Adaptive transmission method based on LoS-MIMO channel cognition

Abstract

The invention discloses a self-adaptive transmission method based on LoS-MIMO channel cognition, which comprises the following steps that: at the moment t, a transmitting end performs data transmission of an LoS-MIMO channel by using an initial frame structure; the receiving end carries out channel estimation on the t moment by using the pilot symbols in the initial frame structure to obtain a channel estimation matrix H1; estimating the position of the air mobile platform by using a Gaussian kernel function method to obtain the position of the air mobile platform at the moment t ', and obtaining a prediction channel matrix H2 at the moment t' according to the characteristics of the LoS-MIMO channel matrix and the position of the air mobile platform at the moment t '; calculating the similarity between the H1 and the H2; and when the similarity between the H1 and the H2 is greater than a preset threshold eta, determining that the channel conversion at the t'moment is greater than the channel conversion at the t moment, and the channel estimation result of the current frame structure is not suitable for balancing the back-end data, thereby shortening the frame structure. Long frame structure transmission is adopted under the conditions of good channel quality and slow channel conversion, and the transmission rate is improved. And a short frame structure is adopted for transmission under the conditions of poor channel quality and fast channel conversion, so that the transmission reliability is ensured.

Description

[0001] System Field

[0002] The present invention belongs to the technical field of millimeter wave communications, and in particular relates to an adaptive transmission method based on LoS-MIMO channel cognition.

[0003] Background System

[0004] The current LoS-MIMO transmission technology usually arranges and designs the antenna before channel transmission, or adaptively adjusts the antenna angle during transmission according to the transmission distance to ensure the number of spatial multiplexing streams. However, the prior art lacks adaptive transmission methods for different transmission environments and channel conditions. Specifically, the existing MIMO spatial division multiplexing technology mainly optimizes the antenna array layout through the characteristics of the LoS-MIMO channel matrix to obtain the maximum spatial division multiplexing gain. For example, the document "A 60GHz LOS MIMO Backhaul Design Combining Spatial Multiplexing and Beamforming for a 100Gbps Throughput" designs a solution based on a layered MIMO system, which uses the upper sub-array scheduling and the lower sub-array arrangement configuration to achieve antenna gain and directivity optimization and ensure the utilization of spatial multiplexing gain. However, this solution fails to fully consider the strong channel correlation and fast channel changes in the actual transmission process.

[0005] In addition, the Chinese patent "Application No. 202311087475.4" discloses a high-speed millimeter-wave multiplexing method and proposes a channel model that combines line-of-sight MIMO spatial multiplexing and polarization multiplexing to solve problems such as limited antenna array layout of aerial mobile platforms and polarization interference in traditional MIMO polarization multiplexing. Although this method realizes multi-stream multiplexing, it does not consider the adaptive frame structure transmission method of LoS-MIMO under the problem of strong channel correlation and fast channel change at different transmission distances.

[0006] Similarly, a UAV beam tracking method based on GP position prediction is proposed in the Chinese patent "Application No. CN201811602233.3", which realizes beam tracking through position prediction. However, this method only considers position prediction, and does not consider the MIMO channel correlation enhancement problem caused by position changes in the LoS-MIMO transmission environment, especially in an environment where the channel changes rapidly, and lacks adaptive transmission scheme design.

[0007] In the existing scheme, for LoS-MIMO transmission technology, only the pre-antenna design based on the characteristics of the LoS-MIMO channel matrix is ​​considered. At present, MIMO space division multiplexing technology in the LoS-MIMO field mainly uses the characteristics of the LoS-MIMO channel matrix to obtain the layout expression of the optimal antenna array under the LoS-MIMO channel, and optimizes the antenna array layout to obtain the maximum space division multiplexing gain. Or for the limited layout of the antenna array of the aerial platform, there are problems such as polarization interference in the traditional MIMO polarization multiplexing, and a joint channel model of line-of-sight MIMO spatial multiplexing and polarization multiplexing is established to break through the optimization layout of the antenna array under the line-of-sight MIMO channel, eliminate the cross-polarization interference of ultra-wideband signals, and perform low-complexity calibration technology for spatial multi-stream multiplexing systems. At the same time, there is a Gaussian kernel function-based position prediction scheme for drones, which realizes the tracking of drone positions. Although these methods theoretically realize the space division multiplexing of LoS-MIMO transmission technology and drone position tracking, there is no corresponding solution strategy for the high correlation between the LoS-MIMO channel matrix and the transmission distance, as well as the strong channel correlation and fast channel change problems in the actual data transmission process.

[0008] In summary, the existing technology has deficiencies in processing LoS-MIMO channel correlation and rapid channel changes, which limits the application of LoS-MIMO technology in complex transmission environments. Summary of the invention

[0009] In view of the above-mentioned deficiencies existing in the existing system, the present invention provides an adaptive transmission method based on LoS-MIMO channel cognition.

