A geometric random and beam-domain wireless channel modeling method for very large-scale MIMO

Abstract

The present invention discloses a method for modeling ultra-large-scale MIMO geometric random and beam-domain wireless channels, and the specific steps include: 1) setting simulation parameters, network layout, etc., generating initial large-scale parameters, and generating initial clusters based on model initialization parameters; 2) updating large-scale and small-scale parameters; 3) generating geometric random channel model channel coefficients, steering matrices, and beam-domain channel matrices; 4) analyzing channel characteristics, and verifying the accuracy of the model by comparing with measured data. The wireless channel modeling method established by the present invention supports spherical waves and spatial non-stationary characteristics, and can realize the conversion from geometric random to beam-domain channel models under near-field conditions. The model accuracy is verified by the consistency of simulation and measurement results of space-time-frequency correlation functions and channel capacity. The present invention has guiding significance for the design and performance evaluation of ultra-large-scale MIMO communication systems.

Description

Technical Field

[0001] The present invention belongs to the technical field of ultra-large-scale MIMO wireless channel modeling, and in particular relates to an ultra-large-scale MIMO geometric random and beam domain wireless channel modeling method. Background Art

[0002] The sixth generation (6G) mobile communication system has proposed a vision of full coverage, full spectrum, full application and strong security. As one of the key technologies for 6G full application scenarios, ultra-large-scale MIMO has attracted more and more attention. When the base station uses hundreds, thousands or even tens of thousands of antennas, the spectrum efficiency is greatly increased; at the same time, as the number of antennas on the base station increases, the channel vectors between users tend to be orthogonal, so that more users can be served and the connection density can be increased. These advantages make it possible for the 6G mobile communication system to achieve the goal of massive traffic and massive connections.

[0003] Channel measurement, characteristic analysis and modeling are the prerequisites for the design and performance evaluation of very large-scale MIMO communication systems. After fully understanding the channel characteristics of very large-scale MIMO through channel measurement, it is very important to propose a corresponding channel model and use corresponding parameters to characterize the channel characteristics. At present, the very large-scale MIMO geometric random channel model only analyzes the channel characteristics from the space-time-frequency domain, but not from the beam domain. The beam domain channel model mainly considers the non-stationary characteristics of the spatial domain, and there are still deficiencies in the modeling of spherical wave characteristics. Therefore, it is of great significance to propose a very large-scale MIMO geometric random and beam domain wireless channel modeling method that considers the spherical wave characteristics and the non-stationary characteristics of the spatial domain. Summary of the invention

[0004] The purpose of the present invention is to provide a very large-scale MIMO geometric random and beam domain wireless channel modeling method to solve the technical problem that the current very large-scale MIMO geometric random channel model only analyzes channel characteristics from the space-time-frequency domain but not from the beam domain, and the beam domain channel model mainly considers the non-stationary characteristics of the spatial domain and there is still insufficient modeling of spherical wave characteristics.

[0005] In order to solve the above technical problems, the specific technical solutions of the present invention are as follows:

[0006] A method for modeling ultra-large-scale MIMO geometric random and beam-domain wireless channels, the method comprising:

[0007] Step S1, setting simulation parameters, network layout and scenarios;

[0008] Step S2: Set the movement speed, acceleration and direction of the transmitter to determine the movement trajectory; according to the initial positions of the transmitter and the receiver, use the sine wave superposition method to generate large-scale parameters with spatial consistency;

[0009] Step S3, according to the large-scale parameters with spatial consistency generated in step S2, the delay, power and angle parameters of the cluster are calculated to generate random cluster position parameters that obey exponential distribution;

[0010] Step S4, applying ellipsoidal Gaussian scatterer distribution to generate angles and distances of scatterers within the initial cluster relative to the transmitting end and the receiving end;

[0011] Step S401, generating parameters of intra-cluster azimuth spread, elevation spread, and cluster-to-transmitter and receiver delay spread that obey Gaussian distribution;

[0012] Step S402, using the ellipsoidal Gaussian scatterer distribution to calculate the azimuth angle, elevation angle, and distance from the scatterer to both ends of the transmitter and receiver within the cluster;

[0013] Step S5: Calculate the time-varying position of the transmitter according to the movement speed, magnitude and direction of the acceleration of the transmitter, and generate large-scale parameters with spatial consistency by using the sine wave superposition method according to the position of the transmitter and the position of the receiver;

[0014] Step S6, updating small-scale parameters by applying the birth and death process of clusters along the array axis;

[0015] Step S7, obtaining the channel coefficient of the geometric random channel model according to the formula;

[0016] Step S8, using high-order Taylor expansion to generate a steering matrix;

[0017] Step S9, calculating the beam domain channel matrix;

[0018] Step S10: Analyze channel characteristics and compare with measurement data to verify model accuracy.

[0019] Further, step S1 includes:

[0020] Step S101, simulation parameters include operating frequency, bandwidth, number of antennas at both ends of the transmitter and receiver, and antenna array arrangement, etc.;

[0021] Step S102, network layout includes setting the actual positions of the transmitting end and the receiving end, setting the placement angle of the transmitting end antenna array and the placement angle of the receiving end antenna array, etc.;

[0022] Step S103: The simulation scene includes an urban scene, a rural scene, and an indoor scene.

