (e.g., from a terminal to a network device) or a downlink for carrying downlink (DL) transmissions (e.g., from a network device to a terminal). UL transmissions may also be referred to as reverse link transmissions, while DL transmissions may also be referred to as forward link transmissions. Downlink transmissions may be carried out using licensed frequency bands, unlicensed frequency bands, or both. Similarly, uplink transmissions may be carried out using licensed frequency bands, unlicensed frequency bands, or both.

[0070] Before describing the specific technical solution, an example of the near-field XL-MIMO uplink system model to which this application applies will be given first. The base station is configured with a uniform planar array, containing... One horizontal antenna and There are [number] vertical antennas, and the total number of receiving antennas is [number]. The antenna spacing is set to half a wavelength. ,in Let be the carrier wavelength. Assume the system has a total of The number of single-antenna users, the first The symbol sent by each user is Then the transmitted symbol vector composed of all users is Each user transmits symbols independently and with unit average symbol energy. The base station receives the following signal:

[0071]

[0072] in, The received signal vector, For the uplink channel matrix, This refers to the number of receiving antennas at the base station. The signal vector sent to the user, It is an additive white Gaussian noise vector. The variance of the complex Gaussian noise is... It is an identity matrix.

[0073] In near-field scenarios, the first The near-field spatial non-stationary channel for an individual user can be modeled as follows:

[0074]

[0075] in, For the first Channel vectors for each user Rician factor For direct path channel components, This refers to the channel components of the non-direct path.

[0076] Direct path channel components Represented as:

[0077]

[0078] Non-direct path channel components Represented as:

[0079]

[0080] In the above expression, Indicates user Propagation distance to the center of the antenna array Indicates wave number, Indicates the number of scattering clusters. Indicates user No. Small-scale fading coefficients of each path Indicates user To the The distance between the scatterers Indicates the first The propagation distance from each scatterer to the center of the antenna array. In the near-field spherical wave model, the array steering vector... Not only dependent on horizontal angle and pitch angle It also depends on the distance of transmission. It is represented as:

[0081]

[0082] in, For users To the array The distance between the antennas is calculated using the following formula:

[0083]

[0084] in, For the first The three-dimensional coordinates of the antenna. It should be noted that for a ULA array, the array steering vector is represented as... It is unaffected by pitch angle.

[0085] In this application, a spatially non-stationary matrix is ​​introduced to address the spatial non-stationarity of the XL-MIMO channel. , which is represented as ,in Indicates user No. The NLOS path in the The visibility weights at each receiving antenna are used to describe the differences in energy distribution of the user signal at different array locations:

[0086]

[0087] in, Indicates user No. The visible antenna region with a scattering path. Indicates the first From one antenna to the visible antenna area The distance span, This is the attenuation control factor outside the region, typically set to 13. , representing the diagonal span of the Uniform Planar Array (UPA) antenna index network. This modeling method is used to characterize the spatial non-stationary phenomena in XL-MIMO arrays, such as different visible areas for different users and uneven energy distribution within the array.

[0088] Based on the above system model, the specific process of the method in this embodiment includes the following steps:

[0089] Please refer to Figure 1This application provides an uplink multi-user signal detection method suitable for XL-MIMO, comprising:

[0090] Step S101: Perform channel estimation based on the received uplink pilot signal to obtain channel estimation information, and obtain the location information of each user to the base station antenna array;

[0091] Step S102: Based on the channel estimation information and the location information, obtain the energy similarity, location similarity and channel correlation between users respectively, and construct a grouped joint metric to characterize the comprehensive similarity between users;

[0092] Step S103: Based on the grouping joint metric, the users are grouped to obtain multiple user groups, and the set of active antennas corresponding to each user group is determined.

[0093] Step S104: In one inter-group iteration, each user group is processed sequentially. For the current user group, based on the detection results of the user groups that have been processed in this iteration and the detection results of the unprocessed user groups in the previous iteration, an approximate expectation propagation iterative detection is performed in the current user group, and the detection results corresponding to the current user group are updated.

[0094] Step S105: If the inter-group iteration satisfies the preset termination condition, then the detection results of each user group are output.

[0095] In this embodiment, the base station can receive the known pilot sequences (i.e., uplink pilot signals) sent by each user, and use estimation algorithms such as least squares or least mean square error to obtain the estimated channel matrix and channel estimation information for detection. This channel estimation information includes the amplitude and phase response of the user on the large-scale antenna array. Simultaneously, the base station also needs to acquire the three-dimensional position information of each user relative to the center of the antenna array. In near-field scenarios, this position information typically includes the user's radial distance, azimuth angle, and elevation angle. Acquiring the channel estimation information and position information provides the necessary data foundation for subsequent joint grouping using channel energy distribution and spatial geometric features. It should be noted that the methods for acquiring position information include, but are not limited to, pilot estimation; it can also be obtained through independent positioning reference signals or sensor-aided information, as long as it reflects the spatial distribution of users in the near-field region.