[0010] The present invention provides an adaptive transmission method based on LoS-MIMO channel cognition, comprising:

[0011] At time t, the transmitter uses the initial frame structure to send data on the LoS-MIMO channel;

[0012] The receiving end uses the pilot symbols in the initial frame structure to perform channel estimation at time t and obtains the channel estimation matrix H1;

[0013] The Gaussian kernel function method is used to estimate the position of the aerial mobile platform and obtain the position of the aerial mobile platform at time t′. According to the characteristics of the LoS-MIMO channel matrix and the position of the aerial mobile platform at time t′, the predicted channel matrix H2 at time t′ is obtained.

[0014] Calculate the similarity between H1 and H2;

[0015] When the similarity between H1 and H2 is greater than a preset threshold η, it is determined that the channel change at time t′ is greater than the channel change at time t, and the channel estimation result of the current frame structure is not suitable for equalization of the subsequent data, and the frame structure is shortened;

[0016] When the similarity between H1 and H2 is less than a preset threshold η, it is determined that the channel change at time t′ is less than the channel change at time t, and the frame structure is not changed.

[0017] Preferably, calculating the similarity between H1 and H2 includes:

[0018] The similarity between H1 and H2 is calculated according to the following formula:

[0019] det(H2 -1 H1-I)

[0020] Where I is the unit matrix. When the channel estimation matrix H1 is exactly the same as the predicted channel matrix H2, det(H2 -1 H1-I)=0.

[0021] Preferably, after the receiving end uses the pilot symbols in the initial frame structure to perform channel estimation at time t and obtains the channel estimation matrix H1, it also includes:

[0022] The correlation of the column vectors of the channel estimation matrix H1 is calculated. If the correlation is lower than the set threshold, a high-code rate and high-order modulation method is used. If the correlation is higher than the set threshold, a low-code rate and low-order modulation method is used.

[0023] Preferably, the initial frame structure includes multiple subframes, one subframe includes multiple time moments, the frame header part of the frame structure includes frame header synchronization information and control information, and the subframe includes 1 pilot symbol and 20 service information symbols.

[0024] Preferably, the receiving end uses the pilot symbols in the initial frame structure to perform channel estimation at time t to obtain a channel estimation matrix H1, including:

[0025] For the transmission distance range and angle range, when the LoS-MIMO fixed antenna spacing is d, a rectangular coordinate system is established with the transmitting antenna as the x-axis horizontal plane, and the angle θ between the first transmitting and receiving antennas, the distance r between the first transmitting and receiving antennas, and the angle between the receiving antenna and the y-axis are calculated. Put it in [θ min ,θ max ]、[r min , r max ]、 The range is divided into multiple candidate points;

[0026] Using the grid search method, θ, r is searched, and its optimization function is expressed as: Where P is the pilot matrix, Y is the received signal, and H is the channel to be estimated;

[0027] by As the goal, the candidate points are selected and optimized, and the channel estimation matrix H1 is obtained according to the optimized results.

[0028] Preferably, the channel estimation matrix H1 is obtained according to the optimized result, including:

[0029] Substitute the optimized angle between the first transmitting and receiving antennas, the distance between the first transmitting and receiving antennas, and the angle between the receiving antenna and the y-axis into the following formula to obtain the distance between the receiving antenna n2 and the transmitting antenna n1

[0030]

[0031] Wherein, d1 is the distance between the n1th transmitting antenna and the first transmitting antenna, d1 = (n1-1) d, d2 is the distance between the n2th receiving antenna and the first receiving antenna, d2 = (n2-1) d;

[0032] The channel element of the channel estimation matrix H1 is expressed as: j is a complex number, f c Indicates the carrier frequency.

[0033] Preferably, the Gaussian kernel function method is used to estimate the position of the airborne mobile platform at time t′ before the next pilot symbol to obtain the position of the airborne mobile platform at time t′, including:

[0034] The positions of the sending and receiving airborne mobile platform MSI position information are denoted as x t =(x t ,y t , z t ) and x r =(x r ,y r , z r ); the input and output of the prediction of the x-coordinate of the aerial mobile platform position are the past x-coordinate information i x ={x r (tT), ..., x r (t-1)} and the future x-coordinate information o x ={x r (t), ..., x r (t+T f )}, where T f represents the number of moments predicted backward from time t, T represents the number of moments of historical information adopted forward from time T; the relationship between the output and input of the position x coordinate is expressed as ox =f(i x ), where f is the potential Gaussian function; the training set of size K consisting of the historical information of the x-coordinate of the aerial mobile platform is recorded as The test set of size S containing future information of x coordinates is denoted as The input and output of the training set are aggregated into matrices I = [i x，1 , ..., i x，K ] T Know O=[o x，1 , ..., o x，K ] T ;

