Downlink channel reconstruction and user positioning method for U6G FDD ultra-large scale MIMO

Abstract

The invention discloses a U6G FDD (Frequency Division Duplexing) super-large-scale MIMO (Multiple Input Multiple Output)-oriented downlink channel reconstruction and user positioning method, which comprises the following steps that: firstly, a user sends an uplink pilot signal to a base station, designs a Cartesian domain codebook-based channel image generation method, and visualizes an uplink received signal into a channel image; secondly, detecting the positions of a user and a scatterer in a channel image by using a target detection network, finely adjusting the detected coordinates, performing uplink path gain estimation by using a least square algorithm, and positioning the user according to the maximum uplink path gain; then designing a downlink beam forming matrix based on a base vector, adjusting the direction of a downlink pilot beam by using the matrix, and estimating a downlink path gain according to a downlink receiving signal; and finally, completing reconstruction of a downlink channel according to the coordinates of the user and the scatterer and the estimated downlink path gain. According to the method, rapid and high-precision multi-cluster multi-path downlink channel reconstruction and user positioning can be realized with relatively low calculation complexity and pilot frequency overhead.

Description

Technical Field

[0001] The present invention relates to a method for downlink channel reconstruction and user positioning for U6G FDD ultra-large-scale MIMO, belonging to the field of wireless communication technology. Background Art

[0002] At present, globally, the sixth-generation mobile communication technology (6G) has set off a research upsurge. As a key force leading future communication reforms, 6G is highly anticipated. 6G is expected to achieve lower latency, higher data transmission rates, and higher spectral efficiency, bringing higher network speeds, stronger connection densities, and more excellent intelligent levels. With the continuous in-depth research of 6G, the ultra-large-scale MIMO technology in the U6G frequency band has become one of the key areas of 6G wireless communication research. In the U6G frequency band, it has the characteristics of high frequency, large bandwidth, and short wavelength, which creates favorable conditions for the hardware deployment of centralized MIMO. The ultra-large-scale MIMO technology, with a larger number of antennas deployed than large-scale MIMO, is expected to significantly improve the capacity and performance of the communication system, bringing a better communication experience for users. However, the practical application of ultra-large-scale MIMO technology in the U6G frequency band faces many challenges.

[0003] Generally speaking, the frequency division duplexing (FDD) system has lower latency and higher spectral utilization rate, and is suitable for simultaneously transmitting a large amount of user data streams in the 6G system. In the FDD mode, the problems of downlink channel reconstruction and user positioning need to be solved urgently. The spatial reciprocity in the FDD system can be used for downlink channel reconstruction, that is, estimating the frequency-independent parameters shared by the uplink and downlink on the uplink. In the U6G urban scenario, the wireless communication environment is complex and usually has the characteristics of dense scattering. Traditional channel models are difficult to accurately describe the actual channel situation. Therefore, the multi-cluster multi-path channel model is more in line with the channels in the actual scenario. However, accurate channel estimation in the multi-path multi-cluster scenario often requires a large number of pilots, resulting in high overhead. In addition, with the increase in the number of antennas in the ultra-large-scale MIMO system, the near-field effect becomes more obvious. Therefore, it is crucial to consider the spherical wavefront in the channel model. The newly emerging channel characteristics bring challenges to the downlink channel reconstruction problem. The channel characteristics in the near-field region are very different from those in the far-field. Traditional far-field-based channel estimation algorithms, such as algorithms using angular domain sparsity, are difficult to be directly applied. Moreover, the significant increase in the number of antennas makes the input data dimension increase sharply. Some deep learning methods that do not introduce channel sparsity characteristics, such as designing neural networks to directly learn the channel from the received signal, have poor fitting effects, and the channel estimation accuracy cannot meet the communication requirements. In the existing research on the near-field region, some methods estimate the channel by introducing a new codebook, but a large amount of calculation is required during the codebook search process, resulting in serious waste of computing resources. To solve this problem, it is necessary to combine technologies such as deep learning and image processing to reduce the computational complexity and improve the channel reconstruction efficiency.