[0023] Furthermore, the large-scale parameters in step S2 include delay spread (DS), angular spread (AS), Rice factor (K-factor, KF) and cross polarization ratio (XPR). For line-of-sight environment, the large-scale parameters must consider the Rice factor KF, and for non-line-of-sight environment, the large-scale parameters do not need to consider the Rice factor KF.

[0024] Further, step S3 includes the following steps:

[0025] Step S301, cluster delay is expressed as:

[0026] τ n =r τ DSln(X n )

[0027] Among them, r τ is the delay distribution scaling factor, X n It follows a uniform distribution between 0 and 1, where n is the number of the cluster;

[0028] Step S302: Cluster power is expressed as:

[0029]

[0030] Among them, Z n The shadow fading of each cluster has a mean of 0 and a variance of ζ 2 The normal distribution Z n ~N(0,ζ 2 );

[0031] Step S303, generating initialization cluster angle parameters, including azimuth angle of departure (AAoD), azimuth angle of arrival (AAoA), elevation angle of departure (EAoD) and elevation angle of arrival (EAoA);

[0032] Step S304: Generate cluster position parameters for initializing exponential distribution.

[0033] Further, step S4 includes the following steps:

[0034] Step S401, the cluster center is located at the coordinate origin, and the distribution model of the scatterer position (x', y', z') in the cluster is:

[0035]

[0036] Among them, σ DS represents the standard deviation of the cluster delay spread that follows the Gaussian distribution, σ AS represents the standard deviation of the azimuthal spread of the clusters that follow the Gaussian distribution, σ ES represents the standard deviation of the cluster pitch angle spread that follows a Gaussian distribution;

[0037] Step S402: In order to indicate the position of the scatterer and reflect the spatial angle relationship, the cluster center is located at the spherical coordinate The relationship between the position of the scatterer in the cluster (x, y, z) and the position of the scatterer at the origin of the cluster center is:

[0038]

[0039] in, represents the distance from the sender or receiver to the center of the cluster, represents the average elevation angle of the scatterers in the cluster, represents the average azimuth angle of scatterers within the cluster;

[0040] Step S403: According to the relationship x=dcos(φ E )cos(φ A ), y = dcos(φ E )sin(φ A ), z = dsin(φ E ) to obtain the angle and distance of the scatterer in the cluster relative to the transmitter or receiver; d represents the distance from the transmitter or receiver to the scatterer in the cluster, φ E represents the elevation angle of the scatterer in the cluster, φ A represents the azimuth of the scatterers within the cluster.

[0041] Further, step S6 includes:

[0042] Step S601: For the transmitting end and the receiving end, the evolution process of the cluster from the pth antenna unit and the qth antenna unit to the p+1th antenna unit and the q+1th antenna unit is expressed as:

[0043]

[0044] and

[0045]

[0046] in, defines the evolution of clusters along the array axis, represents the cluster seen by the pth antenna unit at the transmitter, represents the cluster seen by the p+1th antenna unit, represents the cluster seen by the qth antenna unit at the receiving end, represents the cluster seen by the q+1th antenna unit, M T Indicates the total number of antennas at the transmitting end, M R Indicates the total number of antennas at the receiving end;

[0047] Step S602: define λ G is the birth rate of the cluster, λ R is the cluster combination rate, and the survival probability of the sending and receiving clusters along the array axis is obtained as:

[0048]

[0049] and

[0050]

[0051] in, represents the position difference of the transmitting antenna unit on the array axis, Represents the position difference of the receiving antenna unit on the array axis; represents the scene-dependent correlation factor in the array domain, δ p is the distance from the first antenna unit to the pth antenna unit at the transmitting end, δ q is the distance from the first antenna unit to the qth antenna unit at the receiving end, is the elevation angle of the transmitting antenna array, is the elevation angle of the receiving antenna array;

[0052] Step S603: The joint survival probability of the sending and receiving clusters is:

[0053]

[0054] Step S604: The average number of newly born clusters is expressed as:

[0055]

[0056] At this point, the evolution of the cluster along the array axis is completed;

[0057] Further, step S7 includes:

[0058] Step S701: Generate a channel transfer function of the line-of-sight component of a geometric random channel model:

[0059]

[0060] Among them, f c Indicates the carrier frequency, F p(q),V Indicates the antenna unit Corresponding to the directivity pattern of vertical polarization, Fp(q),H Indicates the antenna unit The directivity diagram corresponding to horizontal polarization is and is a random phase uniformly distributed in (0,2π], τ L (t) is the delay of the line-of-sight path at time t, a R,L is the steering vector at the receiving end, a T,L is the steering vector at the sender, and At time t arrive EAoD and AAoD, and At time t arrive EAoA and AAoA;

[0061] Step S702: Generate a channel transfer function of a non-line-of-sight component of a geometric random channel model:

[0062]

[0063] in, is the cross-polarization power ratio, and is a random phase uniformly distributed in (0,2π], and At time t and The power and delay of the mth subpath passing through the nth cluster, and is the random phase of the non-line-of-sight path in (0,2π] that follows a uniform distribution, and At time t EAoD and AAoD to the mth subpath of the nth cluster, and is the mth subpath of the nth cluster at time t EAoA and AAoA, N(t) and M n (t) are the number of clusters and the number of sub-paths within a cluster at time t, respectively;

[0064] Step S703: The channel transfer function of the geometric random channel model is:

[0065]

[0066] in, is the sight distance component, is the non-line-of-sight component, K R (t) is the KF at time t.