[0035] The kernel function for the position prediction of the aerial mobile platform is selected as K=K1+K2, where K1 is a linear kernel function and K2 is a square exponential kernel function;

[0036] The optimal hyperparameter estimate of the kernel function is obtained by maximizing the marginal likelihood function on a specific data set. Among them, θ GP is a hyperparameter;

[0037] This allows us to obtain the output distribution after training And the predicted distribution of the future x-coordinate of the aerial mobile platform position output on the specific test set in, is the mean of the prediction distribution, is the variance of the prediction distribution, O x* ,I x* represents the output of the test set;

[0038] The future x-coordinate of the aerial mobile platform is obtained by predicting the mean of the distribution Estimated, that is, x ={x r (t), ..., x r (t+T f )} aggregated into a matrix O x =[o x，1 , ..., o x，K ] T The estimated value is

[0039] The estimated value of the x coordinate at the future time t′ is recorded as The prediction method of y coordinate and z coordinate is similar to that of and The predicted position of the aerial mobile platform at time t′ in the future is

[0040] Preferably, according to the characteristics of the LoS-MIMO channel matrix and the position of the aerial mobile platform at time t′, the predicted channel matrix H2 at time t′ is obtained, including:

[0041] according to and the predicted position of the aerial mobile platform at time t′ The predicted channel result H2 at time t′ is obtained.

[0042] Preferably, shortening the frame structure includes:

[0043] Reduce the number of service information symbols or reduce the number of subframes transmitted per unit time.

[0044] Preferably, the method further comprises:

[0045] After shortening the frame structure, if the similarity remains less than a preset threshold value η within a set time period, the frame structure length is increased.

[0046] The present invention uses Gaussian kernel function to predict the position of the aerial mobile platform in the future time slot to obtain the channel matrix information at the future time, compares it with the current state channel information obtained by channel estimation using pilot information, and optimizes the frame structure of the transmitting end according to the similarity between the two channels. In the case of good channel quality and slow channel change, long frame structure transmission is adopted to improve the transmission rate. In the case of poor channel quality and fast channel change, short frame structure transmission is adopted to ensure transmission reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A flowchart of an adaptive transmission method based on LoS-MIMO channel cognition provided in an embodiment of the present invention.

[0048] Figure 2 A schematic diagram of a frame structure design provided by an embodiment of the present invention;

[0049] Figure 3 A schematic diagram of a flow chart of an adaptive modulation and coding method provided in an embodiment of the present invention;

[0050] Figure 4 A schematic diagram of LoS-MIMO transmission provided by an embodiment of the present invention;

[0051] Figure 5 The x-axis predicted value and the true value provided by the embodiment of the present invention;

[0052] Figure 6 Another flow chart of an adaptive transmission method based on LoS-MIMO channel cognition provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0053] The exemplary embodiments will be described more fully below with reference to the accompanying drawings, but the exemplary embodiments may be embodied in different forms and should not be construed as being limited to the embodiments set forth herein. On the contrary, the purpose of providing these embodiments is to make the present disclosure thorough and complete, and to enable those skilled in the art to fully understand the scope of the present disclosure.

[0054] As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed clauses.

[0055] The terms used herein are only used to describe specific embodiments and are not intended to limit the present disclosure. As used herein, the singular forms "a", "an" and "the" are also intended to include the plural forms, unless the context clearly indicates otherwise. It will also be understood that when the terms "comprising" and / or "made of" are used in this specification, the presence of the features, wholes, steps, operations, elements and / or components is specified, but the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or groups thereof is not excluded.

[0056] The embodiments described herein may be described with reference to plan views and / or cross-sectional views by means of idealized schematic diagrams of the present disclosure. Therefore, the example illustrations may be modified according to the manufacturing system and / or tolerances. Therefore, the embodiments are not limited to the embodiments shown in the drawings, but include modifications of the configurations formed based on the manufacturing process. Therefore, the regions illustrated in the drawings have schematic properties, and the shapes of the regions shown in the figures illustrate the specific shapes of the regions of the elements, but are not intended to be limiting.

[0057] The present invention provides an adaptive transmission method based on LoS-MIMO channel recognition, which is described in detail below in conjunction with the accompanying drawings of the embodiments provided by the present invention.

[0058] like Figure 1 As shown, the present invention provides an adaptive transmission method based on LoS-MIMO channel cognition, comprising:

[0059] Step A101, at time t, the transmitting end uses the initial frame structure to send data on the LoS-MIMO channel;

[0060] Step A102, the receiving end uses the pilot symbols in the initial frame structure to perform channel estimation at time t to obtain a channel estimation matrix H1;

[0061] Step A103, using the Gaussian kernel function method, the position of the aerial mobile platform is estimated to obtain the position of the aerial mobile platform at time t′, and according to the characteristics of the LoS-MIMO channel matrix and the position of the aerial mobile platform at time t′, the predicted channel matrix H2 at time t′ is obtained;

[0062] Step A104, calculating the similarity between H1 and H2;

[0063] Step A105, when the similarity between H1 and H2 is greater than a preset threshold, it is determined that the channel change at time t′ is greater than the channel change at time t, and the channel estimation result of the current frame structure is not suitable for equalization of the subsequent data, and the frame structure is shortened;

[0064] Step A106, when the similarity between H1 and H2 is less than a preset threshold, it is determined that the channel change at time t' is less than the channel change at time t, and the frame structure is not changed.