[0004] In order to break through these technical bottlenecks and promote the development of FDD ultra-large-scale MIMO systems in the U6G frequency band, innovative downlink channel reconstruction and user positioning methods have become the key research directions. Combining the powerful feature extraction ability of deep learning and the intuitive expression advantage of image processing technology for channel information, it is expected to design more efficient and accurate downlink channel reconstruction and user positioning schemes. At the same time, designing low-complexity algorithms to reduce computational costs and resource consumption is the key to realizing the wide application of ultra-large-scale MIMO technology in the U6G frequency band. In-depth research on these technologies can not only improve the performance of FDD ultra-large-scale MIMO systems in the U6G frequency band but also provide solid technical support for the application of 6G communication in complex scenarios, helping 6G technology achieve broader commercial and social values. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a downlink channel reconstruction and user positioning method for U6G FDD ultra-large-scale MIMO, which can achieve fast and high-precision multi-cluster multipath downlink channel reconstruction and user positioning in the U6G frequency band in the FDD mode with relatively low computational complexity and pilot overhead.

[0006] The present invention adopts the following technical solutions to solve the above technical problems:

[0007] A downlink channel reconstruction and user positioning method for U6G FDD ultra-large-scale MIMO. The ultra-large-scale MIMO system includes a base station and a user. The base station is configured with N uniformly spaced linear array antennas, and the user is configured with a single antenna. There are L clusters between the base station and the user, and each cluster has S + 1 scatterers. The central scatterer of each cluster is the main path scatterer, and the other scatterers are the side path scatterers. The method includes the following steps:

[0008] Step 1: The user sends an uplink pilot signal to the base station, and the base station visualizes the received uplink pilot signal as a Cartesian domain channel image.

[0009] Step 2: Use the pre-trained object detection network CenterNet to roughly estimate the positions of the user and scatterers in the Cartesian domain channel image, obtain the position coordinates detected by the network, and convert the position coordinates detected by the network into roughly estimated position coordinates in the rectangular coordinate system.

[0010] Step 3: Use orthogonal matching pursuit optimization to optimize the roughly estimated position coordinates obtained in Step 2 to obtain optimized position coordinates.

[0011] Step 4: Based on the optimized position coordinates obtained in Step 3 and the uplink pilot signals received by the base station, use the least squares algorithm to obtain all uplink path gains; take the largest uplink path gain among all uplink path gains as the uplink path gain of the user, and perform user coordinate positioning according to the uplink path gain of the user.

[0012] Step 5: The base station designs a downlink beamforming matrix based on basis vectors based on the optimized position coordinates obtained in Step 3 and all uplink path gains obtained in Step 4, adjusts the downlink pilot beam direction, and sends downlink pilot signals to the user along the adjusted direction.

[0013] Step 6: The user receives the downlink pilot signals transmitted by the base station, estimates the downlink path gain, and reconstructs the downlink multi-cluster multi-path near-field channel.

[0014] Compared with the prior art, the present invention adopting the above technical solutions has the following technical effects:

[0015] 1. By modeling the multi-cluster multi-path channel, the present invention can better simulate the dense scattering characteristics of the wireless communication environment in the U6G urban scenario, and uses a neural network-based method to accurately estimate this complex environment.

[0016] 2. By utilizing the spatial reciprocity of the uplink and downlink channels in the FDD system to estimate the frequency-independent parameters shared by the uplink and downlink on the uplink link, the present invention can reduce the pilot and feedback overhead.

[0017] 3. By using a high-precision target detection algorithm, the present invention can estimate the positions of the user and all scatterers at one time, achieve a detection accuracy beyond that of the orthogonal matching scheme under the same-size codebook, and effectively reduce the computational complexity.

[0018] 4. By designing a downlink beamforming matrix based on basis vectors, only L downlink pilots need to be transmitted, which can reduce the downlink pilot overhead while achieving high reconstruction accuracy.

