Wireless channel assessment method and system

Abstract

The application discloses a wireless channel evaluation method and system, which is used to solve the technical problem of low efficiency of wireless channel evaluation. A wireless channel evaluation scheme based on beam space reference signal received power measurement data adopts a dual driving mode of model driving and data driving, and establishes a geo-statistical channel model. By establishing a statistical relationship between low-dimensional reference signal received power and high-dimensional channel vector, a sparse signal processing technology is used in a single grid to efficiently solve the angle power spectrum statistical parameters of each transmission path in a three-dimensional space. Since only reference signal received power is used instead of channel matrix, the required calculation complexity is very low. The wireless channel evaluation method provided by the application can quickly and accurately model the statistical characteristics of the wireless channel in a geo-communication scenario, thereby quickly evaluating the wireless channel quality and improving the network optimization efficiency.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a wireless channel evaluation method and system. Background Technology

[0002] A wireless channel is a mathematical description of the communication environment between a signal transmitter and a signal receiver. The characteristics of the channel determine the effectiveness and accuracy of information transmission, playing a decisive role in communication quality. Channel modeling is a crucial step in 5G network optimization. An accurate channel model empowers network optimization algorithms, enabling them to accurately perceive the channel quality of each user in the network topology and provide optimization strategies to ultimately improve overall communication quality.

[0003] In the process of developing the existing technology, the inventors discovered that:

[0004] Existing channel models include deterministic channel models. While deterministic channel models (such as the WINNER channel model and the IMT-advanced channel model) can accurately characterize the digital features of the channel, they often have numerous parameters, complex structures, and require a large amount of computational and storage resources.

[0005] For localized communication scenarios, deterministic channel models require modeling the channel matrix, which involves high computational complexity, low efficiency in wireless channel evaluation, and difficulty in meeting the response requirements of network optimization tasks.

[0006] Therefore, a new wireless channel evaluation scheme is needed to solve the technical problem of low efficiency in wireless channel evaluation. Summary of the Invention

[0007] This application provides a new wireless channel evaluation scheme to solve the technical problem of low efficiency in wireless channel evaluation.

[0008] Specifically, a wireless channel evaluation method includes the following steps:

[0009] The antenna array transmits a reference signal according to a preset antenna gain and transmit power;

[0010] The received power of the reference signal is obtained by measuring at a single grid position at a preset location;

[0011] The environmental multipath structure of the wireless channel is obtained by using a localized statistical channel model.

[0012] Based on the environmental multipath structure, the angular power spectrum statistical characteristics of the wireless channel are obtained.

[0013] The quality of a wireless channel is assessed based on the statistical characteristics of its angular power spectrum.

[0014] Furthermore, using a localized statistical channel model, the environmental multipath structure of the wireless channel is obtained, specifically including:

[0015] The environmental multipath structure of a wireless channel is described by the channel impulse response of a single grid using an antenna array.

[0016] The antenna's channel impulse response to a single grid is as follows:

[0017]

[0018] In the formula, N V N represents the total number of divisions of the vertical plane in the angular domain. H This represents the total number of divisions of the horizontal plane in the angular domain. G represents path loss. i，d Indicates antenna gain. d represents the angle of departure of the channel relative to the vertical line of the ground. x and d y Indicates the spacing between adjacent antennas. This indicates that the phase error between angles follows a uniform distribution in the interval [-π, π]. This indicates that the phase error between antennas follows a mean of 0 and a variance of σ. 2 The Gaussian distribution.

[0019] Furthermore, based on the environmental multipath structure, the angular power spectrum statistical characteristics of the wireless channel are obtained, specifically including:

[0020] Based on the environmental multipath structure, the relationship between the received power of the reference signal and the statistical characteristics of the angular power spectrum of the wireless channel is obtained:

[0021]

[0022] In the formula, N V N represents the total number of divisions of the vertical plane in the angular domain. H RSRP represents the total number of divisions of the horizontal plane in the angular domain. l,k Indicates the received power of the reference signal. Represents the coefficient matrix. This represents the angular power spectrum statistical characteristics of a wireless channel;

[0023] and

[0024]

[0025] In the formula, P represents the transmission power, and g i，j This indicates the antenna gain.

[0026] Furthermore, the method also includes:

[0027] We use the weighted nonnegative orthogonal matching pursuit method to estimate sparse vectors as statistical properties of the angular power spectrum.

[0028] Furthermore, a weighted nonnegative orthogonal matching pursuit method is used to estimate sparse vectors as statistical properties of the angular power spectrum, specifically including:

[0029] definition

[0030] enter

[0031] definition x = 0, r0 = y are the initial parameters;

[0032] repeat

[0033] Until max(A) T r k ) < 0 or |S| squals to K,

[0034] Calculate the sparse vector x as a statistical property of the angular power spectrum.