[0065] In step A102, the receiving end uses the pilot symbol in the initial frame structure to perform channel estimation at time t, and records the subframe where the pilot symbol is located as the first subframe. In step A103, in addition to using pilot information to perform channel estimation, the LoS-MIMO channel matrix can be established by performing position prediction based on the characteristics of the LoS-MIMO channel that is highly correlated with distance and the historical position of the aerial target. Therefore, through the channel prediction scheme based on the Gaussian learning process, the channel matrix can be obtained by simply predicting the position through the motion trajectory. At the current time t, by predicting the channel at time t′ before the next pilot position of the previous data symbol (service information symbol), the channel matrix information H2 of the predicted next pilot position can be obtained. As shown Figure 2 As shown, the frame structure, i.e., a superframe, includes several subframes. One pilot symbol is inserted into one subframe. Here, the next pilot position refers to the pilot position of the next subframe of the first subframe.

[0066] In step A105-step A106, the similarity of H1 and H2 is compared. When the similarity of H1 and H2 is greater than the preset threshold η, that is, the similarity of the two channel matrices is low, it is determined that the channel has a large change in the prediction time, indicating that the position change of the aerial platform during this period of time will cause the channel matrix to change rapidly, and the current frame structure will cause the back-end data symbol equalization effect to decrease. The frame structure needs to be shortened to ensure transmission accuracy. When the similarity of H1 and H2 is less than the preset threshold η, that is, the similarity of the two channel matrices is high, it is considered that the position change of the aerial platform during this period of time will not cause the rapid change of the channel matrix, and will not cause a serious decrease in the MIMO equalization effect, and there is no need to change the current frame structure.

[0067] Under the existing LoS-MIMO technology, there is a lack of methods to achieve adaptive transmission through channel cognition. The present invention considers the adaptive transmission scheme of channel cognition under LoS-MIMO transmission technology in the scenario of aerial mobile platform to ground fixed base station, aiming to propose a LoS-MIMO technical solution that can achieve adaptive transmission under different channel conditions, so as to solve the deficiencies in the existing technology.

[0068] The present invention uses Gaussian kernel function to predict the position of the aerial mobile platform in the future time slot to obtain the channel matrix information at the future time, compares it with the current state channel information obtained by channel estimation using pilot information, and optimizes the frame structure of the transmitting end according to the similarity between the two channels. In the case of good channel quality and slow channel change, long frame structure transmission is adopted to improve the transmission rate. In the case of poor channel quality and fast channel change, short frame structure transmission is adopted to ensure transmission reliability.

[0069] Preferably, calculating the similarity between H1 and H2 includes:

[0070] The similarity between H1 and H2 is calculated according to the following formula:

[0071] det(H2 -1 H1-I)

[0072] Where I is the unit matrix. When H1 and H2 are exactly the same, det(H2 -1 H1-I)=0.

[0073] It should be noted that if the aerial mobile platform is stationary and the noise is not considered, H1 and H2 should be exactly the same, that is, H2 -1 H1=I,or det(H2 -1 H1-I)=0.

[0074] Preferably, after the receiving end uses the pilot symbols in the initial frame structure to perform channel estimation at time t and obtains the channel estimation matrix H1, it also includes:

[0075] The correlation of the column vectors of the channel estimation matrix H1 is calculated. If the correlation is lower than the set threshold, a high-code rate and high-order modulation method is used. If the correlation is higher than the set threshold, a low-code rate and low-order modulation method is used.

[0076] In the above steps, after obtaining the channel estimation result H1, the correlation of the column vector of H1 is calculated. If the correlation of the column vector is higher than the set threshold, it means that the channel quality is poor at this time. The transmitter will adaptively use a low-code rate and low-order modulation method for transmission at the next moment to ensure transmission accuracy. If the correlation of the column vector is lower than the set threshold, it means that the channel quality is good at this time. The transmitter will adaptively use a high-code rate and high-order modulation method for transmission at the next moment to improve transmission efficiency. Figure 3 As shown, it is a schematic diagram of the process of adaptive modulation and coding.