[0096] Then, due to the significant spatial non-stationarity of near-field XL-MIMO channels, single-dimensional similarity metrics are often insufficient to accurately characterize the interference relationships between users. Therefore, this embodiment constructs a joint metric from three complementary dimensions: energy similarity reflects the degree of overlap of users in the visible area of ​​the antenna array; positional similarity characterizes the geometric proximity of users in the near-field three-dimensional space; and channel correlation directly characterizes the orthogonality or interference intensity between user channel vectors. By fusing information from these three dimensions to construct a grouped joint metric, erroneous clustering caused by relying on only a single indicator can be avoided. This ensures that subsequent grouping results can guarantee that users within a group share similar channel structures while maintaining sufficient separability between groups, thus creating favorable conditions for subsequent inter-group iterative detection.

[0097] Furthermore, joint grouping of users and antennas is performed. On one hand, users with similar channel characteristics and spatial locations are grouped into the same user group based on a joint grouping metric, making users within the same group more likely to generate strong signals on the same set of antennas. On the other hand, for each user group, the antennas that contribute the most to the signal of the group are selected as the active antenna set based on the energy concentration areas of its members on the array. This joint grouping mechanism decomposes the high-dimensional detection problem, which originally operated on a large-scale global channel matrix, into several low-dimensional local detection sub-problems. Each local sub-problem involves only one user group and its corresponding active antenna set, thus significantly reducing the computational dimensionality of subsequent signal detection from the source.

[0098] Furthermore, this embodiment constructs a hierarchical detection architecture that combines outer-layer inter-group serial interference cancellation with inner-layer intra-group nonlinear iterative detection. In the outer-layer iteration, each user group is processed sequentially according to a predetermined order. Using the detection results of the already processed user groups, including symbol estimation results, inter-group interference processing is performed on the currently unprocessed user groups. In the inner-layer detection, based on the local channel matrix corresponding to the active antenna set of the current user group, approximate expectation propagation iterative detection is performed. Since the dimension of the local channel matrix is ​​much smaller than the global channel matrix, and inter-group interference has been partially suppressed in the outer-layer iteration, the inner-layer approximate expectation propagation algorithm can converge quickly with low computational overhead and obtain high-precision detection results, which are then used as the basis for the next round of inter-group interference cancellation. This hierarchical collaborative mechanism effectively solves the problems of excessive complexity in traditional global nonlinear detection and insufficient inter-group interference suppression in existing local detection schemes.

[0099] Finally, a preset termination condition is set to determine whether the outer iteration has converged. This can be achieved by reaching a preset maximum number of inter-group iterations, or by the difference between the detection results obtained from two adjacent iterations being less than a preset threshold. In this embodiment, reaching the preset maximum number of inter-group iterations is taken as an example. When the termination condition is met, the system uses the latest symbol estimation results of each group as the final output, thus completing the uplink multi-user signal detection process.

[0100] In summary, the embodiments of this application establish a clear technical logic chain for joint grouping to serve hierarchical detection. Through multi-dimensional feature fusion and user-antenna joint grouping, the high-dimensional global problem is precisely decomposed into low-dimensional local sub-problems. Based on this, through the nested execution of outer-layer interference cancellation and inner-layer fine detection, near-globally optimal detection performance is achieved in the low-dimensional subspace. This design not only fully exploits the spatial non-stationary structure of the near-field XL-MIMO channel and reduces computational complexity, but also achieves an effective trade-off between detection complexity and bit error rate performance.

[0101] Preferably, in the uplink multi-user signal detection method for XL-MIMO as described above, the step of obtaining energy similarity, location similarity, and channel correlation among users based on the channel estimation information and the location information, and constructing a grouped joint metric to characterize the comprehensive similarity among users, includes:

[0102] Based on the channel estimation information matrix, the normalized energy vector of each user is obtained, and the energy similarity between any two users is calculated.

[0103] Based on the location information, calculate the location similarity between any two users;

[0104] Based on the channel estimation information, calculate the channel correlation between any two users;

[0105] The group joint metric is constructed based on the energy similarity, the location similarity, and the channel correlation.

[0106] This implementation provides an example of the steps for obtaining the group joint metric. First, the entire base station array is divided into... The nth subarray is used to calculate the energy vector of each user on each subarray. Let the nth subarray be the energy vector of each user on each subarray. The antenna set of each subarray is Then the user The energy on this subarray is:

[0107]

[0108] in, For users to the subarray The channel coefficient of the root antenna.

[0109] Based on this, users are constructed. Normalized energy vector:

[0110]

[0111] in, For users In the Signal energy on each subarray This represents the total number of subarrays.

[0112] Based on the above normalized energy vector, the user and users The energy similarity between them can be expressed as:

[0113]

[0114] in, and users respectively and users The normalized energy vector. This energy similarity reflects the degree of overlap between the visible areas of two users on the array; higher overlap indicates stronger potential interference.

[0115] Secondly, the positional similarity between users is calculated based on location information such as the user's three-dimensional spatial coordinates. Its physical meaning lies in characterizing the geometric proximity of users in the near field region. The higher the positional similarity, the closer the users' physical locations are, and thus they have similar spherical wavefront features.

[0116] Then, channel correlation is calculated directly based on the channel estimation vector. Its physical meaning lies in characterizing the orthogonality between user channels or the actual interference intensity, which can be specifically expressed as:

[0117]

[0118] in, For users The channel vector, For users The channel vector, This indicates the conjugate transpose operation.