[0035] Furthermore, the method also includes:

[0036] The sparse Bayesian learning method is used to estimate sparse vectors as statistical properties of the angular power spectrum.

[0037] Furthermore, a sparse Bayesian learning method is used to estimate sparse vectors as statistical properties of the angular power spectrum, specifically including:

[0038] Using the sparse Bayesian learning method, the probabilistic model is defined as follows:

[0039] RSRP = Ax + n,

[0040] In the formula, n is a Gaussian distribution. noise;

[0041] Based on the likelihood function, we obtain

[0042]

[0043] In the formula, the inverse variance parameter β follows a Gamma distribution:

[0044]

[0045] In the formula, It is the Gamma function;

[0046] According to Bayes' theorem, the posterior probability distribution of x is:

[0047]

[0048] Assume the prior probability of x follows a Gaussian distribution:

[0049]

[0050] The posterior probability of x follows a Gaussian distribution.

[0051] In the formula,

[0052] The maximum posterior probability can then be expressed as:

[0053]

[0054] The log-likelihood function is obtained as follows:

[0055]

[0056] In the formula, C = β -1 I M +A[diag(α)] -1 A T ;

[0057] Based on the log-likelihood function, taking the partial derivatives with respect to hyperparameters α and β respectively, and setting the partial derivatives of hyperparameters α and β to zero, we obtain:

[0058]

[0059]

[0060] In the formula,

[0061] This application also provides a wireless channel evaluation system.

[0062] Specifically, a wireless channel assessment system includes:

[0063] Antenna array, used to transmit reference signals according to preset antenna gain and transmit power;

[0064] A wireless channel evaluation device is used to measure the received power of a reference signal at a single grid position at a preset location; it is also used to obtain the environmental multipath structure of the wireless channel using a localized statistical channel model; it is also used to obtain the angular power spectrum statistical characteristics of the wireless channel based on the environmental multipath structure; and it is also used to perform quality evaluation of the wireless channel based on the angular power spectrum statistical characteristics of the wireless channel.

[0065] Furthermore, the wireless channel evaluation device uses a localized statistical channel model to obtain the environmental multipath structure of the wireless channel, specifically including:

[0066] The environmental multipath structure of a wireless channel is described by the channel impulse response of a single grid through an antenna array. The channel impulse response of the antenna to a single grid is as follows:

[0067]

[0068] In the formula, N V N represents the total number of divisions of the vertical plane in the angular domain. H This represents the total number of divisions of the horizontal plane in the angular domain. G represents path loss. i，j Indicates antenna gain. d represents the angle of departure of the channel relative to the vertical line of the ground. x and d y Indicates the spacing between adjacent antennas. This indicates that the phase error between angles follows a uniform distribution in the interval [-π, π]. This indicates that the phase error between antennas follows a mean of 0 and a variance of σ. 2 The Gaussian distribution.

[0069] Furthermore, the wireless channel evaluation device obtains the angular power spectrum statistical characteristics of the wireless channel based on the environmental multipath structure, specifically including:

[0070] Based on the environmental multipath structure, the relationship between the received power of the reference signal and the statistical characteristics of the angular power spectrum of the wireless channel is obtained:

[0071]

[0072] In the formula, N V N represents the total number of divisions of the vertical plane in the angular domain. H RSRP represents the total number of divisions of the horizontal plane in the angular domain. l,k Indicates the received power of the reference signal. Represents the coefficient matrix. This represents the angular power spectrum statistical characteristics of a wireless channel;

[0073] and

[0074] In the formula, P represents the transmission power, and g i，j This indicates the antenna gain.

[0075] The technical solution provided in this application has at least the following beneficial effects:

[0076] Based on reference signal received power measurement data in beamspace, a localized statistical channel model was established using a dual-drive approach of model-driven and data-driven methods. By establishing the statistical relationship between low-dimensional reference signal received power and high-dimensional channel vectors, sparse signal processing techniques were employed in a single grid to efficiently solve for the angular power spectrum statistical parameters of each transmission path in three-dimensional space. Since only the reference signal received power was used instead of the channel matrix, the required computational complexity was very low. The wireless channel evaluation method provided in this application can perform fast and accurate modeling of wireless channel statistical characteristics in localized communication scenarios, thereby enabling rapid evaluation of wireless channel quality and improving network optimization efficiency. Attached Figure Description

[0077] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0078] Figure 1 A flowchart illustrating a wireless channel evaluation method provided in an embodiment of this application;

[0079] Figure 2 A schematic diagram of the antenna array reference coordinate system and multipath channel provided in the embodiments of this application;

[0080] Figure 3 A schematic diagram of the reference signal received power in the beam domain provided in an embodiment of this application;

[0081] Figure 4 A schematic diagram of the main lobe and side lobes of coefficient matrix A provided in an embodiment of this application;

[0082] Figure 5 This is a schematic diagram of the structure of a wireless channel evaluation system provided in an embodiment of this application.