[0019] 5. The method of the present invention has good robustness in different signal-to-noise ratio environments, effectively ensuring the stability and efficiency of communication between the base station and the user. Description of the Drawings

[0020] Figure 1 is a schematic diagram of the U6G ultra-large-scale MIMO system model proposed by the present invention;

[0021] Figure 2 is a schematic diagram of the coordinate detection scheme based on the target detection network proposed by the present invention;

[0022] Figure 3It is the flowchart of the downlink channel reconstruction and user positioning method of the present invention for U6G FDD ultra-large-scale MIMO. Detailed implementation manners

[0023] The following details the implementation manners of the present invention, and the examples of the implementation manners are shown in the accompanying drawings. The implementation manners described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as a limitation to the present invention.

[0024] The present invention proposes a low-overhead downlink channel reconstruction and user positioning method for a U6G FDD ultra-large-scale MIMO system. It designs a channel image generation method based on a Cartesian domain codebook to convert the uplink received signal into a sparse channel image, designs a target detection network based on CenterNet to roughly estimate the positions of users and scatterers, and designs a coordinate fine-tuning scheme based on the orthogonal matching pursuit algorithm. The least squares algorithm is used to obtain the uplink path gain, and user positioning is performed according to the maximum uplink path gain. A beamforming matrix based on basis vectors is designed to greatly reduce the downlink pilot overhead in a multi-path multi-cluster scenario. This matrix is used to adjust the downlink pilot beam direction and estimate the downlink path gain, thereby achieving high-precision and fast reconstruction of the U6G band multi-cluster multi-path downlink channel in an FDD mode with low overhead and low complexity. In addition, this method can exhibit good performance in different signal-to-noise ratio scenarios, ensuring the stability and efficiency of communication between the base station and the user equipment. The following further illustrates the present invention in conjunction with embodiments.

[0025] In this embodiment, the ultra-large-scale MIMO system has one base station and one user. There are L clusters between the base station and the user, and there are S + 1 scatterers in each cluster, as Figure 1 shown. The base station is configured with N uniformly spaced linear array antennas, and the user is equipped with a single antenna. The center of each cluster is the main path, and the other paths within the cluster are side paths.

[0026] The uplink multi-cluster multi-path near-field channel can be expressed as:

[0027]

[0028] In the downlink, compared with the uplink, the positions of the user and the scatterers remain unchanged, but the path gains are different. Since the positions of the scatterers within a cluster are close, the downlink gains of the scatterers within a cluster change similarly. The downlink channel can be expressed as:

[0029]

[0030] where g l represents the uplink complex gain of the main path in the l-th cluster, g s,l is the uplink complex gain of the s-th side path in the l-th cluster, and η lis the downlink complex gain coefficient of the l-th cluster. (z l , x l ) are the coordinates of the user or the main path scatterer of the l-th cluster, and (z s,l , x s,l ) are the coordinates of the s-th side path scatterer in the l-th cluster. a(z, x) represents the steering vector, and the n-th element in the steering vector can be modeled as:

[0031]

[0032] where k c represents the wave number, and D n (z, x) represents the distance between the coordinates (z, x) and the n-th antenna, which can be expressed as:

[0033]

[0034] Set the uplink transmitted pilot signal to 1 and the transmit power to P UE . If the additive Gaussian complex noise n follows a complex Gaussian distribution, the uplink signal received by the base station can be expressed as:

[0035]