[0119] Finally, to balance intra-group similarity and detectability, the information from the three dimensions mentioned above is weighted or linearly combined to construct a joint grouping metric:

[0120]

[0121] This joint metric incorporates energy similarity. Location similarity Channel correlation Three dimensions can avoid erroneous clustering caused by relying on only a single indicator (such as relying only on angle or only on energy), so that subsequent grouping results can ensure that users within the group share similar channel structures and maintain sufficient separability between groups.

[0122] It should be noted that this application does not limit the order of the steps for obtaining energy similarity, location similarity and channel correlation, and they can be processed synchronously or asynchronously.

[0123] See Figure 2 Specifically, in the uplink multi-user signal detection method for XL-MIMO described above, the step of calculating the location similarity between any two users based on the location information includes:

[0124] Step S201: Based on the location information, normalize the distance difference, azimuth difference, and pitch difference between any two users to obtain a normalized location difference quantity. The location information includes: the user's estimated distance, estimated azimuth, and estimated pitch.

[0125] Step S202: Perform kernel function mapping based on the normalized position difference to obtain the bounded position similarity.

[0126] This application provides a specific example of the method for obtaining location similarity. Since radial distance is a core physical feature determining the spatial non-stationarity of the channel in near-field communication scenarios, this embodiment introduces radial distance difference as a calculation element for location similarity. The user's location information is used as the distance... Azimuth and pitch angle For example, location information is specifically represented as Then the user and users The dimensionless normalized positional difference between them can be expressed as:

[0127]

[0128]

[0129]

[0130] in, To normalize the distance difference, To normalize the azimuth difference, To normalize the pitch angle difference, and users respectively and users Distance estimate, and These are azimuth difference and elevation difference, respectively. , , This is a normalization factor for the distance, azimuth, and elevation dimensions. It is important to note that the location difference calculation here must be based on parameters from two different users, not duplicate parameters from the same user.

[0131] Subsequently, a kernel function is used to exponentially map the sum of squares of the normalized differences to obtain the positional similarity. Taking the Gaussian kernel function as an example, its specific expression is as follows:

[0132]

[0133] This metric integrates radial distance difference, azimuth difference, and pitch difference, and maps them using a Gaussian kernel function. Similarity values ​​between intervals.

[0134] This mapping transforms unbounded geometric differences into positional similarity values ​​within the [0,1] interval. This mapping method not only smoothly reflects the proximity relationship between users but also better corresponds to their proximity in the visible area on the antenna array. Compared to the simple Euclidean distance criterion, it is more suitable for group decision-making in near-field spatial non-stationary channels.

[0135] See Figure 3 Preferably, in the uplink multi-user signal detection method for XL-MIMO as described above, the grouping of the users based on the grouped joint metric to obtain multiple user groups includes:

[0136] Step S301: Based on the grouped joint metric, calculate the local density and global coupling degree of each user;

[0137] Step S302: Based on the difference between the local density and the global coupling degree, construct a representative metric for each user;

[0138] Step S303: Determine a preset number of seed users based on the representativeness metric and the grouped joint metric;

[0139] Step S304: Take each seed user as the group center of a user group, and calculate the intra-group affinity between the candidate user and each user group, and the inter-group coupling degree of the candidate user in one user group pointing to other user groups.

[0140] Step S305: Based on the allocation metric constructed from the difference between the intra-group affinity and the inter-group coupling, the candidate user is assigned to the target user group that maximizes the allocation metric.

[0141] In this embodiment, an example is given of how to group the users based on the grouping joint metric to obtain multiple user groups. Specifically, let the first... The neighborhood set of a user is composed of the users closest to it under the joint metric. Composed of individual users, then users Local density It can be represented as:

[0142]

[0143] in, For the number of users in the neighboring area, For users The set of neighboring users For users With users Grouping joint metric between them.

[0144] At the same time, to avoid the selected seed users maintaining a high degree of similarity with a large number of users globally, users The global coupling degree can be expressed as:

[0145]

[0146] in, This represents the total number of users.

[0147] Users can be constructed based on local density and global coupling. The representativeness measure can be expressed as .

[0148] Local density measures a user's representativeness to similar users in their neighborhood; higher density indicates a user is more likely to serve as the center of a local cluster. Global coupling measures the average similarity between a user and the entire user set; high coupling suggests the user may be located at the boundary of multiple clusters or belong to globally uniformly distributed background noise, making them unsuitable as the center for distinguishing different user groups. The difference between these two measures yields a representativeness metric. This metric, combined with the grouping joint metric, facilitates the selection of high-quality seed users who are both highly representative locally and unique globally. In one specific embodiment, limiting the number of seed users to a preset limit avoids the impact of too many or too few groups on the grouping results.

[0149] For the allocation of non-seed users, let the first... Each user group is denoted as For candidate users Define its pointer to the first The in-group affinity of the group is:

[0150]

[0151] in, For the first The number of users in each user group For users With users Grouping joint measure. Intragroup affinity. Reflecting users The higher the average match between the user and the existing members of the group, the better the user's match. The more consistent one's internal structure is with the group, the more suitable one is to join the group.

[0152] At the same time, to prevent users If a group maintains a strong association with other groups after joining a certain group, its pointer is defined as pointing to the first group. The inter-group coupling degree is:

[0153]

[0154] in: Total number of user groups. Inter-group coupling. Used to depict when users Assigned to the When grouped, this value represents the average similarity between the user and other groups. A higher value indicates stronger coupling between the user and other groups, which is not conducive to forming clear inter-group boundaries.