[0083] The reference numerals in the figure are as follows:

[0084] 100 Wireless Channel Evaluation System

[0085] 11-antenna array

[0086] 12. Wireless channel evaluation device. Detailed Implementation

[0087] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0088] Please refer to Figure 1 To address the technical problem of low efficiency in wireless channel assessment, this application provides a wireless channel assessment method, comprising the following steps:

[0089] S110: The antenna array transmits a reference signal according to the preset antenna gain and transmit power.

[0090] It is understood that the antenna array is installed on a base station for propagating wireless signals. Typically, the antenna array consists of several identical antenna elements arranged according to a certain pattern. The antenna array has preset engineering parameters such as antenna gain, antenna spacing, and transmit power.

[0091] To ensure communication quality, base stations typically use antenna arrays to transmit reference signals to determine which specific frequency ranges within the base station's coverage area have better quality, and thus prioritize their allocation to user terminals.

[0092] Specifically, the antenna array transmits reference signals based on preset antenna gain and transmit power to provide a reference for the base station's resource scheduling.

[0093] S120: Measure the received power of the reference signal at a single grid position at a preset location.

[0094] It should be noted that, in order to improve the accuracy of channel modeling, this application further subdivides the coverage area of ​​the base station. Specifically, this application divides the coverage area of ​​the base station into a grid. In the application scenario where the base station is a 5G base station, since the coverage radius of a 5G base station is approximately 100-300 meters, the grid division of the base station's coverage area in this application can be represented as: subdividing the base station's coverage area into several 10-meter by 10-meter square grids.

[0095] Based on the gridded division of the base station coverage area, an antenna array reference coordinate system can be established. Then, based on this reference coordinate system, the coordinates of any single grid can be determined as its positional attribute. Furthermore, the positional attribute of any single grid can be defined using preset position coordinates. The received power of a reference signal can be measured at the location of a single grid to characterize the wireless channel.

[0096] S130: Using a localized statistical channel model, the environmental multipath structure of the wireless channel is obtained.

[0097] It is understandable that the environmental multipath structure of a wireless channel can be characterized by large-scale fading (path loss, shadowing fading) or small-scale fading. This application characterizes the large-scale fading of a channel with arbitrary grids using a localized statistical channel model.

[0098] For further details, please refer to Figure 2 Assume the base station's antenna array contains N x ×N y If there are multiple antennas, then using a localized statistical channel model, the environmental multipath structure of the wireless channel can be obtained, specifically including:

[0099] The environmental multipath structure of a wireless channel is described by the channel impulse response of a single grid through an antenna array, where the channel impulse response of antenna (x, y) to grid l is:

[0100]

[0101] In the formula, N V N represents the total number of divisions of the vertical plane in the angular domain. H This represents the total number of divisions of the horizontal plane in the angular domain;

[0102] This represents path loss, and the path loss follows a log-normal distribution, that is...

[0103]

[0104] g i，j This represents the antenna gain for the corresponding antenna (x, y);

[0105] Indicates the angle of departure of the channel relative to the vertical line of the ground;

[0106] d x and d y Indicates the spacing between adjacent antennas;

[0107] This indicates that the phase error between angles follows a uniform distribution in the interval [-π, π]. This indicates that the phase error between antennas follows a mean of 0 and a variance of σ. 2 The Gaussian distribution.

[0108] It should be noted that the unknown parameter in the above formula (1) is the path loss. These are the non-zero elements in the angular power spectrum x that need to be estimated.

[0109] S140: Obtain the angular power spectrum statistical characteristics of the wireless channel based on the environmental multipath structure.

[0110] Understandably, the angular power spectrum x is a statistical characteristic of channel fading on a large scale and can be used to characterize the environment. Specifically, the angular power spectrum x is a sparse vector whose dimension is the number of equal divisions of the angle. The positions of the non-zero elements correspond to the departure angles of the channel multipath, and the values ​​of the non-zero elements correspond to the channel gain of each path.

[0111] Furthermore, based on the environmental multipath structure, the angular power spectrum statistical characteristics of the wireless channel are obtained, specifically including:

[0112] Based on the environmental multipath structure, the relationship between the received power of the reference signal and the statistical characteristics of the angular power spectrum of the wireless channel is obtained:

[0113]

[0114] In the formula, N V N represents the total number of divisions of the vertical plane in the angular domain. H RSRP represents the total number of divisions of the horizontal plane in the angular domain. l,k Indicates the received power of the reference signal. Represents the coefficient matrix. This represents the angular power spectrum statistical characteristics of a wireless channel;

[0115] and

[0116]

[0117] In the formula, P represents the transmission power, and g i，j This indicates the antenna gain.