[0067] Further, step S8 includes:

[0068] Step S801: Generate a steering vector of the sight distance component and

[0069] and

[0070]

[0071] in, and The initial time arrive EAoD, AAoD and distance; At time t arrive The distance and The initial time arrive EAoA, AAoA and distance, At time t arrive The distance, λ is the carrier wavelength;

[0072] Step S802: Generate a steering vector of a non-line-of-sight component and

[0073]

[0074] and

[0075]

[0076] in, and The initial time EAoD, AAoD and distance to the mth subpath of the nth cluster; At time t The distance to the mth subpath of the nth cluster, and They are the mth subpath of the nth cluster at the initial moment to EAoA, AAoA and distance, is the mth subpath of the nth cluster at time t distance;

[0077] Step S803: The line-of-sight distance calculation method is the same as the non-line-of-sight distance calculation method, and the distance vector is calculated. and They are: and Using the vector calculation rule, we can deduce:

[0078]

[0079] and

[0080]

[0081] in,

[0082]

[0083]

[0084]

[0085] and

[0086]

[0087] in, For the initial moment The vector distance to the mth subpath of the nth cluster, For the initial moment arrive The vector distance, is the vector velocity of the transmitter at time t, is the vector velocity of the sending cluster at time t, is the mth subpath of the nth cluster at the initial moment The distance For the initial moment arrive The distance is the vector velocity of the receiving end at time t, is the vector velocity of the receiving cluster at time t, is the azimuth of the transmitting antenna array, is the azimuth angle of the receiving antenna array, α T is the azimuth of the vector velocity difference between the transmitter and the transmitter cluster, α R is the azimuth of the vector velocity difference between the receiving end and the receiving cluster, υ T is the scalar velocity of the transmitter, υ R is the scalar speed at the receiving end;

[0088] Step S804: Get the final steering vector and

[0089]

[0090] and

[0091]

[0092] in,

[0093]

[0094] and

[0095]

[0096] Among them, δ T and δ R is the distance between adjacent antenna elements of the transmitting antenna array and the receiving antenna array;

[0097] Step S805: The non-line-of-sight component channel matrix can be written as:

[0098]

[0099] Among them, S t,p The path set angle at the sender is close to:

[0100]

[0101] S r,q The path set angle at the receiving end is close to:

[0102]

[0103] Step S806: Get the final steering matrix:

[0104]

[0105] and

[0106]

[0107] Step S807: When the condition is met and here and represents a constant, and the high-order terms of the steering vector become 0. At this time, the steering matrix A T and A R is the unitary matrix that satisfies the near-field spherical wave condition.

[0108] Further, step S9 includes:

[0109] Step S901: In the steering matrix A T and A R When the unitary matrix under the near-field spherical wave condition is satisfied, the beam domain channel matrix is:

[0110]

[0111] Step S902: Line-of-sight component of beam domain channel matrix b q row and b p Column elements are represented as:

[0112]

[0113] Step S903: Non-line-of-sight component of beam domain channel matrix b q row and b p Column elements are represented as:

[0114]

[0115] Step S904: The beam domain channel matrix is ​​expressed as:

[0116]

[0117] Further, step S10 includes:

[0118] Step S1001, analyzing channel characteristics, including spatial cross-correlation function, temporal autocorrelation function, frequency correlation function and channel capacity, and fitting simulation results with measurement data;

[0119] Step S1002: Compare the simulation results and the measurement data, verify the accuracy of the model, and draw conclusions.

[0120] The ultra-large-scale MIMO geometric random and beam-domain wireless channel modeling method of the present invention has the following advantages:

[0121] The present invention supports the configuration of ultra-large-scale MIMO antenna arrays at the transmitting and receiving ends; the spherical wave and spatial non-stationary characteristics are taken into account, a high-order Taylor expansion method is used to derive the spherical wave guide vector, and the cluster birth and death process along the array axis is used to model the spatial non-stationary characteristics. The steering matrix is ​​converted into a unitary matrix through angular distance ring sampling, thereby realizing the conversion from the geometric random channel model to the beam domain channel model under near-field conditions. At the same time, under far-field conditions, the steering matrix is ​​simplified to a unitary matrix, which also realizes the conversion from the geometric random channel model to the beam domain channel model under far-field conditions. The accuracy of the model is verified by the consistency between the simulation results and the measured data, which has guiding significance for the design and performance evaluation of ultra-large-scale MIMO communication systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0122] Figure 1 A schematic flow chart of a method for modeling ultra-large-scale MIMO geometric random and beam-domain wireless channels provided in Embodiment 1 of the present invention;

[0123] Figure 2 This is a geometric random channel model diagram of ultra-large-scale MIMO in Example 1 of the present invention;

[0124] Figure 3 This is a result diagram of the spatial cross-correlation function of the ultra-large-scale MIMO geometric random and beam-domain channel models in the line-of-sight environment in Example 1 of the present invention;

[0125] Figure 4 This is a result diagram of the time autocorrelation function of the ultra-large-scale MIMO geometric random and beam-domain channel models in Example 1 of the present invention;

[0126] Figure 5 This is a frequency correlation function result diagram of the ultra-large-scale MIMO geometric random and beam domain channel models in the line-of-sight environment in Example 1 of the present invention;

[0127] Figure 6 This is a diagram showing the channel capacity results of the ultra-large-scale MIMO geometric random and beam-domain channel models in a line-of-sight environment in Example 1 of the present invention;

[0128] Figure 7 This is a graph showing the channel capacity results of the ultra-large-scale MIMO geometric random and beam-domain channel models in a non-line-of-sight environment in Example 1 of the present invention. DETAILED DESCRIPTION

[0129] In order to better understand the purpose, structure and function of the present invention, the following is a further detailed description of a very large-scale MIMO geometric random and beam domain wireless channel modeling method of the present invention in conjunction with the accompanying drawings.