[0155] Then, an assignment metric is constructed based on the difference between intra-group affinity and inter-group coupling: The larger this value, the stronger the user's... Join the Grouping is more conducive to enhancing intra-group consistency and reducing inter-group coupling.

[0156] user The final grouping results are determined according to the following criteria:

[0157]

[0158] in, Indicates that the user The allocation metric is the target user group with the highest allocation value. Here, intra-group affinity reflects the average matching degree between the user to be assigned and the existing members of the target group, while inter-group coupling reflects the average association strength between the user and all other non-target groups. By maximizing the allocation metric, users can be guided to the target group that best matches them and has the least interference from other groups, thus achieving a grouping effect of "tight aggregation within groups and clear separation between groups," effectively reducing the interference coupling burden in subsequent signal detection.

[0159] See Figure 4 Preferably, in the uplink multi-user signal detection method for XL-MIMO as described above, determining a preset number of seed users based on the representativeness metric and the grouped joint metric includes:

[0160] Step S401: Select the user with the largest representative metric as the first seed user in the seed set;

[0161] Step S402: Obtain the representativeness metric of the candidate user minus the difference in similarity between the candidate user and the seed set;

[0162] Step S403: Determine the candidate user with the largest difference as the new seed user and store it in the seed set until a preset number of seed users are determined.

[0163] This embodiment illustrates the steps for determining a preset number of seed users. Specifically, it is a seed selection process based on a greedy strategy. The first seed user is selected by directly choosing the maximum representative metric, ensuring the priority establishment of the strongest cluster center.

[0164] Then, a separation constraint is introduced for the selection of subsequent seed users. The difference between the representativeness metric of the candidate user and the similarity between the candidate user and the seed set is denoted as the seed selection metric. Let the set of currently selected seeds be... For candidate users Obtain seed selection metric The formula can be expressed as:

[0165]

[0166] in, As a representativeness measure, the second term is the similarity between the candidate user and the selected seed set. That is, while considering the representativeness of the candidate user itself, the cumulative similarity between the candidate user and the selected seed set is deducted.

[0167] By maximizing This mechanism allows the algorithm to prioritize users who are both highly representative and significantly distinguishable from existing seeds as new seeds. It forces the algorithm to avoid already covered high-density areas and instead search for candidates that represent the new user cluster while maintaining sufficient distance from existing seeds. This avoids the problem of multiple seed users clustering in the same area or selecting isolated outliers, significantly improving the stability and rationality of the initial grouping centers.

[0168] See Figure 5 Preferably, in the uplink multi-user signal detection method for XL-MIMO as described above, determining the set of active antennas corresponding to each user group includes:

[0169] Step S501: Obtain the total energy of the target user group on each antenna subarray, and sort the total energy corresponding to each antenna subarray from largest to smallest to obtain a subarray sequence, wherein the target user group is any one of the user groups;

[0170] Step S502: Determine the number of subarrays required for the target user group based on the cumulative energy threshold criterion;

[0171] Step S503: Select target subarrays corresponding to the number of subarrays sequentially from the subarray sequence, and take the union of the antennas contained in the target subarrays as the active antenna set of the target user group.

[0172] This embodiment illustrates the steps for determining the active antenna set corresponding to each user group. Due to the spatial non-stationarity of the near-field channel, the signal energy of each user group is typically concentrated only in a portion of the antenna region. Therefore, obtaining the set of active antennas for each user group... The target user group (i.e., the target user group) in the th The total energy across the subarrays is Then, sort the subarrays in descending order of total energy to obtain the subarray sequence. Let the index of the sorted subarray sequence be denoted as . .

[0173] Then, based on the cumulative energy threshold criterion, the number of subarrays required for the target user group is determined. The formula for the cumulative energy threshold criterion is expressed as:

[0174]

[0175] in, The preset energy threshold, For the sorted number The group energy of each subarray This represents the total number of subarrays. Then, the corresponding target subarray can be selected sequentially from the subarray sequence based on this number of subarrays.

[0176] For example, if the cumulative energy threshold η=0.9 is set, then the top-ranked subarrays are selected as target subarrays, and their cumulative energy accounts for more than 90% of the total energy of the group. Meanwhile, the edge subarrays with a remaining energy ratio of less than 10% are regarded as noise-dominated or inactive areas and are removed.

[0177] Finally, the union of the antennas contained in these selected target subarrays is taken, which constitutes the set of active antennas for this user group. This dynamic antenna selection mechanism based on energy concentration ensures that the local detection matrix retains most of the effective signal components while significantly reducing dimensionality, avoiding the performance loss caused by fixed antenna grouping, and providing a low-dimensional and high signal-to-noise ratio input model for subsequent hierarchical iterative detection.