[0118] The following describes the derivation of the relationship between the received power of the reference signal and the angular power spectrum statistical characteristics of the wireless channel:

[0119] In massive MIMO antennas, the channel impulse response has a large dimension. To reduce complexity, this application considers the reference signal received power in the beamspace. Assume the channel matrix from the antenna array to the l-th grid is... The precoding matrix of the k-th beam is So if Figure 3 As shown, the reference signal received power from the k-th beam to the l-th grid can be expressed as:

[0120]

[0121] Where P represents the transmit power. Further expansion of rsrp l，k (t) can be obtained

[0122]

[0123] By taking the expectation of time i on both sides of the above equation, we can obtain

[0124]

[0125] in,

[0126]

[0127] Thus, this application establishes the statistical relationship between the received power of the low-dimensional reference signal and the angular power spectrum of the wireless channel. For any single grid, this relationship can be expressed as:

[0128]

[0129] The main lobe and side lobes of coefficient matrix A are as follows: Figure 4 As shown, the path between the antenna array and the grid can only exist at the angles corresponding to the main lobe and the side lobe. x is a sparse vector with generally less than 10 non-zero elements, corresponding to a multipath channel with generally less than 10 paths.

[0130] Thus, this application can estimate x using RSRP and coefficient matrix A.

[0131] Furthermore, based on the reference signal received power (RSRP) measurement data in the beam space, this application provides a model-driven or data-driven approach to establish a localized statistical channel model.

[0132] In the embodiment of this application that employs model-driven localized statistical channel modeling, the method further includes:

[0133] We use the weighted nonnegative orthogonal matching pursuit method to estimate sparse vectors as statistical properties of the angular power spectrum.

[0134] Specifically, in the embodiment of this application that uses a model-driven approach to establish a localized statistical channel model, this application proposes an optimization problem to solve for the l0 norm constraint in order to recover x:

[0135]

[0136] Where K represents the maximum number of non-zero elements in x.

[0137] Since the coefficient matrix A contains some columns with large amplitudes, it affects the accuracy of finding non-zero elements. To solve this problem (6), this application adopts the weighted non-negative orthogonal matching pursuit (WNOMP) method, and designs a dynamic weight λ using the WNOMP method. k This reduces the impact of columns with large amplitude.

[0138] Furthermore, a weighted nonnegative orthogonal matching pursuit method is used to estimate sparse vectors as statistical properties of the angular power spectrum, specifically including:

[0139] definition

[0140] enter

[0141] Define k = 0 x = 0, r0 = y are the initial parameters;

[0142] repeat

[0143] until

[0144] Calculate the sparse vector x as a statistical property of the angular power spectrum.

[0145] Furthermore, in addition to the model-driven approach using the weighted non-negative orthogonal matching pursuit method, this application can also employ a data-driven approach to estimate x. In the embodiment of this application that uses a data-driven approach to establish a localized statistical channel model, the method further includes:

[0146] The sparse Bayesian learning method is used to estimate sparse vectors as statistical properties of the angular power spectrum.

[0147] It is understandable that the sparse Bayesian learning is a data-driven approach that automatically learns key parameters and can effectively replace manual parameter tuning.

[0148] Furthermore, a sparse Bayesian learning method is used to estimate sparse vectors as statistical properties of the angular power spectrum, specifically including:

[0149] Using the sparse Bayesian learning method, the probabilistic model is defined as follows:

[0150] RSRP = Ax + n,

[0151] In the formula, n is a Gaussian distribution. noise;

[0152] Based on the likelihood function, we obtain

[0153]

[0154] In the formula, the inverse variance parameter β follows a Gamma distribution:

[0155]

[0156] In the formula, It is the Gamma function;

[0157] According to Bayes' theorem, the posterior probability distribution of x is:

[0158]

[0159] Assume the prior probability of x follows a Gaussian distribution:

[0160]

[0161] The posterior probability of x follows a Gaussian distribution.

[0162] In the formula,

[0163] The maximum posterior probability can then be expressed as:

[0164]

[0165] The log-likelihood function is obtained as follows:

[0166]

[0167] In the formula, C = β -1 I M +A[diag(α)] -1 A T ;

[0168] Based on the log-likelihood function, taking the partial derivatives with respect to hyperparameters α and β respectively, and setting the partial derivatives of hyperparameters α and β to zero, we obtain:

[0169]

[0170]

[0171] In the formula,

[0172] The derivation process is explained below:

[0173] In order to use the sparse Bayesian learning method, this application introduces a probabilistic model:

[0174] RSRP = Ax + n, (7)

[0175] Where n follows a Gaussian distribution The noise.