[0036] Based on the above background description, as Figure 3 shown, a low-overhead downlink channel reconstruction and user positioning method for a U6G FDD ultra-large-scale MIMO system disclosed in an embodiment of the present invention designs a channel image generation method based on a Cartesian codebook, draws the uplink received signal into a channel image, designs an object detection network based on CenterNet to roughly estimate the positions of users and scatterers, designs a coordinate fine-tuning scheme based on the orthogonal matching pursuit algorithm, uses the least squares algorithm to obtain the uplink path gain, performs user positioning according to the maximum uplink path gain, designs a beamforming matrix based on basis vectors to reduce the downlink pilot overhead, uses this matrix to adjust the downlink pilot beam direction, and estimates the downlink path gain, thereby realizing high-precision and fast reconstruction of the multi-cluster multi-path downlink channel in the U6G frequency band in the FDD mode with low cost and low complexity. In addition, this method can exhibit good performance in different signal-to-noise ratio scenarios, ensuring the stability and efficiency of communication between the base station and the user equipment. Specifically, the user sends an uplink pilot signal to the base station, and the received signal at the base station can be written as y ul , and the converted channel image matrix can be written as Use the target detection network CenterNet to detect the positions of users and scatterers from the channel images, then use the orthogonal matching pursuit algorithm to refine the detected coordinates, and use the least squares algorithm to obtain the complex gains of the uplink paths. Then, user positioning is performed based on the maximum uplink path gain. In the downlink, a basis vector-based beamforming matrix is designed to reduce the pilot overhead, the downlink pilot beam directions are adjusted, and the users estimate the downlink path gain coefficients based on the downlink received signals. Finally, the multi-cluster multipath downlink channel reconstruction in the FDD mode is achieved by combining the estimated coordinates and the estimated path gains.

[0037] The method specifically includes the following steps:

[0038] Step 1, uplink channel image generation stage: The user sends a pilot signal to the base station, and the base station visualizes the received uplink signal as a Cartesian domain channel image. When generating the channel image, the near-field Cartesian domain codebook used can be expressed as U C , and each element of it is the steering vector at the sampling point position.

[0039] Perform the following conversion on the uplink received signal using the Cartesian domain codebook:

[0040] y C = U C y ul

[0041] y C y is reshaped into matrix Y C , normalize the magnitude of each term of the matrix, and convert the matrix into a grayscale image matrix in the following way

[0042]

[0043] Use the grayscale value of the image to represent the received signal strength. Each "X"-shaped beam in the image represents a propagation path, and the intersection point is the position of the user or scatterer.

[0044] Step 2, rough coordinate estimation stage: Use such as Figure 2The target detection network CenterNet shown above first trains the model on a training set with annotated target boxes. During testing, the input is the channel image converted in Step 1. The network first extracts general image features through the backbone network ResNet-50, then passes through a decoder composed of an upsampling layer, a normalization layer, and a ReLU activation layer, and finally passes through three detection heads composed of a convolutional layer, a ReLU activation layer, and a convolutional layer, which respectively output a heatmap, a coordinate prediction offset feature map, and a target size feature map. The heatmap is responsible for predicting the category and confidence of the detected object, the coordinate prediction offset feature map is used to predict the horizontal and vertical coordinate offsets of the center point, and the target size feature map is used to predict the width and height of the detection box. In addition, the final predicted coordinates can be obtained through some simple post-processing steps, that is, the rough estimates of the positions of the user and all scatterers can be obtained through one network inference.

[0045] Then, the image coordinates detected by the network are converted into the rough estimate position coordinates in the Cartesian coordinate system

[0046] Step 3, Coordinate Fine-Tuning Stage: Based on the rough estimate position coordinates obtained in Step 2 design the orthogonal matching pursuit optimization of a small-range and fine-grained codebook near each path to obtain the fine-tuned position coordinates

[0047]

[0048] Step 4, User Localization Stage: At the base station side, based on the fine-tuned position coordinates obtained in Step 3 and the received signal y ul , use the least squares algorithm to estimate the uplink path gain Among them, the largest uplink path gain among all uplink path gains corresponds to the uplink path gain of the user, that is, the coordinate localization of the user is performed according to the estimated uplink path gain of the user.