[0077] The present invention aims at the channel characteristics in the LoS-MIMO transmission process, and proposes an adaptive transmission scheme based on channel cognition of pilot information. The scheme collects pilot information to perceive the changes in channel status in real time, especially in an environment with strong channel correlation, and adaptively adjusts the modulation mode. The present invention mainly considers the transmission of actual aerial mobile platforms and fixed ground base stations. In view of the obvious change of LoS-MIMO channel matrix with distance, the pilot information is first used to obtain the LoS-MIMO channel at the current moment for channel estimation. After obtaining the estimation result, the correlation between channel column vectors is calculated. When the channel correlation is lower than the set threshold, it means that the channel quality is good at this time, and the transmitter adaptively adjusts to a high-code rate and high-order modulation mode to improve the transmission rate. When the channel correlation is higher than the set threshold, it means that the channel quality is poor at this time, and the transmitter adaptively adjusts to a low-code rate and low-order modulation mode to ensure transmission accuracy.

[0078] Preferably, the initial frame structure includes multiple subframes, one subframe includes multiple time points, and the frame header part of the frame structure includes frame header synchronization information (used for frame header synchronization) and control information (including MCS information, subframe type information, etc.). The specific frame structure is as follows: Figure 2 As shown. Currently, one pilot symbol and 20 service information symbols are inserted into one subframe, and each pilot symbol or service information symbol uses 6144 effective subcarriers. Every 200 subframes form a superframe, i.e., a frame structure. For different antennas, an additional CAZAC sequence with different root sequences is inserted into each superframe header for frame header synchronization.

[0079] The characteristics of LoS-MIMO channels are introduced below. The main application scenarios of LoS-MIMO channels are air-to-ground and air-to-air long-distance transmission scenarios. In this scenario, only the direct path is considered and the reflection path is ignored. The LoS-MIMO channel is modeled using the ray tracing principle, and its channel elements are expressed as: Where r(m, n) is the distance between the corresponding transmitting and receiving antennas. Using the channel capacity calculation formula and matrix analysis, it can be seen that for the LoS-MIMO channel, the maximum channel capacity condition is that the channel matrix column correlation is 0, that is:

[0080]

[0081] From the above formula, we can see that there are many optimal solutions to maximize the channel capacity. However, in practical applications, the shortest antenna spacing is often considered, and the constraint condition for the maximum channel capacity is:

[0082]

[0083] Among them, θ r and θ t is the angle of the transmitting and receiving antennas, cosθ r , cosθt Project it on the x-axis.

[0084] The MIMO channel capacity calculation formula is:

[0085] Where U is N t and N r The smaller number, I U is the unit matrix of U×U, det() is the determinant operation of the matrix, Q is the channel gain matrix, expressed as:

[0086]

[0087] Perform singular value decomposition on the channel matrix H to obtain:

[0088] H=UDV H

[0089] Where U and V are unitary matrices, and D is a singular value diagonal matrix. Substituting the above decomposition into the previous channel capacity formula yields:

[0090]

[0091] in, is the eigenvalue of matrix Q, that is, the square of the eigenvalue of matrix D, and r represents the rank of the channel gain matrix. According to the above analysis, the key factor determining the capacity of the LoS-MIMO channel is the distance between the transmitting and receiving antennas.

[0092] The following is a schematic diagram of a two-dimensional LoS-MIMO transmission scenario. Figure 4 As shown, preferably, the receiving end uses the pilot symbols in the initial frame structure to perform channel estimation at time t to obtain a channel estimation matrix H1, including:

[0093] For the transmission distance range and angle range, when the LoS-MIMO fixed antenna spacing is d, a rectangular coordinate system is established with the transmitting antenna as the x-axis horizontal plane, and the angle θ between the first transmitting and receiving antennas, the distance r between the first transmitting and receiving antennas, and the angle between the receiving antenna and the y-axis are calculated. Put it in [θ min ,θ max ]、[r min , r max ]、 The range is divided into multiple candidate points;

[0094] Using the grid search method, θ, r is searched, and its optimization function is expressed as:

[0095] Where P is the pilot matrix, Y is the received signal, and H is the channel to be estimated;

[0096] by As the goal, the candidate points are selected and optimized, and the channel estimation matrix H1 is obtained according to the optimized results.

[0097] Since d is known, for the LoS-MIMO channel, only the above By knowing the three parameters θ, r, we can get the complete LoS-MIMO channel information.

[0098] It should be noted that The search results of θ, r have certain errors, and the error accuracy can be preset. High accuracy leads to high complexity and long search time, while low accuracy leads to low complexity and short search time. The error accuracy can be selected according to the required accuracy, allowed search time and system computing power.

[0099] Preferably, the channel estimation matrix H1 is obtained according to the optimized result, including:

[0100] Substitute the optimized angle between the first transmitting and receiving antennas, the distance between the first transmitting and receiving antennas, and the angle between the receiving antenna and the y-axis into the following formula to obtain the distance between the receiving antenna n2 and the transmitting antenna n1

[0101]

[0102] Wherein, d1 is the distance between the n1th transmitting antenna and the first transmitting antenna, d1 = (n1-1) d, d2 is the distance between the n2th receiving antenna and the first receiving antenna, d2 = (n2-1) d;

[0103] The channel element of the channel estimation matrix H1 is expressed as: j is a complex number, f c Indicates the carrier frequency.