[0130] Example 1

[0131] See also Figure 1-Figure 7 This embodiment provides a method for modeling ultra-large-scale MIMO geometric random and beam-domain wireless channels. The process of the method is as follows: Figure 1 As shown, the method specifically includes:

[0132] Step S1, setting simulation parameters, network layout and scenarios;

[0133] Specifically, in this embodiment, step S1 includes the following steps:

[0134] Step S101, setting the operating frequency to 5.3 GHz, the bandwidth to 160 MHz, the number of transmitting antennas to 8, the spacing between adjacent antennas to 5 cm, and making the horizontal polarization and vertical polarization arranged alternately, the number of receiving antennas to 128, the spacing between adjacent antenna units to 0.6 times the wavelength, and arranged into a uniform linear array;

[0135] Step S102, set the receiving end antenna array position coordinates to (0m, 0m, 20m), the transmitting end antenna position coordinates to (5m, -37m, 1.5m), (-20m, -37m, 1.5m) and (-25m, -37m, 1.5m), corresponding to route 1 position point 1, route 1 position point 6, route 1 position point 7 respectively;

[0136] Step S103: setting the simulation scene to a city scene.

[0137] Step S2: Setting the motion trajectory of the sending end to generate initial large-scale parameters;

[0138] Specifically, in this embodiment, step S2 includes the following steps:

[0139] Step S201: Set the movement speed of the transmitter to 2 m / s, the movement direction to (-1, 0, 0), and the acceleration to 0 m / s. 2 , at this time the motion type is uniform linear motion;

[0140] Step S202: According to the location of the transmitter, a sine wave superposition method is applied to generate large-scale parameters with spatial consistency, including delay spread DS, angle spread AS, Rice factor KF and cross-polarization ratio XPR. It should be noted that for line-of-sight environment, large-scale parameters should consider KF, while for non-line-of-sight environment, large-scale parameters do not need to consider KF.

[0141] Step S3, generating the delay, power, angle and position parameters of the initial cluster according to the initial large-scale parameters;

[0142] Specifically, in this embodiment, step S3 includes the following steps:

[0143] Step S301, cluster delay is expressed as:

[0144] τ n =r τ DSln(X n )

[0145] Among them, r τ is the delay distribution scaling factor, X n It follows a uniform distribution between 0 and 1, where n is the number of the cluster;

[0146] Step S302: Cluster power is expressed as:

[0147]

[0148] Among them, z n The shadow fading of each cluster has a mean of 0 and a variance of ζ 2 The normal distribution Z n~N(0,ζ 2 );

[0149] Step S303, generating initialization cluster angle parameters, including azimuth departure angle AAoD, azimuth arrival angle AAoA, elevation departure angle EAoD and elevation arrival angle EAoA;

[0150] Step S304: Generate cluster position parameters for initializing exponential distribution.

[0151] Step S4, applying ellipsoidal Gaussian scatterer distribution to generate angles and distances of scatterers in the initial cluster relative to the transmitting end and the receiving end;

[0152] Specifically, in this embodiment, step S4 includes the following steps:

[0153] Step S401: The cluster center is located at the coordinate origin, and the distribution of the scatterer positions (x', y', z') within the cluster can be modeled as:

[0154]

[0155] Among them, σ DS , σ AS and σ ES They represent the standard deviations of the cluster delay spread, cluster azimuth spread, and cluster elevation spread that obey Gaussian distribution respectively;

[0156] Step S402: In order to indicate the position of the scatterer and reflect the spatial angle relationship, the cluster center is located at the spherical coordinate The relationship between the position of the scatterer in the cluster (x, y, z) and the position of the scatterer at the origin of the cluster center is:

[0157]

[0158] in, represents the distance from the sender or receiver to the center of the cluster, and They represent the average elevation angle and average azimuth angle of the scatterers in the cluster respectively;

[0159] Step S403: According to the relationship x=dcos(φ E )cos(φ A ), y = dcos(φ E )sin(φ A ), z = dsin(φ E ) to obtain the angle and distance of the scatterer in the cluster relative to the transmitter or receiver. d represents the distance from the transmitter or receiver to the scatterer in the cluster, φ E and φ A represent the elevation angle and azimuth angle of the scatterers in the cluster respectively;

[0160] Step S5, calculating the position of the sending end according to the speed of the sending end, and updating the time-varying large-scale parameters;

[0161] Specifically, in this embodiment, step S5 includes the following steps:

[0162] Step S501, calculating the time-varying position of the sending end according to the movement speed, magnitude and direction of the acceleration of the sending end;

[0163] Step S502: Apply the sine wave superposition method to update the large-scale parameters with spatial consistency, including DS, AS, KF and XPR.