[0049] Step 5, Downlink Beamforming Stage: At the base station side, based on the results of the uplink coordinate and path gain estimation, design a beamforming matrix based on basis vectors. The base station transmits pilot signals along the direction of each cluster to adjust the direction of the pilot beam through the beamforming matrix to achieve low overhead. For each cluster estimated in the uplink, use b l to represent the basis vector of the l-th cluster, which can be expressed as:

[0050]

[0051] Then stack b l | l＝1,…,L into a matrix B. F is the downlink beamforming matrix in the conjugate transpose form of matrix B, where the l-th row of F is the beamforming vector of the l-th cluster. The downlink received signal ydl It can be expressed as:

[0052]

[0053] where P BS represents the transmission power of the base station, and z is the downlink complex Gaussian noise with zero mean and unit variance.

[0054] Step 6, the downlink channel reconstruction phase: The user equipment receives the downlink pilot signal y dl transmitted by the base station, and uses the least squares algorithm to estimate the downlink gain coefficient η:

[0055]

[0056] Finally, the downlink channel is reconstructed as:

[0057]

[0058] Based on the same inventive concept, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the foregoing downlink channel reconstruction and user positioning method for U6G FDD ultra-large-scale MIMO are implemented.

[0059] Based on the same inventive concept, an embodiment of the present application provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the steps of the foregoing downlink channel reconstruction and user positioning method for U6G FDD ultra-large-scale MIMO are implemented.

[0060] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0061] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices produce means for implementing the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0062] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means for implementing the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0063] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0064] The above embodiments are only used to illustrate the technical idea of the present invention, and the protection scope of the present invention cannot be limited thereby. Any modification made on the basis of the technical solution according to the technical idea proposed by the present invention falls within the protection scope of the present invention.

Claims

1. A downlink channel reconstruction and user positioning method for U6G FDD ultra-large-scale MIMO. The ultra-large-scale MIMO system includes a base station and a user. The base station is configured with N uniform linear array antennas, and the user is configured with a single antenna. There are L clusters between the base station and the user, and each cluster has S + 1 scatterers. The central scatterer of each cluster is the main path scatterer, and the other scatterers are the side path scatterers. It is characterized in that, The method includes the following steps: Step 1, the user sends an uplink pilot signal to the base station, and the base station visualizes the received uplink pilot signal as a Cartesian domain channel image; Step 2, use the pre-trained object detection network CenterNet to roughly estimate the positions of the user and the scatterers in the Cartesian domain channel image, obtain the position coordinates detected by the network, and convert the position coordinates detected by the network into the roughly estimated position coordinates in the rectangular coordinate system; Step 3, use orthogonal matching pursuit optimization to optimize the roughly estimated position coordinates obtained in Step 2 to obtain the optimized position coordinates; Step 4, based on the optimized position coordinates obtained in Step 3 and the uplink pilot signal received by the base station, use the least squares algorithm to obtain all uplink path gains; take the largest uplink path gain among all uplink path gains as the uplink path gain of the user, and perform user coordinate positioning according to the uplink path gain of the user; Step 5, the base station designs a downlink beamforming matrix based on basis vectors based on the optimized position coordinates obtained in Step 3 and all uplink path gains obtained in Step 4, adjusts the downlink pilot beam direction, and sends a downlink pilot signal to the user along the adjusted direction; Step 6, the user receives the downlink pilot signal transmitted by the base station, estimates the downlink path gain and reconstructs the downlink multi-cluster multi-path near-field channel.

2. The downlink channel reconstruction and user positioning method for U6G FDD ultra-large-scale MIMO according to claim 1, characterized in that The specific process of Step 1 is as follows: Use the near-field Cartesian domain codebook to perform the following conversion on the uplink pilot signal: y C = U C y ul Among them, y C represents the Cartesian domain received signal, and U C represents the near-field Cartesian domain codebook; y ul represents the uplink pilot signal received by the base station, and the formula is as follows: where P UE is the transmit power of the user, n is the additive white Gaussian complex noise; h ul represents the uplink multi-cluster multipath near-field channel, and the formula is as follows: where g l represents the uplink complex gain of the main path in the l-th cluster, and g s,l is the uplink complex gain of the s-th side path in the l-th cluster. S is the number of side-path scatterers, and a(z l , x l ) and a(z s,l , x s,l ) are both steering vectors. The n-th element [a(z, x)] n is modeled as: where \(j\) is the imaginary unit and \(k\) c is the wave number, \(a(z,x)=a(z\) l ,x l ) or \(a(z\) s,l ,x s,l ), \(D\) n (z,x) represents the distance between the coordinate \((z,x)\) and the \(n\)-th antenna, \(d\) is the antenna spacing, \(n = 1,\ldots,N\); Reshape y C into matrix Y C , normalize matrix Y C , and convert it into a Cartesian domain channel image in the following way: Among them, represents the Cartesian domain channel image, ∥Y C ∥ represents the matrix Y C matrix after taking the modulus of each term.