[0104] The Gaussian learning process is introduced below: The main task of the Gaussian learning process is to learn the mapping relationship between input and output from the empirical data (historical position information of the aerial mobile platform), and based on this, establish a prediction model for the future MAI of the airship.

[0105] Preferably, the Gaussian kernel function method is used to estimate the position of the airborne mobile platform at time t′ before the next pilot symbol to obtain the position of the airborne mobile platform at time t′, including:

[0106] The positions of the sending and receiving airborne mobile platform MSI position information are represented by X t =(x t ,y t , z t ) and X r =(x r ,yr , z r ); the input and output of the prediction of the x-coordinate of the aerial mobile platform position are the past x-coordinate information i x ={x r (tT), ..., x r (t-1)} and future x-coordinate information o x ={x r (t), ..., x r (t+T f )}, where T f represents the number of moments predicted backward from time t, T represents the number of moments of historical information adopted forward from time T; the relationship between the output and input of the position x coordinate is expressed as o x =f(i x ), where f is the potential Gaussian function; the training set of size K consisting of the historical information of the x-coordinate of the aerial mobile platform is denoted as The test set of size S containing future information of x coordinates is denoted as The input and output of the training set are aggregated into matrices I = [i x，1 , ..., i x，K ] T Know O=[o x，1 , ..., o x，K ] T ;

[0107] The kernel function for the position prediction of the aerial mobile platform is selected as K=K1+K2, where K1 is a linear kernel function and K2 is a square exponential kernel function;

[0108] The optimal hyperparameter estimate of the kernel function is obtained by maximizing the marginal likelihood function on a specific data set. Among them, θ GP is a hyperparameter;

[0109] This allows us to obtain the output distribution after training And the predicted distribution of the future x-coordinate of the aerial mobile platform position output on the specific test set in, is the mean of the prediction distribution, is the variance of the prediction distribution, O x* ,I x* Represents the output of the test set;

[0110] The future x-coordinate of the aerial mobile platform is obtained by predicting the mean of the distribution Estimation, that is, x ={x r (t), ..., x r (t+T f )} aggregated into a matrix Ox =[o x，1 , ..., 0 x，K ] T The estimated value is

[0111] The estimated value of the x coordinate at the future time t′ is recorded as The prediction method of y coordinate and z coordinate is similar to that of and The predicted position of the aerial mobile platform at time t′ in the future is

[0112] like Figure 5 As shown, it is a schematic diagram for comparing the predicted value on the x-axis with the true value.

[0113] In the above steps, according to the Gaussian process definition, f(i x ) is a Gaussian process determined by its mean and variance, expressed as follows:

[0114]

[0115] Among them, m(i x ) is the mean function, k(i x ,i x ′) is the variance (kernel) function;

[0116] I x* and O x* They are defined as the matrix forms of the input and output of the test set in the same way. It is known that the input I x , the output distribution can be written as

[0117]

[0118] Among them, m(I x ) is the mean matrix of the training set, and the i-th element is m(i x，i ), K(I x , I x ) is the covariance matrix, whose elements are expressed as K(i, j)=k(i x，i ,i x，j );

[0119] Since a zero-mean function is usually used, the training set output is O x And the test set output O x* The joint prior distribution of is expressed as follows:

[0120]

[0121] Among them, K(I x , I x*) represents the covariance of the training and test points in K×A dimensions, K(I x* , I x ) and K(I x* , I x* ) is similar to the definition, and the predicted distribution of the test set output is expressed as follows:

[0122]

[0123] The mean value of the output μ x =K(I x* , I x )K(I x , I x ) -1 O x As the predicted value, the above covariance ∑ x =K(I x* , I x* )-K(I x* , I x )K(I x , I x ) -1 K(I x , I x* ) represents the statistical uncertainty of the prediction. The predicted mean and variance of the test set output are related to the kernel function K, so it is set to fit the data pattern of the aerial mobile platform position.

[0124] It should be noted that for the sake of convenience, it is assumed that the positions and observations obtained by the airship are noise-free, but the Gaussian process-based learning method can also handle the prediction of noisy observations.

[0125] For the position prediction of the aerial mobile platform, the linear kernel function K1 is used to fit the linear relationship between the positions at adjacent moments caused by inertia. Its (i, j)th element

[0126] The square exponential kernel function K2 is used to model the smooth change of position due to mechanical adjustment, and its (i, j)th element

[0127]

[0128] Preferably, according to the characteristics of the LoS-MIMO channel matrix and the position of the aerial mobile platform at time t′, the predicted channel matrix H2 at time t′ is obtained, including:

[0129] according to and the predicted position of the aerial mobile platform at time t′ The predicted channel result H2 at time t′ is obtained.