[0176] Furthermore, the likelihood function can be obtained as follows:

[0177]

[0178] Wherein, the inverse variance parameter β follows a Gamma distribution:

[0179]

[0180] in It is the Gamma function.

[0181] According to Bayes' theorem, the posterior probability distribution of x can be obtained as follows:

[0182]

[0183] Next, we need to determine the prior probability p(x) of x, and then calculate the posterior distribution of x according to (9).

[0184] Suppose that the prior probability of x follows a Gaussian distribution, that is...

[0185]

[0186] Therefore, the posterior probability of x also follows a Gaussian distribution.

[0187] in

[0188] The problem of maximizing the posterior probability can be expressed as:

[0189]

[0190] At the same time, the log-likelihood function can be obtained as follows:

[0191]

[0192] Where C = β -1 I M +A[diag(α)] -1 A T .

[0193] To determine the hyperparameters α and β, we take the partial derivatives of the log-likelihood function in (13) with respect to these two hyperparameters, and set their partial derivatives to zero, thus obtaining...

[0194]

[0195] in,

[0196] The sparse Bayesian learning method begins by randomly assigning initial values ​​to the hyperparameters α and β, then updating the mean and covariance matrix of the posterior distribution of x according to (11), followed by updating the hyperparameters α and β according to (14). (11) and (14) are iterated alternately in this manner until the algorithm converges and stabilizes. During this learning process, a portion of {α} i The value of} will become very large, this part {α} i This corresponds to a Gaussian distribution with zero mean and zero variance, which is the zero element corresponding to x. The other part {α} i The value of} will become very small, ensuring the existence of non-zero elements of x. The algorithm can converge by iterating (11) and (14) several times, but since there is a matrix inversion operation in (11), the overall algorithm running time is not short.

[0197] S150: Based on the angular power spectrum statistical characteristics of the wireless channel, perform a quality assessment of the wireless channel.

[0198] Understandably, based on the angular power spectrum statistical characteristics of the wireless channel, a channel quality assessment can be performed, generating a channel quality assessment result. After obtaining the channel quality assessment result, the base station can select an appropriate scheduling algorithm and downlink data block size to ensure that the user terminal obtains the best downlink performance in different wireless environments.

[0199] In summary, the wireless channel evaluation method provided in this application, based on reference signal received power (RSRP) measurement data in beamspace, employs a dual-drive approach of model-driven and data-driven methods to establish a localized statistical channel model. By establishing the statistical relationship between low-dimensional reference signal received power and high-dimensional channel vectors, sparse signal processing techniques (such as sparse optimization and sparse Bayesian learning) are used in a single grid to efficiently solve for the angular power spectrum statistical parameters of each transmission path in three-dimensional space. Since only the reference signal received power is used instead of the channel matrix, the required computational complexity is very low. The wireless channel evaluation method provided in this application can perform fast and accurate modeling of wireless channel statistical characteristics in localized communication scenarios, thereby enabling rapid evaluation of wireless channel quality and improving network optimization efficiency.

[0200] Please refer to Figure 5 To support wireless channel evaluation methods, this application also provides a wireless channel evaluation system 100, comprising:

[0201] Antenna array 11 is used to transmit reference signals according to preset antenna gain and transmission power;

[0202] The wireless channel evaluation device 12 is used to measure the received power of a reference signal at a single grid position at a preset location; it is also used to obtain the environmental multipath structure of the wireless channel using a localized statistical channel model; it is also used to obtain the angular power spectrum statistical characteristics of the wireless channel based on the environmental multipath structure; and it is also used to perform quality evaluation of the wireless channel based on the angular power spectrum statistical characteristics of the wireless channel.

[0203] It is understood that the antenna array 11 is installed on a base station for transmitting wireless signals. Typically, the antenna array 11 consists of several identical antenna elements arranged according to a certain pattern. The antenna array 11 has preset engineering parameters such as antenna gain, antenna spacing, and transmit power.

[0204] To ensure communication quality, base stations typically use antenna array 11 to transmit reference signals to determine which specific frequency ranges within the base station's coverage area have better quality, and thus prioritize their allocation to user terminals.

[0205] Specifically, the antenna array 11 transmits a reference signal based on the preset antenna gain and transmit power to provide a reference for the scheduling resources of the base station.

[0206] The wireless channel evaluation device 12 measures the received power of the reference signal at a single grid position at a preset location.

[0207] It should be noted that, in order to improve the accuracy of channel modeling, this application further subdivides the coverage area of ​​the base station. Specifically, this application divides the coverage area of ​​the base station into a grid. In the application scenario where the base station is a 5G base station, since the coverage radius of a 5G base station is approximately 100-300 meters, the grid division of the base station's coverage area in this application can be represented as: subdividing the base station's coverage area into several 10-meter by 10-meter square grids.