[0178] See Figure 6 Preferably, in the uplink multi-user signal detection method for XL-MIMO as described above, the step of performing approximate expected propagation iterative detection within the current user group based on the detection results of the user group processed in the current iteration and the detection results of the unprocessed user group in the previous iteration, and updating the detection results corresponding to the current user group, includes:

[0179] Step S601: Subtract the interference generated on the active antenna set corresponding to the current user group by the detection results of the user group that have been processed in the current iteration and the detection results of the user group that have not been processed in the previous iteration from the received signal received by the base station to obtain the residual signal;

[0180] Step S602: Based on the residual signal and the channel of the current user group on the corresponding active antenna set, construct a corresponding local detection model, and perform the approximate expected propagation iterative detection within the local detection model to obtain the symbol soft estimate of the current user group;

[0181] Step S603: Perform constellation hard decision on the symbol soft estimation, update the detection result corresponding to the current user group based on the symbol estimation after hard decision, and mark the current user group as the processed user group.

[0182] In this embodiment, let , , respectively representing the first Number of users in a group and number of active antennas. Let the number of users in the group be [number of users in the group] and The local reception model corresponding to the group of user antennas is as follows:

[0183]

[0184] in, For the first Received signals on the active antennas For the first Group of users in the first Local channel matrix on active antennas, For the first The sent symbol vectors of the group users, For the first The sent symbol vectors of the group users, For the first Noise vector on the active antenna group.

[0185] In each round of iteration between outer groups, the system processes each user group one by one in a predetermined order (e.g., the order in which seed users are determined). For the th The first outer iteration The residual signal of the group is expressed as:

[0186]

[0187] in, For the first The symbol estimation of the user groups processed in the next iteration is the detection result. For the first Symbol estimation for user groups not processed in the next iteration.

[0188] In other words, for the user group currently awaiting processing, the system utilizes the latest detection results of previously processed user groups, combined with the local channel matrix of these processed groups on the active antenna set corresponding to the current target user group, to reconstruct the interference components they generate in that local antenna region, and subtracts them from the corresponding components of the original received signal. It is crucial to emphasize that the interference reconstruction and cancellation operations here are strictly limited to the dimension corresponding to the "active antenna set," rather than using the global channel matrix. Performing interference cancellation only on the active antenna set avoids introducing additional noise enhancement effects and reduces the computational complexity of interference cancellation from the global antenna count to the local active antenna count, achieving precise dimensionality reduction at the physical level. The residual signal obtained after the above processing can be considered as an equivalent local observation signal containing only the current target user group signal, residual inter-group interference, and noise.

[0189] Furthermore, using the residual signal obtained above and the local channel matrix of the current user group on its active antenna set, a local detection model with significantly reduced dimensionality is constructed:

[0190]

[0191] in, This includes noise and residual inter-group interference.

[0192] Then, within this local detection model, the AEP algorithm is used for refined nonlinear detection. The core idea of ​​the AEP algorithm is based on the maximum a posteriori probability criterion, using a Gaussian distribution to approximate the discrete posterior distribution of the transmitted symbols, and using the moment matching rule to continuously update the mean and variance parameters of the approximate distribution during the iteration process, making it gradually approach the true posterior distribution. Compared with performing global expectation propagation detection directly on the complete large-scale channel matrix, this embodiment performs AEP iteration within a local low-dimensional model, which has significant convergence advantages: on the one hand, the condition number of the local channel matrix is ​​usually better than that of the global matrix, resulting in better numerical stability in the moment matching process; on the other hand, since strong inter-group interference has been partially eliminated in the step of acquiring the residual signal, the equivalent signal-to-noise ratio within the local model is improved, further accelerating the convergence speed of the AEP algorithm and improving the accuracy of soft estimation.

[0193] It should be noted that although this embodiment uses approximate expectation propagation as an example, other nonlinear iterative detection algorithms based on Gaussian approximation or message passing can be used instead in other embodiments, as long as they can use the local detection model for soft symbol estimation.

[0194] Furthermore, after completing a preset number of inner AEP iterations or satisfying the convergence condition, the final soft symbol estimate is mapped to the nearest discrete constellation point to obtain the hard decision symbol estimate for the current user group. This hard decision result not only serves as the final detection output for the current user group but is also immediately used in the interference cancellation process for subsequent user groups. Then, the current user group is marked as "processed." This serial update mechanism ensures that subsequent user groups always have access to the latest and more accurate interference cancellation information, thereby progressively improving overall detection performance.

[0195] Preferably, the uplink multi-user signal detection method for XL-MIMO described above further includes, before sequentially processing each user group within one inter-group iteration:

[0196] Based on the received signal and the channel estimation information, the initial detection results for each user are obtained through low-complexity linear minimum mean square error detection.

[0197] Specifically, nonlinear iterative detection algorithms are typically sensitive to initial values. If iteration starts directly from a zero vector or a random vector, it is prone to getting trapped in local optima or resulting in slow convergence. In this embodiment, before entering the hierarchical iterative loop, a low-complexity linear minimum mean square error (LMMSE) detector is used to perform a coarse estimate of the transmitted symbols for all users:

[0198]

[0199] in, This is the initial test result. For the uplink channel matrix, For noise variance, for 3D identity matrix This is the received signal vector.

[0200] While LMMSE detection cannot completely eliminate multi-user interference, it has low computational complexity and provides relatively reliable soft information. Using this initial detection result as the starting point for subsequent inter-group serial interference cancellation and intra-group approximate expectation propagation iterations is equivalent to providing a "hot start" location within the neighborhood of the global optimum for nonlinear detection. This design not only significantly reduces the number of subsequent nonlinear iterations required to converge but also improves the algorithm's robustness under different signal-to-noise ratio conditions, ensuring the stability of the hierarchical detection architecture in practical engineering applications.