3. The downlink channel reconstruction and user positioning method for U6G FDD ultra-large-scale MIMO according to claim 1, characterized in that, In Step 2, the pre-trained object detection network CenterNet includes a backbone network ResNet-50, a decoder, and three detection heads with the same structure. The decoder includes an upsampling layer, a normalization layer, and a first ReLU activation layer connected in sequence. The detection head includes a first convolutional layer, a second ReLU activation layer, and a second convolutional layer connected in sequence; The Cartesian domain channel image is subjected to feature extraction using the backbone network ResNet-50. After the decoder decodes the extracted features, they are sent into three detection heads. The three detection heads respectively output a heat map, a coordinate prediction offset feature map, and an object size feature map. The heat map is used to predict the category and confidence of the detected object. The coordinate prediction offset feature map is used to predict the horizontal and vertical coordinate offsets of the center point of the detected object. The object size feature map is used to predict the width and height of the detection box corresponding to the detected object. Through a score threshold filter, the coordinate points with a predicted probability less than the preset threshold are deleted to obtain the position coordinates finally detected by the network. Then is converted into the rough estimated position coordinates in the Cartesian coordinate system 4. The downlink channel reconstruction and user positioning method for U6G FDD ultra-large-scale MIMO according to claim 1, characterized in that In the step 5, the optimized position coordinates are clustered into L clusters, and the base vector b of the l-th cluster l is expressed as: Among them, represents the uplink complex gain of the main path in the estimated $l$-th cluster, represents the uplink complex gain of the $s$-th side path in the estimated $l$-th cluster, represents the steering vector corresponding to the main path coordinates in the estimated $l$-th cluster, represents the steering vector corresponding to the $s$-th side path coordinates in the estimated $l$-th cluster; Stack the b of all clusters l into matrix B. The downlink beamforming matrix F is the conjugate transpose form of matrix B. The l-th row of F is the beamforming vector of the l-th cluster. The downlink pilot signal y dl is expressed as: where P BS is the transmission power of the base station, z is the downlink complex Gaussian noise with zero mean and unit variance; h dl represents the downlink multi-cluster multipath near-field channel, and the formula is as follows: Among them, η l is the downlink complex gain coefficient of the l-th cluster, and g l represents the uplink complex gain of the main path in the l-th cluster, and g s,l is the uplink complex gain of the s-th side path in the l-th cluster. S is the number of side-path scatterers, and a(z l , x l ) and a(z s,l , x s,l ) are both steering vectors.

5. The downlink channel reconstruction and user positioning method for U6G FDD ultra-large-scale MIMO according to claim 1, characterized in that In Step 6, the user receives the downlink pilot signal transmitted by the base station and uses the least squares method to estimate the downlink path gain coefficient η: where F represents the downlink beamforming matrix, and B represents the matrix stacked by all clusters of b l y dl is the downlink pilot signal; Reconstruct the downlink multi-cluster multi-path near-field channel according to η: Among them, represents the reconstructed downlink multi-cluster multipath near-field channel.

6. A computer device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the downlink channel reconstruction and user positioning method for U6G FDD ultra-large-scale MIMO according to any one of claims 1 to 5.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the downlink channel reconstruction and user positioning method for U6G FDD ultra-large-scale MIMO according to any one of claims 1 to 5.

Images