[0164] Step S6, updating small-scale parameters by applying the cluster birth and death process;

[0165] Specifically, in this embodiment, step S6 includes the following steps:

[0166] Step S601: For the transmitting end and the receiving end, the evolution process of the cluster from the pth antenna unit and the qth antenna unit to the p+1th antenna unit and the q+1th antenna unit is expressed as:

[0167]

[0168] and

[0169]

[0170] in, defines the evolution of clusters along the array axis, and They represent the clusters seen by the pth and p+1th antenna units at the transmitter, respectively. and They represent the clusters seen by the qth and q+1th antenna units at the receiving end, respectively, and M T and M R Represent the total number of antennas at the transmitting end and the receiving end respectively;

[0171] Step S602: define λ G is the birth rate of the cluster, λ R is the cluster combination rate, the survival probability of the sending and receiving clusters along the array axis can be obtained as:

[0172]

[0173] and

[0174]

[0175] in, and Represents the position difference between the transmitting and receiving antenna elements on the array axis. represents the scene-dependent correlation factor in the array domain, δ p and δ q are the distance from the first antenna unit to the pth antenna unit at the transmitting end and the distance from the first antenna unit to the qth antenna unit at the receiving end, respectively. and are the elevation angles of the transmitting antenna array and the receiving antenna array respectively;

[0176] Step S603: The joint survival probability of the sending and receiving clusters is:

[0177]

[0178] Step S604: The average number of newly born clusters is expressed as:

[0179]

[0180] At this point, the evolution of the cluster along the array axis is completed;

[0181] Step S605: Update corresponding small-scale parameters according to existing clusters.

[0182] Step S7, obtaining the channel coefficient of the geometric random channel model according to the formula;

[0183] Specifically, in this embodiment, step S7 includes the following steps:

[0184] Step S701: Generate a channel transfer function of the line-of-sight component of a geometric random channel model:

[0185]

[0186] Among them, f c Indicates the carrier frequency, F p(q),V and F p(q),H Indicates the antenna unit Corresponding to the directional diagram of vertical polarization and horizontal polarization, and is a random phase uniformly distributed in (0,2π], τL(t) is the delay of the line-of-sight path at time t, and a R,L and a T,L are the steering vectors at the receiving and sending ends respectively, and At time t arrive EAoD and AAoD, and At time t arrive EAoA and AAoA;

[0187] Step S702: Generate a channel transfer function of a non-line-of-sight component of a geometric random channel model:

[0188]

[0189] in, is the cross-polarization power ratio, and is a random phase uniformly distributed in (0,2π], and At time t and The power and delay of the mth subpath passing through the nth cluster, and is the random phase of the non-line-of-sight path in (0,2π] that follows a uniform distribution, and At time t EAoD and AAoD to the mth subpath of the nth cluster, and is the mth subpath of the nth cluster at time t EAoA and AAoA, N(t) and M n (t) are the number of clusters and the number of sub-paths within a cluster at time t, respectively;

[0190] Step S703: The channel transfer function of the geometric random channel model is:

[0191]

[0192] in, is the sight distance component, is the non-line-of-sight component, K R (t) is the KF at time t.

[0193] Step S8, generating a steering matrix using a high-order Taylor expansion;

[0194] Specifically, in this embodiment, step S8 includes the following steps:

[0195] Step S801: Generate a steering vector of the sight distance component and

[0196] and

[0197]

[0198] in, and The initial time arrive EAoD, AAoD and distance, At time t arrive The distance and The initial time arrive EAoA, AAoA and distance, At time t arrive The distance, λ is the carrier wavelength;

[0199] Step S802: Generate a steering vector of a non-line-of-sight component and

[0200]

[0201] and

[0202]

[0203] in, and The initial time EAoD, AAoD and distance to the mth subpath of the nth cluster, At time t The distance to the mth subpath of the nth cluster, and They are the mth subpath of the nth cluster at the initial moment to EAoA, AAoA and distance, is the mth subpath of the nth cluster at time t distance;

[0204] Step S803: Since the methods for calculating line-of-sight distance and non-line-of-sight distance are the same, we take non-line-of-sight as an example to calculate the distance vector and They are: and Using the vector calculation rule, we can further deduce:

[0205]

[0206] and

[0207]

[0208] in,

[0209]

[0210]

[0211]

[0212] and

[0213]

[0214] in, For the initial moment The vector distance to the mth subpath of the nth cluster, For the initial moment arrive The vector distance, is the vector velocity of the transmitter at time t, is the vector velocity of the sending cluster at time t, is the mth subpath of the nth cluster at the initial moment The distance For the initial moment arrive The distance is the vector velocity of the receiving end at time t, is the vector velocity of the receiving cluster at time t, is the azimuth of the transmitting antenna array, is the azimuth angle of the receiving antenna array, α T is the azimuth of the vector velocity difference between the transmitter and the transmitter cluster, α R is the azimuth of the vector velocity difference between the receiving end and the receiving cluster, υ T is the scalar velocity of the transmitter, υ R is the scalar speed at the receiving end;

[0215] Step S804: Get the final steering vector and

[0216]

[0217] and

[0218]

[0219] in,

[0220]

[0221] and

[0222]

[0223] Among them, δ T and δ Ris the distance between adjacent antenna elements of the transmitting antenna array and the receiving antenna array. Step S805, the non-line-of-sight component channel matrix can be written as:

[0224]

[0225] Among them, S t,p The path set angle at the sender is close to:

[0226]

[0227] S r,q The path set angle at the receiving end is close to:

[0228]

[0229] Step S806: Get the final steering matrix:

[0230]

[0231] and

[0232]

[0233]

[0234] Step S807: When the condition is met and here and represents a constant, and the high-order terms of the steering vector become 0. At this time, the steering matrix A T and A R is the unitary matrix that satisfies the near-field spherical wave condition.