[0208] Based on the gridded division of the base station's coverage area, the wireless channel evaluation device 12 can establish a reference coordinate system for the antenna array 11. Then, based on the antenna array 11 reference coordinate system, the coordinates of any single grid can be determined as the positional attribute of that single grid. Furthermore, the wireless channel evaluation device 12 can define the positional attribute of any single grid using preset position coordinates. The received power of a reference signal can be measured at the location of a single grid to characterize the wireless channel.

[0209] The wireless channel evaluation device 12 uses a localized statistical channel model to obtain the environmental multipath structure of the wireless channel.

[0210] It is understandable that the environmental multipath structure of a wireless channel can be characterized by large-scale fading (path loss, shadowing fading) or small-scale fading. The wireless channel evaluation device 12 characterizes the large-scale fading of the channel in an arbitrary grid by using a localized statistical channel model.

[0211] For further details, please refer to Figure 2 Assuming antenna array 11 contains N x ×N y If there are one antenna, then the wireless channel evaluation device 12 uses a localized statistical channel model to obtain the environmental multipath structure of the wireless channel, specifically including:

[0212] The environmental multipath structure of the wireless channel is described by the channel impulse response of a single grid using antenna array 11, where the channel impulse response of antenna (x, y) to grid l is:

[0213]

[0214] In the formula, N V N represents the total number of divisions of the vertical plane in the angular domain. H This represents the total number of divisions of the horizontal plane in the angular domain;

[0215] This represents path loss, and the path loss follows a log-normal distribution, that is...

[0216]

[0217] g i，j This represents the antenna gain for the corresponding antenna (x, y);

[0218] Indicates the angle of departure of the channel relative to the vertical line of the ground;

[0219] d x and d y Indicates the spacing between adjacent antennas;

[0220] This indicates that the phase error between angles follows a uniform distribution in the interval [-π, π]. This indicates that the phase error between antennas follows a mean of 0 and a variance of σ. 2 The Gaussian distribution.

[0221] It should be noted that the unknown parameter in the above formula (1) is the path loss. These are the non-zero elements in the angular power spectrum x that need to be estimated.

[0222] The wireless channel evaluation device 12 obtains the angular power spectrum statistical characteristics of the wireless channel based on the environmental multipath structure.

[0223] Understandably, the angular power spectrum x is a statistical characteristic of channel fading on a large scale and can be used to characterize the environment. Specifically, the angular power spectrum x is a sparse vector whose dimension is the number of equal divisions of the angle. The positions of the non-zero elements correspond to the departure angles of the channel multipath, and the values ​​of the non-zero elements correspond to the channel gain of each path.

[0224] Furthermore, the wireless channel evaluation device 12 obtains the angular power spectrum statistical characteristics of the wireless channel based on the environmental multipath structure, specifically including:

[0225] Based on the environmental multipath structure, the relationship between the received power of the reference signal and the statistical characteristics of the angular power spectrum of the wireless channel is obtained:

[0226]

[0227] In the formula, N V N represents the total number of divisions of the vertical plane in the angular domain. H RSRPl represents the total number of divisions of the horizontal plane in the angular domain. ,k Indicates the received power of the reference signal. Represents the coefficient matrix. This represents the angular power spectrum statistical characteristics of a wireless channel;

[0228] and

[0229]

[0230] In the formula, P represents the transmission power, and g i，j This indicates the antenna gain.

[0231] The following describes the derivation of the relationship between the received power of the reference signal and the angular power spectrum statistical characteristics of the wireless channel:

[0232] In massive MIMO antennas, the channel impulse response has a large dimension. To reduce complexity, this application considers the reference signal received power in the beam space. Assume the channel matrix from antenna array 11 to the l-th grid is... The precoding matrix of the k-th beam is So if Figure 3 As shown, the reference signal received power from the k-th beam to the l-th grid can be expressed as:

[0233]

[0234] Where P represents the transmit power. Further expansion of rsrp l，k (t) can be obtained

[0235]

[0236] By taking the expectation of time t on both sides of the above equation, we can obtain

[0237]

[0238] in,

[0239]

[0240] Thus, this application establishes the statistical relationship between the received power of the low-dimensional reference signal and the angular power spectrum of the wireless channel. For any single grid, this relationship can be expressed as:

[0241]

[0242] The main lobe and side lobes of coefficient matrix A are as follows: Figure 4 As shown, the path between antenna array 11 and the grid can only exist at the angles corresponding to the main lobe and side lobe. x is a sparse vector with generally less than 10 non-zero elements, corresponding to a multipath channel with generally less than 10 paths.

[0243] Thus, this application can estimate x using RSRP and coefficient matrix A.

[0244] Furthermore, based on the reference signal received power (RSRP) measurement data in the beam space, the wireless channel evaluation device 12 establishes a localized statistical channel model using model-driven or data-driven methods.