[0130] Preferably, shortening the frame structure includes:

[0131] Reduce the number of service information symbols or reduce the number of subframes transmitted per unit time.

[0132] It should be noted that reducing the number of service information symbols means reducing the number of data symbols included in each frame header.

[0133] The present invention is mainly aimed at the scenario of communication between ground fixed base stations and aerial mobile platforms using LoS-MIMO technology. According to the channel characteristics in the LoS-MIMO transmission process, a scheme is proposed to achieve adaptive frame length transmission based on the speed of channel change by comparing the similarity between the pilot information channel estimation result and the position prediction channel prediction result. Because the LoS-MIMO channel only contains line-of-sight channels, its channel elements only depend on the distance between each transceiver antenna. At the same time, the internal antenna spacing of the transceiver antenna is known during design, so the transmission distance between all antennas is obtained by the relationship between the distance between the first transceiver antenna and the angle between the transceiver antennas. Only The parameters θ, r can be used to obtain complete LoS-MIMO channel information. In view of this feature, the entire LoS-MIMO channel can be reconstructed by combining the historical motion trajectory of the aerial platform and using the next pilot position predicted by the Gaussian kernel function. Compare the channel estimated by the pilot channel at the current moment with the predicted channel. If the two channel matrices are close, it means that the channel change in the future time slot is not obvious, and there is no need to change the transmission frame structure. If the two channels are not similar, it means that the channel change in the future time slot is obvious, and it is necessary to shorten the number of service information symbols in the frame structure or reduce the number of subframes transmitted per unit time to ensure transmission accuracy. Compare the channel information estimated by the current channel with the channel conditions in the future time slot obtained by position prediction to determine the speed of channel change and realize adaptive transmission under frame structures of different lengths.

[0134] Preferably, the method further comprises:

[0135] After shortening the frame structure, if the similarity remains less than a preset threshold value η within a set time period, the frame structure length is increased.

[0136] like Figure 6 As shown, it is another flow chart of the adaptive transmission method based on LoS-MIMO channel cognition provided by an embodiment of the present invention. Repeat the above process to ensure that when the channel changes slowly, the channel estimation is accurate, each frame contains more service information symbols, and the transmission rate is improved. When the channel changes quickly, the channel estimation is inaccurate, the length of each frame is shortened, the number of service information symbols is reduced, or the number of subframes transmitted per unit time is reduced to ensure transmission accuracy. After shortening the frame structure, if the H1 and H2 received within a period of time are relatively close, it means that the channel state changes slowly during this period of time, and the frame structure can be increased to increase the transmission efficiency.

[0137] It is to be understood that the above embodiments are merely exemplary embodiments for illustrating the principles of the present invention, but the present invention is not limited thereto. For ordinary system personnel in this field, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered as the protection scope of the present invention.

Claims

1. An adaptive transmission method based on LoS-MIMO channel cognition, characterized in that: include: At time t, the transmitter uses the initial frame structure to send data on the LoS-MIMO channel; The receiving end uses the pilot symbols in the initial frame structure to perform channel estimation at time t and obtains the channel estimation matrix H1; The Gaussian kernel function method is used to estimate the position of the aerial mobile platform and obtain the position of the aerial mobile platform at time t′. According to the characteristics of the LoS-MIMO channel matrix and the position of the aerial mobile platform at time t′, the predicted channel matrix H2 at time t′ is obtained. Calculate the similarity between H1 and H2; When the similarity between H1 and H2 is greater than a preset threshold η, it is determined that the channel change at time t′ is greater than the channel change at time t, and the channel estimation result of the current frame structure is not suitable for equalization of the subsequent data, and the frame structure is shortened; When the similarity between H1 and H2 is less than a preset threshold η, it is determined that the channel change at time t′ is less than the channel change at time t, and the frame structure is not changed.

2. The adaptive transmission method based on LoS-MIMO channel cognition according to claim 1, characterized in that: Calculate the similarity between H1 and H2, including: The similarity between H1 and H2 is calculated according to the following formula: it(H2 -1 H1-I) Where I is the unit matrix. When H1 and H2 are exactly the same, det(H2 -1 H1-I)=0.

3. The adaptive transmission method based on LoS-MIMO channel cognition according to claim 1, characterized in that: The receiving end uses the pilot symbols in the initial frame structure to perform channel estimation at time t, and after obtaining the channel estimation matrix H1, it also includes: The column vector correlation of the channel estimation matrix H1 is calculated. If the correlation is lower than the set threshold, a high-code rate and high-order modulation method is adopted. If the correlation is higher than the set threshold, a low-code rate and low-order modulation method is adopted.