[0201] To verify the effectiveness of the hierarchical iterative detection scheme proposed in this embodiment, this application compares and illustrates it using several different signal detection algorithms.

[0202] First, a rigorous Monte Carlo simulation platform was built based on the standard 6G XL-MIMO communication environment for verification, and the channel modeling was as described above. The number of users per antenna was set, and the simulation parameters for the UPA array were uniformly set as follows: number of horizontal antennas, number of vertical antennas, total number of receiving antennas, antenna array spacing, carrier frequency, and the channel consisted of a direct path and 10 randomly generated scattering paths, with a distance range of 5 to 30 meters. During the detection process, the number of groups was 4, the outer layer Successive Interference Cancellation (SIC) iterations were 3, and the inner layer AEP iterations were 6.

[0203] See Figure 7 , Figure 7This diagram illustrates the bit error rate (BER) of different signal detection algorithms in an XL-MIMO multi-user uplink scenario. The horizontal axis represents the system's signal-to-noise ratio (SNR), ranging from 0 dB to 16 dB, and the vertical axis represents the bit error rate (BER). The diagram shows comparison curves for various schemes, including: the classic ZF and MMSE detection algorithms, the EP detection algorithm, the Partial Zero-Forcing (PZF) detection algorithm based on visible region (VR) overlap map grouping, the AEP-SIC detection algorithm based on VR overlap map grouping, and the AEP-SIC detection algorithm based on multi-dimensional similarity vector grouping proposed in this application. Combined with... Figure 7 As can be seen, under different signal-to-noise ratio (SNR) conditions, the bit error rate (BER) performance curve (magenta solid line) of the AEP-SIC detection method based on multidimensional similarity grouping proposed in this application is very close to that of the Expectation Propagation (EP) detection algorithm (red solid line), and the two almost overlap when the SNR is greater than 6 dB. Furthermore, it significantly outperforms the traditional ZF and MMSE linear detection algorithms. This indicates that through reasonable user-antenna joint grouping and inter-group interference cancellation, this application successfully retains the performance advantage of global nonlinear detection within a local detection framework. It should be noted that when obtaining the BER in this application, the detection result is first demodulated and bit recovered, and then compared one by one with the original bits from the transmitting end to count the total number of erroneous bits. Then, based on the bit error rate formula: The final bit error rate is obtained, where, This represents the total number of bits sent.

[0204] See Figure 8 , Figure 8 This diagram illustrates the normalized computational complexity of different detection algorithms. The horizontal axis represents the change in the number of receiving antennas (64 users per antenna, 4 groups), and the vertical axis represents the normalized computational complexity. The diagram shows comparison curves for various schemes, including: the classic ZF detection algorithm and MMSE detection algorithm, the global EP detection algorithm, the AEP-SIC detection algorithm based on VR overlap map grouping, and the AEP-SIC detection algorithm based on multi-dimensional similarity vector grouping proposed in this application. Figure 8As can be seen, with the increase of the number of receiving antennas, the normalized complexity of the global EP detection algorithm remains constant at 1.0, while the complexity of the method in this application (yellow solid line) stabilizes at around 0.19, which is less than one-fifth of that of the global EP, and significantly lower than the comparative scheme (purple solid line) that does not adopt the grouping strategy of this application. This result fully demonstrates that this embodiment, through the design of "local channel extraction" and "hierarchical iterative collaboration", significantly reduces the computational complexity of XL-MIMO uplink signal detection while ensuring detection performance, achieving the optimal trade-off between performance and efficiency.

[0205] The various methods of the embodiments of this application have been described above. Apparatus for implementing the above methods will now be provided.

[0206] Please refer to Figure 9 Another embodiment of this application provides a control device for uplink multi-user signal detection suitable for XL-MIMO, comprising:

[0207] The first processing module 901 is used to perform channel estimation based on the received uplink pilot signal, obtain channel estimation information, and acquire the location information of each user to the base station antenna array.

[0208] The second processing module 902 is used to obtain the energy similarity, location similarity and channel correlation between users based on the channel estimation information and the location information, and to construct a grouped joint metric to characterize the comprehensive similarity between users.

[0209] The third processing module 903 is used to group each user based on the grouping joint metric to obtain multiple user groups and determine the set of active antennas corresponding to each user group.

[0210] The fourth processing module 904 is used to process each user group sequentially in one inter-group iteration, wherein, for the current user group, based on the detection results of the user groups that have been processed in the current iteration and the detection results of the unprocessed user groups in the previous iteration, an approximate expectation propagation iterative detection is performed in the current user group, and the detection results corresponding to the current user group are updated.

[0211] The fifth processing module 905 is used to output the detection results of each user group if the inter-group iteration meets the preset termination condition.

[0212] Preferably, in the control device for uplink multi-user signal detection of XL-MIMO as described above, the second processing module includes:

[0213] The first sub-processing module is used to obtain the normalized energy vector of each user based on the channel estimation information matrix, and to calculate the energy similarity between any two users.

[0214] The second sub-processing module is used to calculate the location similarity between any two users based on the location information;

[0215] The third sub-processing module is used to calculate the channel correlation between any two users based on the channel estimation information;

[0216] The fourth sub-processing module is used to construct the group joint metric based on the energy similarity, the location similarity, and the channel correlation.