[0235] Step S9, calculating the beam domain channel matrix;

[0236] Specifically, in this embodiment, step S9 includes the following steps:

[0237] Step S901: In the steering matrix A T and A R When the unitary matrix under the near-field spherical wave condition is satisfied, the beam domain channel matrix is:

[0238]

[0239] Step S902: Line-of-sight component of beam domain channel matrix b q row and b p Column elements are represented as:

[0240]

[0241] Step S903: Non-line-of-sight component of beam domain channel matrix b q row and b p Column elements are represented as:

[0242]

[0243] Step S904: The beam domain channel matrix is ​​expressed as:

[0244]

[0245] Step S10: Analyze channel characteristics and compare with measurement data to verify model accuracy;

[0246] Specifically, in this embodiment, the wireless channel characteristics are further analyzed by the following method, which includes:

[0247] Step S1001: Figure 3 As shown in the figure, for the spatial cross-correlation function, the geometric random channel model considering the near-field spherical wave fits the measured data better, indicating that the geometric random channel model is more accurate when considering the near-field conditions than the far-field conditions. In addition, the correlation of the beam domain channel model is lower than that of the geometric random channel model, indicating that the beam domain correlation is smaller than the array domain correlation. With the increase of antenna spacing and beam spacing, the geometric random channel model and the beam domain channel model show a downward trend, indicating the non-stationary characteristics of the spatial domain.

[0248] Step S1002: Figure 4 As shown in the figure, for the time autocorrelation function, the geometric random channel model and beam domain channel model considering near-field spherical waves fit the measured data better, indicating that considering near-field conditions is more accurate than considering far-field conditions. In addition, the correlation of the beam domain channel model and the geometric random channel model considering near-field conditions is lower than that considering far-field conditions, indicating that in a line-of-sight environment, the simplified far-field conditions reduce accuracy due to the loss of some multipaths. As the time interval increases, the time autocorrelation function shows a downward trend, indicating the non-stationary characteristics of the time domain.

[0249] Step S1003: Figure 5 As shown in the figure, for the frequency correlation function, the geometric random channel model and beam domain channel model considering near-field spherical waves fit the measured data better, indicating that considering near-field conditions is more accurate than considering far-field conditions. In addition, the correlation of the beam domain channel model and the geometric random channel model considering near-field conditions is lower than that considering far-field conditions, indicating that in a line-of-sight environment, the simplified far-field conditions reduce accuracy due to the loss of some multipaths. As the frequency interval increases, the frequency correlation function shows a downward trend, indicating the non-stationary characteristics of the frequency domain.

[0250] Step S1004: Figure 6 and Figure 7 As shown in the figure, for channel capacity, in line-of-sight environment, the geometric random channel model and beam-domain channel model considering near-field spherical waves fit the measured data better, indicating that considering near-field conditions in line-of-sight environment is more accurate than considering far-field conditions. In non-line-of-sight environment, the beam-domain channel model and the geometric random channel model both consider near-field conditions and far-field conditions and fit the measured data better, indicating that in non-line-of-sight environment, considering near-field conditions does not improve system performance.

[0251] In summary, the present invention supports the configuration of a transmitting end and a receiving end with a super-large-scale MIMO antenna array through a super-large-scale MIMO geometric random and beam domain wireless channel modeling method; the spherical wave and spatial non-stationary characteristics are taken into account, and the high-order Taylor expansion method is used to derive the spherical wave guide vector, and the cluster birth and death process along the array axis is used to model the spatial non-stationary characteristics. The steering matrix is ​​converted into a unitary matrix through angular distance ring sampling, thereby realizing the conversion from the geometric random channel model to the beam domain channel model under near-field conditions. At the same time, under far-field conditions, the steering matrix is ​​simplified to a unitary matrix, which also realizes the conversion from the geometric random channel model to the beam domain channel model under far-field conditions. The accuracy of the model is verified by the consistency between the simulation results and the measured data, which has guiding significance for the design and performance evaluation of super-large-scale MIMO communication systems.

[0252] It is to be understood that the present invention is described by some embodiments, and it is known to those skilled in the art that various changes or equivalent substitutions may be made to these features and embodiments without departing from the spirit and scope of the present invention. In addition, under the teachings of the present invention, these features and embodiments may be modified to adapt to specific circumstances and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the scope of protection of the present invention.