[0245] In an embodiment where the wireless channel evaluation device 12 employs a model-driven approach to establish a localized statistical channel model, the wireless channel evaluation device 12 is further configured to:

[0246] We use the weighted nonnegative orthogonal matching pursuit method to estimate sparse vectors as statistical properties of the angular power spectrum.

[0247] Specifically, in the embodiment where the wireless channel evaluation device 12 employs a model-driven approach to establish a localized statistical channel model, this application proposes an optimization problem to solve for the l0 norm constraint to recover x:

[0248]

[0249] Where K represents the maximum number of non-zero elements in x.

[0250] Since the coefficient matrix A contains some columns with large amplitudes, it affects the accuracy of finding non-zero elements. To solve this problem (6), the wireless channel evaluation device 12 adopts the weighted non-negative orthogonal matching pursuit (WNOMP) method, and designs a dynamic weight λ using the WNOMP method.k This reduces the impact of columns with large amplitude.

[0251] Furthermore, the wireless channel evaluation device 12 uses a weighted non-negative orthogonal matched pursuit method to estimate sparse vectors as statistical characteristics of the angular power spectrum, specifically including:

[0252] definition

[0253] enter

[0254] Define k = 0, x = 0, r0 = y are the initial parameters;

[0255] repeat

[0256] until

[0257] Calculate the sparse vector x as a statistical property of the angular power spectrum.

[0258] Furthermore, in addition to the model-driven approach using the weighted non-negative orthogonal matching pursuit method, the wireless channel evaluation device 12 can also use a data-driven approach to estimate x. In an embodiment where the wireless channel evaluation device 12 uses a data-driven approach to establish a localized statistical channel model, the wireless channel evaluation device 12 is also used for:

[0259] The sparse Bayesian learning method is used to estimate sparse vectors as statistical properties of the angular power spectrum.

[0260] It is understandable that the sparse Bayesian learning is a data-driven approach that automatically learns key parameters and can effectively replace manual parameter tuning.

[0261] Furthermore, the wireless channel evaluation device 12 uses a sparse Bayesian learning method to estimate sparse vectors as statistical properties of the angular power spectrum, specifically including:

[0262] Using the sparse Bayesian learning method, the probabilistic model is defined as follows:

[0263] RSRP = Ax + n, where n is a Gaussian distribution. noise;

[0264] Based on the likelihood function, we obtain

[0265]

[0266] In the formula, the inverse variance parameter β follows a Gamma distribution:

[0267]

[0268] In the formula, It is the Gamma function;

[0269] According to Bayes' theorem, the posterior probability distribution of x is:

[0270]

[0271] Assume the prior probability of x follows a Gaussian distribution:

[0272]

[0273] The posterior probability of x follows a Gaussian distribution.

[0274] In the formula,

[0275] The maximum posterior probability can then be expressed as:

[0276]

[0277] The log-likelihood function is obtained as follows:

[0278]

[0279] In the formula, C = β -1 I M +A[diag(α)] -1 A T ;

[0280] Based on the log-likelihood function, taking the partial derivatives with respect to hyperparameters α and β respectively, and setting the partial derivatives of hyperparameters α and β to zero, we obtain:

[0281]

[0282]

[0283] In the formula,

[0284] The derivation process is explained below:

[0285] In order to use the sparse Bayesian learning method, this application introduces a probabilistic model:

[0286] RSRP=Ax+n (7)

[0287] Where n follows a Gaussian distribution The noise.

[0288] Furthermore, the likelihood function can be obtained as follows:

[0289]

[0290] Wherein, the inverse variance parameter β follows a Gamma distribution:

[0291]

[0292] in It is the Gamma function.

[0293] According to Bayes' theorem, the posterior probability distribution of x can be obtained as follows:

[0294]

[0295] Next, we need to determine the prior probability p(x) of x, and then calculate the posterior distribution of x according to (9).

[0296] Suppose that the prior probability of x follows a Gaussian distribution, that is...

[0297]

[0298] Therefore, the posterior probability of x also follows a Gaussian distribution.

[0299] in

[0300] The problem of maximizing the posterior probability can be expressed as:

[0301]

[0302] At the same time, the log-likelihood function can be obtained as follows:

[0303]

[0304] Where C = β -1 I M +A[diag(α)] -1 A T .

[0305] To determine the hyperparameters α and β, we take the partial derivatives of the log-likelihood function in (13) with respect to these two hyperparameters, and set their partial derivatives to zero, thus obtaining...