[0217] Specifically, in the control device for uplink multi-user signal detection applicable to XL-MIMO as described above, the second sub-processing module is used to include:

[0218] The first processing unit is used to normalize the distance difference, azimuth difference and pitch difference between any two users based on the location information to obtain a normalized location difference quantity. The location information includes: the user's estimated distance, estimated azimuth and estimated pitch.

[0219] The second processing unit is used to perform kernel function mapping based on the normalized position difference to obtain the bounded position similarity.

[0220] Preferably, in the control device for uplink multi-user signal detection of XL-MIMO as described above, the third processing module includes:

[0221] The fifth sub-processing module is used to calculate the local density and global coupling degree of each user based on the grouped joint metric;

[0222] The sixth sub-processing module is used to construct a representative metric for each user based on the difference between the local density and the global coupling degree.

[0223] The seventh sub-processing module is used to determine a preset number of seed users based on the representativeness metric and the grouped joint metric.

[0224] The eighth sub-processing module is used to take each seed user as the group center of a user group and calculate the intra-group affinity between the candidate user and each user group, as well as the inter-group coupling degree of the candidate user in one user group pointing to other user groups.

[0225] The ninth sub-processing module is used to assign the candidate user to the target user group that maximizes the assignment metric based on the allocation metric constructed from the difference between the intra-group affinity and the inter-group coupling.

[0226] Preferably, in the control device for uplink multi-user signal detection of XL-MIMO as described above, the seventh sub-processing module includes:

[0227] The third processing unit is used to select the user with the largest representative metric as the first seed user in the seed set.

[0228] The fourth processing unit is used to obtain the representativeness metric of the candidate user minus the difference in similarity between the candidate user and the seed set;

[0229] The fifth processing unit is used to determine the candidate user that makes the largest difference as the new seed user and store it in the seed set until a preset number of seed users are determined.

[0230] Preferably, in the control device for uplink multi-user signal detection of XL-MIMO as described above, the third processing module further includes:

[0231] The tenth sub-processing module is used to obtain the total energy of the target user group on each antenna subarray, and sort the total energy corresponding to each antenna subarray from large to small to obtain a subarray sequence, wherein the target user group is any one of the user groups;

[0232] The eleventh sub-processing module is used to determine the number of sub-arrays required for the target user group based on the cumulative energy threshold criterion.

[0233] The twelfth sub-processing module is used to sequentially select target subarrays corresponding to the number of subarrays from the subarray sequence, and to take the union of the antennas contained in the target subarrays as the active antenna set of the target user group.

[0234] Preferably, in the control device for uplink multi-user signal detection of XL-MIMO as described above, the fourth processing module includes:

[0235] The thirteenth sub-processing module is used to subtract the interference generated on the active antenna set corresponding to the current user group by the detection results of the user group that have been processed in the current iteration and the detection results of the user group that have not been processed in the previous iteration from the received signal received from the base station, so as to obtain the residual signal;

[0236] The fourteenth sub-processing module is used to construct a corresponding local detection model based on the residual signal and the channel of the current user group on the corresponding active antenna set, and to perform the approximate expected propagation iterative detection within the local detection model to obtain the symbol soft estimate of the current user group.

[0237] The fifteenth sub-processing module is used to perform constellation hard decision on the symbol soft estimation, update the detection result corresponding to the current user group according to the symbol estimation after hard decision, and mark the current user group as the processed user group.

[0238] Preferably, the control device for uplink multi-user signal detection for XL-MIMO as described above further includes:

[0239] The fifth processing module is used to obtain the initial detection results for each user based on the received signal and the channel estimation information through low-complexity linear minimum mean square error detection.

[0240] The apparatus embodiments of this application are apparatuses corresponding to the embodiments of the methods described above. All implementation means in the method embodiments described above are applicable to the apparatus embodiments and can achieve the same technical effects. The apparatus provided in this application embodiments can implement all the method steps implemented in the method embodiments described above and can achieve the same technical effects. Therefore, the parts and beneficial effects that are the same as those in the method embodiments in this embodiment will not be described in detail here.

[0241] Another embodiment of this application provides an electronic device, including: a processor, a memory, and a program stored in the memory and executable on the processor. When the program is executed by the processor, it implements the steps of the uplink multi-user signal detection method for XL-MIMO as described above, and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0242] Another embodiment of this application provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the steps of the uplink multi-user signal detection method for XL-MIMO described above, achieving the same technical effect. To avoid repetition, it will not be described again here. The computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0243] Another embodiment of this application provides a computer program product, including computer instructions that, when executed by a processor, implement the steps of the uplink multi-user signal detection method applicable to XL-MIMO as described above, and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0244] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0245] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0246] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method for uplink multi-user signal detection suitable for XL-MIMO, characterized in that, include: Channel estimation is performed based on the received uplink pilot signal to obtain channel estimation information, and the location information of each user to the base station antenna array is obtained. Based on the channel estimation information and the location information, the energy similarity, location similarity and channel correlation between users are obtained respectively, and a group joint metric for characterizing the comprehensive similarity between users is constructed. Based on the grouping joint metric, each user is grouped to obtain multiple user groups, and the set of active antennas corresponding to each user group is determined. In one inter-group iteration, each user group is processed sequentially. For the current user group, based on the detection results of the user groups that have been processed in this iteration and the detection results of the unprocessed user groups in the previous iteration, an approximate expectation propagation iterative detection is performed in the current user group, and the detection results corresponding to the current user group are updated. If the inter-group iterations meet the preset termination condition, the detection results for each user group are output.