Claims

1. A method for modeling ultra-large-scale MIMO geometric random and beam-domain wireless channels, characterized in that: The method comprises: Step S1, setting simulation parameters, network layout and scenarios; Step S2: Set the movement speed, acceleration and direction of the transmitter to determine the movement trajectory; according to the initial positions of the transmitter and the receiver, use the sine wave superposition method to generate large-scale parameters with spatial consistency; Step S3, according to the large-scale parameters with spatial consistency generated in step S2, the delay, power and angle parameters of the cluster are calculated to generate random cluster position parameters that obey exponential distribution; Step S4, applying ellipsoidal Gaussian scatterer distribution to generate angles and distances of scatterers within the initial cluster relative to the transmitting end and the receiving end; Step S401, generating parameters of intra-cluster azimuth spread, elevation spread, and cluster-to-transmitter and receiver delay spread that obey Gaussian distribution; Step S402, using the ellipsoidal Gaussian scatterer distribution to calculate the azimuth angle, elevation angle, and distance from the scatterer to both ends of the transmitter and receiver within the cluster; Step S5: Calculate the time-varying position of the transmitter according to the movement speed, magnitude and direction of the acceleration of the transmitter, and generate large-scale parameters with spatial consistency by using the sine wave superposition method according to the position of the transmitter and the position of the receiver; Step S6, updating small-scale parameters by applying the birth and death process of clusters along the array axis; Step S7, obtaining the channel coefficient of the geometric random channel model according to the formula; Step S8, using high-order Taylor expansion to generate a steering matrix; Step S9, calculating the beam domain channel matrix; Step S10: Analyze channel characteristics and compare with measurement data to verify model accuracy.

2. The ultra-large-scale MIMO geometric random and beam-domain wireless channel modeling method according to claim 1, characterized in that: The step S1 comprises: Step S101, simulation parameters include operating frequency, bandwidth, number of antennas at both ends of the transmitter and receiver, and antenna array arrangement; Step S102, network layout includes setting the actual positions of the transmitting end and the receiving end, setting the placement angle of the transmitting end antenna array and the placement angle of the receiving end antenna array; Step S103: The simulation scenes include urban scenes, rural scenes, and indoor scenes.

3. The ultra-large-scale MIMO geometric random and beam-domain wireless channel modeling method according to claim 2, characterized in that: The large-scale parameters in step S2 include delay spread DS, angle spread AS, Ricean factor KF and cross-polarization ratio XPR. For a line-of-sight environment, the large-scale parameters need to consider the Ricean factor KF, and for a non-line-of-sight environment, the large-scale parameters do not need to consider the Ricean factor KF.

4. The method for modeling ultra-large-scale MIMO geometric random and beam-domain wireless channels according to claim 3, characterized in that: Step S3 includes the following steps: Step S301, cluster delay is expressed as: t n =r τ DSln(X n ) Among them, r τ is the delay distribution scaling factor, X n It follows a uniform distribution between 0 and 1, where n is the number of the cluster; Step S302: Cluster power is expressed as: Among them, Z n The shadow fading of each cluster has a mean of 0 and a variance of ζ 2 The normal distribution Z n ~N(0,ζ 2 ); Step S303, generating initialization cluster angle parameters, including azimuth departure angle AAoD, azimuth arrival angle AAoA, elevation departure angle EAoD and elevation arrival angle EAoA; Step S304: Generate cluster position parameters for initializing exponential distribution.

5. The method for modeling ultra-large-scale MIMO geometric random and beam-domain wireless channels according to claim 4, characterized in that: Step S4 includes the following steps: Step S401: The cluster center is located at the coordinate origin, and the distribution model of the scatterer position (x′, y′, z′) in the cluster is: Among them, σ DS represents the standard deviation of the cluster delay spread that follows the Gaussian distribution, σ AS represents the standard deviation of the azimuthal spread of the clusters that follow the Gaussian distribution, σ ES represents the standard deviation of the cluster pitch angle spread that follows a Gaussian distribution; Step S402: In order to indicate the position of the scatterer and reflect the spatial angle relationship, the cluster center is located at the spherical coordinate The relationship between the position of the scatterer in the cluster (x, y, z) and the position of the scatterer at the origin of the cluster center is: in, represents the distance from the sender or receiver to the center of the cluster, represents the average elevation angle of the scatterers in the cluster, represents the average azimuth angle of scatterers within the cluster; Step S403: According to the relationship x=dcos(φ E )cos(φ A ), y = dcos(φ E )sin(φ A ), z = dsin(φ E ) to obtain the angle and distance of the scatterer in the cluster relative to the transmitter or receiver; d represents the distance from the transmitter or receiver to the scatterer in the cluster, φ E represents the elevation angle of the scatterer in the cluster, φ A represents the azimuth of the scatterers within the cluster.

6. The method for modeling ultra-large-scale MIMO geometric random and beam-domain wireless channels according to claim 5, characterized in that: The step S6 comprises: Step S601: For the transmitting end and the receiving end, the evolution process of the cluster from the pth antenna unit and the qth antenna unit to the p+1th antenna unit and the q+1th antenna unit is expressed as: and in, defines the evolution of clusters along the array axis, represents the cluster seen by the pth antenna unit at the transmitter, represents the cluster seen by the p+1th antenna unit, represents the cluster seen by the qth antenna unit at the receiving end, represents the cluster seen by the q+1th antenna unit, M T Indicates the total number of antennas at the transmitting end, M R Indicates the total number of antennas at the receiving end; Step S602: define λ G is the birth rate of the cluster, λ R is the cluster combination rate, and the survival probability of the sending and receiving clusters along the array axis is obtained as: and in, represents the position difference of the transmitting antenna unit on the array axis, Represents the position difference of the receiving antenna unit on the array axis; represents the scene-dependent correlation factor in the array domain, δ p is the distance from the first antenna unit to the pth antenna unit at the transmitting end, δ q is the distance from the first antenna unit to the qth antenna unit at the receiving end, is the elevation angle of the transmitting antenna array, is the elevation angle of the receiving antenna array; Step S603: The joint survival probability of the sending and receiving clusters is: Step S604: The average number of newly born clusters is expressed as: At this point, the evolution of the cluster along the array axis is completed; Step S605: Update corresponding small-scale parameters according to existing clusters.