4. The adaptive transmission method based on LoS-MIMO channel cognition according to claim 1, characterized in that: The initial frame structure includes multiple subframes, one subframe includes multiple time points, the frame header part of the frame structure includes frame header synchronization information and control information, and the subframe includes 1 pilot symbol and 20 service information symbols.

5. The adaptive transmission method based on LoS-MIMO channel cognition according to claim 1, characterized in that: The receiving end uses the pilot symbols in the initial frame structure to perform channel estimation at time t and obtains the channel estimation matrix H1, including: For the transmission distance range and angle range, when the LoS-MIMO fixed antenna spacing is d, a rectangular coordinate system is established with the transmitting antenna as the x-axis horizontal plane, and the angle θ between the first transmitting and receiving antennas, the distance r between the first transmitting and receiving antennas, and the angle between the receiving antenna and the y-axis are calculated. Put it in [θ min ,θ max ]、[r min , r max ]、 The range is divided into multiple candidate points; Using the grid search method, θ, r is searched, and its optimization function is expressed as: Where P is the pilot matrix, Y is the received signal, and H is the channel to be estimated; by As the goal, the candidate points are selected and optimized, and the channel estimation matrix H1 is obtained according to the optimized results.

6. The adaptive transmission method based on LoS-MIMO channel cognition according to claim 5, characterized in that: According to the optimized results, the channel estimation matrix H1 is obtained, including: Substitute the optimized angle between the first transmitting and receiving antennas, the distance between the first transmitting and receiving antennas, and the angle between the receiving antenna and the y-axis into the following formula to obtain the distance between the receiving antenna n2 and the transmitting antenna n1 Wherein, d1 is the distance between the n1th transmitting antenna and the first transmitting antenna, d1 = (n1-1) d, d2 is the distance between the n2th receiving antenna and the first receiving antenna, d2 = (n2-1) d; The channel element of the channel estimation matrix H1 is expressed as: j is a complex number, f c Indicates the carrier frequency.

7. The adaptive transmission method based on LoS-MIMO channel cognition according to claim 1, characterized in that: The Gaussian kernel function method is used to estimate the position of the airborne mobile platform at time t′ before the next pilot symbol, and the position of the airborne mobile platform at time t′ is obtained, including: The positions of the sending and receiving airborne mobile platform MSI position information are represented by X t =(x t ,y t , z t ) and X r =(x r ,y r , z r ); the input and output of the prediction of the x-coordinate of the aerial mobile platform position are the past x-coordinate information i x ={x r (tT), ..., x r (t-1)} and the future x-coordinate information o x ={x r (t), ..., x r (t+T f )}, where T f represents the number of moments predicted backward from time t, T represents the number of moments of historical information adopted forward from time T; the relationship between the output and input of the position x coordinate is expressed as o x =f(i x ), where f is the potential Gaussian function; the training set of size K consisting of the historical information of the x-coordinate of the aerial mobile platform is recorded as The test set of size S containing future information of x coordinates is denoted as The input and output of the training set are aggregated into matrices I = [i x，1 , ..., i x，K ] T andO=[o x，1 , ..., o x，K ] T ; The kernel function for the position prediction of the aerial mobile platform is selected as K=K1+K2, where K1 is a linear kernel function and K2 is a square exponential kernel function; The optimal hyperparameter estimate of the kernel function is obtained by maximizing the marginal likelihood function on a specific data set. Among them, θ GP is a hyperparameter; This allows us to obtain the output distribution after training And the predicted distribution of the future x-coordinate of the aerial mobile platform position output on the specific test set in, is the mean of the prediction distribution, is the variance of the prediction distribution, O x* ,I x* Represents the output of the test set; The future x-coordinate of the aerial mobile platform is obtained by predicting the mean of the distribution Estimated, that is, x ={x r (t), ..., x r (t+T f )} aggregated into a matrix O x =[o x，1 , ..., o x，K ] T The estimated value is The estimated value of the x coordinate at the future time t′ is recorded as The prediction method of y coordinate and z coordinate is similar to that of and The predicted position of the aerial mobile platform at time t′ in the future is 8. The adaptive transmission method based on LoS-MIMO channel cognition according to claim 7, characterized in that: According to the characteristics of the LoS-MIMO channel matrix and the position of the aerial mobile platform at time t′, the predicted channel matrix H2 at time t′ is obtained, including: according to and the predicted position of the aerial mobile platform at time t′ The predicted channel result H2 at time t′ is obtained.

9. The adaptive transmission method based on LoS-MIMO channel cognition according to claim 1, characterized in that: Shortening the frame structure, comprising: Reduce the number of service information symbols or reduce the number of subframes transmitted per unit time.

10. The adaptive transmission method based on LoS-MIMO channel cognition according to claim 1, characterized in that: The method further comprises: After shortening the frame structure, if the similarity remains less than a preset threshold value η within a set time period, the frame structure length is increased.

Images