2. The method according to claim 1, characterized in that, Based on the channel estimation information and the location information, the energy similarity, location similarity, and channel correlation between users are obtained respectively, and a grouped joint metric for characterizing the comprehensive similarity between users is constructed, including: Based on the channel estimation information matrix, the normalized energy vector of each user is obtained, and the energy similarity between any two users is calculated. Based on the location information, calculate the location similarity between any two users; Based on the channel estimation information, calculate the channel correlation between any two users; The group joint metric is constructed based on the energy similarity, the location similarity, and the channel correlation.

3. The method according to claim 2, characterized in that, The step of calculating the location similarity between any two users based on the location information includes: Based on the location information, the distance difference, azimuth difference, and pitch difference between any two users are normalized to obtain a normalized location difference quantity. The location information includes: the user's estimated distance, estimated azimuth, and estimated pitch. The kernel function is used to map the normalized positional differences to obtain the bounded positional similarity.

4. The method according to claim 1, characterized in that, The users are grouped based on the grouping joint metric to obtain multiple user groups, including: Based on the grouped joint metric, calculate the local density and global coupling degree of each user; Based on the difference between the local density and the global coupling degree, a representative metric for each user is constructed; Based on the representativeness metric and the grouped joint metric, a preset number of seed users are determined; Each seed user is used as the group center of a user group, and the intra-group affinity of the candidate user with each user group and the inter-group coupling of the candidate user within one user group to other user groups are calculated. Based on the allocation metric constructed from the difference between the intra-group affinity and the inter-group coupling, the candidate user is assigned to the target user group that maximizes the allocation metric.

5. The method according to claim 4, characterized in that, The step of determining a preset number of seed users based on the representativeness metric and the grouped joint metric includes: The user with the highest representative metric is selected as the first seed user in the seed set. The representativeness metric of the candidate user is subtracted from the difference in similarity between the candidate user and the seed set; The candidate user with the largest difference is identified as the new seed user and stored in the seed set until a preset number of seed users are determined.

6. The method according to claim 1, characterized in that, Determining the set of active antennas corresponding to each user group includes: The total energy of the target user group on each antenna subarray is obtained, and the total energy corresponding to each antenna subarray is sorted from largest to smallest to obtain a subarray sequence, wherein the target user group is any one of the user groups; The number of subarrays required for the target user group is determined based on the cumulative energy threshold criterion. Target subarrays corresponding to the number of subarrays are selected sequentially from the subarray sequence, and the union of the antennas contained in the target subarrays is taken as the active antenna set of the target user group.

7. The method according to claim 1, characterized in that, The step of performing approximate expectation propagation iterative detection within the current user group based on the detection results of the user group already processed in the current iteration and the detection results of the unprocessed user group in the previous iteration, and updating the detection results corresponding to the current user group, includes: The residual signal is obtained by subtracting the interference generated on the active antenna set corresponding to the current user group by the detection results of the user group that have been processed in the current iteration and the detection results of the user group that have not been processed in the previous iteration from the received signal received by the base station. Based on the residual signal and the channel of the current user group on the corresponding active antenna set, a corresponding local detection model is constructed, and the approximate expected propagation iterative detection is performed within the local detection model to obtain the symbol soft estimate of the current user group. Perform constellation hard decision on the symbol soft estimate, update the detection result corresponding to the current user group based on the symbol estimate after hard decision, and mark the current user group as the processed user group.

8. The method according to claim 7, characterized in that, Before processing each user group sequentially within one inter-group iteration, the process further includes: Based on the received signal and the channel estimation information, the initial detection results for each user are obtained through low-complexity linear minimum mean square error detection.

9. A control device for uplink multi-user signal detection suitable for XL-MIMO, characterized in that, include: The first processing module is used to perform channel estimation based on the received uplink pilot signal, obtain channel estimation information, and acquire the location information of each user to the base station antenna array. The second processing module is used to obtain the energy similarity, location similarity and channel correlation between users based on the channel estimation information and the location information, and to construct a grouped joint metric to characterize the comprehensive similarity between users. The third processing module is used to group each user based on the grouping joint metric to obtain multiple user groups, and to determine the set of active antennas corresponding to each user group. The fourth processing module is used to process each user group sequentially in one inter-group iteration, wherein, for the current user group, based on the detection results of the user groups that have been processed in the current iteration and the detection results of the unprocessed user groups in the previous iteration, an approximate expectation propagation iterative detection is performed in the current user group, and the detection results corresponding to the current user group are updated. The fifth processing module is used to output the detection results of each user group if the inter-group iteration meets the preset termination condition.

10. An electronic device, characterized in that, include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the uplink multi-user signal detection method for XL-MIMO as described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the uplink multi-user signal detection method for XL-MIMO as described in any one of claims 1 to 8.

12. A computer program product, characterized in that, Includes computer instructions that, when executed by a processor, implement the steps of the uplink multi-user signal detection method for XL-MIMO as described in any one of claims 1 to 8.

Images