[0306]

[0307] in,

[0308] The sparse Bayesian learning method begins by randomly assigning initial values ​​to the hyperparameters α and β, then updating the mean and covariance matrix of the posterior distribution of x according to (11), followed by updating the hyperparameters α and β according to (14). (11) and (14) are iterated alternately in this manner until the algorithm converges and stabilizes. During this learning process, a portion of {α} iThe value of} will become very large, this part {α} i This corresponds to a Gaussian distribution with zero mean and zero variance, which is the zero element corresponding to x. The other part {α} i The value of} will become very small, ensuring the existence of non-zero elements of x. The algorithm can converge by iterating (11) and (14) several times, but since there is a matrix inversion operation in (11), the overall algorithm running time is not short.

[0309] Finally, the wireless channel evaluation device 12 performs a quality assessment of the wireless channel based on the angular power spectrum statistical characteristics, generating a channel quality assessment result. After obtaining the channel quality assessment result, the base station can select an appropriate scheduling algorithm and downlink data block size to ensure that the user terminal obtains the best downlink performance in different wireless environments.

[0310] In summary, the wireless channel evaluation system 100 provided in this application, based on beamspace reference signal received power (RSRP) measurement data, employs a dual-drive approach of model-driven and data-driven methods to establish a localized statistical channel model. By establishing the statistical relationship between low-dimensional reference signal received power and high-dimensional channel vectors, sparse signal processing techniques (such as sparse optimization and sparse Bayesian learning) are used in a single grid to efficiently solve the angular power spectrum statistical parameters of each transmission path in three-dimensional space. Since only the reference signal received power is used instead of the channel matrix, the required computational complexity is very low. The wireless channel evaluation system 100 provided in this application can perform fast and accurate modeling of wireless channel statistical characteristics in localized communication scenarios, thereby enabling rapid evaluation of wireless channel quality and improving network optimization efficiency.

[0311] It should be noted that 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 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 said element.

[0312] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0313] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A wireless channel evaluation method, characterized in that, Includes the following steps: The antenna array transmits a reference signal according to a preset antenna gain and transmit power; The received power of the reference signal is obtained by measuring at a single grid position at a preset location; The environmental multipath structure of the wireless channel is obtained by using a localized statistical channel model. Based on the environmental multipath structure, the angular power spectrum statistical characteristics of the wireless channel are obtained. The quality of the wireless channel is assessed based on the statistical characteristics of its angular power spectrum. Among them, based on the environmental multipath structure, the angular power spectrum statistical characteristics of the wireless channel are obtained, specifically including: Based on the environmental multipath structure, the relationship between the received power of the reference signal and the statistical characteristics of the angular power spectrum of the wireless channel is obtained: ； In the formula, This represents the total number of divisions of the vertical plane in the angular domain. This represents the total number of divisions of the horizontal plane in the angular domain. Indicates the received power of the reference signal. Represents the coefficient matrix. This represents the angular power spectrum statistical characteristics of a wireless channel; and ; ； In the formula Represents transmission power. Indicates antenna gain. : Indicates the spacing between adjacent antennas, : Indicates that the antenna array is in direction, The number of antenna elements in the direction, : indicates the first The first beam, the first The first angle, the second Phase values ​​of each antenna, : indicates the first The first beam, the first The first angle, the second Phase values ​​of each antenna, : indicates the first The beam at the 1st Precoded phase on each antenna : Indicates the tilt angle of the channel relative to the vertical line of the ground. : Indicates the azimuth angle of the channel relative to the vertical line of the ground; The method further includes: definition ; enter ; definition These are the initial parameters; repeat , until We can use a sparse Bayesian learning method to compute a sparse vector x as a statistical property of the angular power spectrum.

2. The wireless channel evaluation method as described in claim 1, characterized in that, Using a localized statistical channel model, the environmental multipath structure of the wireless channel is obtained, specifically including: The environmental multipath structure of a wireless channel is described by the channel impulse response of a single grid through an antenna array. The channel impulse response of the antenna to a single grid is as follows: ； In the formula, This represents the total number of divisions of the vertical plane in the angular domain. This represents the total number of divisions of the horizontal plane in the angular domain. Indicates path loss. Indicates antenna gain. Indicates the angle of departure of the channel relative to the vertical line of the ground. and Indicates the spacing between adjacent antennas. This indicates that the phase error between angles is within the interval [ The distribution in the middle follows a uniform distribution. This indicates that the phase error between antennas follows a mean of 0 and a variance of . The Gaussian distribution.

3. A wireless channel evaluation system based on the method of any one of claims 1-2, characterized in that, include: Antenna array, used to transmit reference signals according to preset antenna gain and transmit power; A wireless channel evaluation device is used to measure the received power of a reference signal at a single grid position at a preset location; it is also used to obtain the environmental multipath structure of the wireless channel using a localized statistical channel model; it is also used to obtain the angular power spectrum statistical characteristics of the wireless channel based on the environmental multipath structure; and it is also used to perform quality evaluation of the wireless channel based on the angular power spectrum statistical characteristics of the wireless channel.

Images