7. The method for modeling ultra-large-scale MIMO geometric random and beam-domain wireless channels according to claim 6, characterized in that: The step S7 comprises: Step S701: Generate a channel transfer function of the line-of-sight component of a geometric random channel model: Among them, f c Indicates the carrier frequency, F p(q)，V Indicates the antenna unit Corresponding to the vertical polarization pattern, F p(q)，H Indicates the antenna unit The directivity diagram corresponding to horizontal polarization is and is a random phase uniformly distributed in (0, 2π], τ L (t) is the delay of the line-of-sight path at time t, a R，L is the steering vector at the receiving end, a T，L is the steering vector at the sender, and At time t arrive EAoD and AAoD, and At time t arrive EAoA and AAoA; Step S702: Generate a channel transfer function of a non-line-of-sight component of a geometric random channel model: in, is the cross-polarization power ratio, and At time t and The power and delay of the mth subpath passing through the nth cluster, and is the random phase of the non-line-of-sight path in (0, 2π] that follows a uniform distribution, is the steering vector at the receiving end of the non-line-of-sight component, is the steering vector of the non-line-of-sight component transmitter, and At time t EAoD and AAoD to the mth subpath of the nth cluster, and is the mth subpath of the nth cluster at time t EAoA and AAoA, N(t) and M n (t) are the number of clusters and the number of sub-paths within a cluster at time t, respectively; Step S703: The channel transfer function of the geometric random channel model is: in, is the sight distance component, is the non-line-of-sight component, K R (t) is the KF at time t.

8. The method for modeling ultra-large-scale MIMO geometric random and beam-domain wireless channels according to claim 7, characterized in that: The step S8 comprises: Step S801: Generate a steering vector of the sight distance component and and in, and The initial time arrive EAoD, AAoD and distance; At time t arrive The distance and The initial time arrive EAoA, AAoA and distance, At time t arrive The distance, λ is the carrier wavelength; Step S802: Generate a steering vector of a non-line-of-sight component and and in, and The initial time EAoD, AAoD and distance to the mth subpath of the nth cluster; At time t The distance to the mth subpath of the nth cluster, and They are the mth subpath of the nth cluster at the initial moment to EAoA, AAoA and distance, is the mth subpath of the nth cluster at time t distance; Step S803: The line-of-sight distance calculation method is the same as the non-line-of-sight distance calculation method, and the distance vector is calculated. and They are: and Using the vector calculation rule, we can deduce: and in, and in, For the initial moment The vector distance to the mth subpath of the nth cluster, For the initial moment arrive The vector distance, is the vector velocity of the transmitter at time t, is the vector velocity of the sending cluster at time t, is the mth subpath of the nth cluster at the initial moment The distance For the initial moment arrive The distance is the vector velocity of the receiving end at time t, is the vector velocity of the receiving cluster at time t, is the azimuth of the transmitting antenna array, is the azimuth angle of the receiving antenna array, α T is the azimuth of the vector velocity difference between the transmitter and the transmitter cluster, α R is the azimuth of the vector velocity difference between the receiving end and the receiving cluster, υ T is the scalar velocity of the transmitter, υ R is the scalar speed at the receiving end; Step S804: Get the final steering vector and and in, and Among them, δ T and δ R is the distance between adjacent antenna elements of the transmitting antenna array and the receiving antenna array; Step S805: The non-line-of-sight component channel matrix is ​​written as: Among them, S t，p The path set angle at the sender is close to: S r，q The path set angle at the receiving end is close to: Step S806: Get the final steering matrix: and Step S807: When the condition is met and here and represents a constant, and the high-order terms of the steering vector become 0. At this time, the steering matrix A T and A R is the unitary matrix that satisfies the near-field spherical wave condition.

9. The method for modeling ultra-large-scale MIMO geometric random and beam-domain wireless channels according to claim 8, characterized in that: The step S9 comprises: Step S901: In the steering matrix A T and A R When the unitary matrix under the near-field spherical wave condition is satisfied, the beam domain channel matrix is: Step S902: Line-of-sight component of beam domain channel matrix b q row and b p Column elements are represented as: Step S903: Non-line-of-sight component of beam domain channel matrix b q row and b p Column elements are represented as: Step S904: The beam domain channel matrix is ​​expressed as:

10. The method for modeling ultra-large-scale MIMO geometric random and beam-domain wireless channels according to claim 9, characterized in that: The step S10 comprises: Step S1001, analyzing channel characteristics, including spatial cross-correlation function, temporal autocorrelation function, frequency correlation function and channel capacity, and fitting simulation results with measurement data; Step S1002: Compare the simulation results and the measurement data, verify the accuracy of the model, and draw conclusions.

Images