Antenna angle adaptive LoS-MIMO robust adaptive transmission method and system

Abstract

The invention discloses an antenna angle self-adaptive LoS-MIMO robust adaptive transmission method and system, and the method comprises the steps: enabling a transmitting and receiving end to exchange historical distance information between the transmitting and receiving ends at a t moment, and constructing a data set; the sending end carries out Gaussian process model training at the t moment based on a Gaussian process machine learning method to obtain a kernel function and an optimal hyper-parameter; the sending end predicts the distance between the first antennas of the receiving and sending end at the t + 1 moment according to the data set, the kernel function and the optimal hyper-parameter at the t moment; the sending end calculates a corresponding LoS-MIMO channel matrix according to the distance between the first antennas of the receiving and sending end at the t + 1 moment; when the correlation of the LoS-MIMO channel matrix exceeds a preset threshold, the transmitting end adaptively adjusts an antenna angle, so that the LoS-MIMO channel capacity is kept maximized within a dynamic transmission distance; and the return sending end predicts the distance between the receiving end and the sending end according to the data set, the kernel function and the optimal hyper-parameter at the moment t and continues to be executed. The spatial degree of freedom is effectively utilized, and the spectrum efficiency of a wireless communication system is improved.

Description

Technical Field

[0001] The present invention relates to the field of millimeter wave communication technology, and in particular to a LoS-MIMO robust adaptive transmission method and system with adaptive antenna angle. Background Art

[0002] In practical applications, by reasonably configuring MIMO technical parameters, designing optimization algorithms and adopting adaptive array antenna methods, the transmission efficiency, coverage and reliability of the communication system can be improved, thereby further improving the system performance and user experience. For example, in "Reconfigurable ULAs for Line-of-Sight MIMO Transmission", the upper bound of the LoS-MIMO channel capacity under the linear antenna array arrangement is calculated, and an antenna rotation adaptive scheme is proposed. Without changing the antenna spacing, the antenna array configuration can achieve close to the LoS-MIMO channel capacity. The difference from the present invention is that this technology assumes that the channel capacity information is known and only involves linear array design, and cannot be applied to higher-dimensional antenna array configurations. The patent "A UAV Millimeter Wave Communication Beam Tracking Method" discloses a UAV prediction beam tracking algorithm based on Gaussian process machine learning, which predicts the future beam angle information of the UAV based on the historical position data of the UAV, thereby realizing UAV beam tracking. The method of this technology can achieve high-precision position prediction in high-dynamic scenarios, ensure the communication quality of high-speed terminals, and improve spectrum efficiency. The difference from the present invention is that the purpose of position prediction in this technology is mainly to achieve drone beam tracking after obtaining position information, and the position information is not used to predict channel capacity information. The patent "A high-speed millimeter wave multiplexing method" studies a high-speed millimeter wave multiplexing method, which establishes a joint channel model of LoS-MIMO spatial multiplexing and polarization multiplexing, and realizes the multiplexing method of LoS-MIMO under polarized antennas through the optimized layout of antenna arrays, elimination of cross-polarization interference of ultra-wideband signals, and low-complexity calibration of spatial multi-stream multiplexing systems. The difference from the present invention is that this technology does not take into account the characteristic that the channel capacity of LoS-MIMO is easily affected by the relative position of the transmitter and receiver, resulting in a reduction in its channel capacity.

[0003] In broadband and high-speed LoS-MIMO communication scenarios, existing technical solutions lack channel estimation methods suitable for actual air-to-ground and air-to-air dynamic long-distance transmission scenarios, which makes it difficult to estimate the true value of channel capacity information and it is impossible to obtain the maximum spatial division multiplexing gain by optimizing the antenna array layout. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, an embodiment of the present invention provides a LoS-MIMO robust adaptive transmission method and system with adaptive antenna angle.

[0005] In a first aspect, an embodiment of the present invention provides a LoS-MIMO robust adaptive transmission method with adaptive antenna angle, comprising:

[0006] The sender and receiver exchange the historical distance information between the sender and receiver at time t to build a data set;

[0007] At time t, the sender performs Gaussian process model training based on Gaussian process machine learning method to obtain the kernel function and optimal hyperparameters;

[0008] At time t, the transmitter predicts the first antenna spacing of the transmitter and receiver at time t+1 based on the data set, kernel function and optimal hyperparameters;

[0009] The transmitting end calculates the corresponding LoS-MIMO channel matrix according to the first antenna spacing of the transmitting and receiving ends at the time t+1;

[0010] When the correlation of the LoS-MIMO channel matrix exceeds the preset threshold, the transmitter adaptively adjusts the antenna angle to maximize the LoS-MIMO channel capacity within the dynamic transmission distance;

[0011] Return to the sending end and continue to execute the step of predicting the distance between the sending and receiving ends at time t according to the data set, the kernel function and the optimal hyperparameters.

[0012] Preferably, the transceiver exchanges historical distance information between the transceivers at time t to construct a data set, including:

[0013] Construct the input and output i composed of position coordinates x / y / z and x / y / z , the training set is composed of the first antenna spacing of the historical transceiver The test set is composed of the first antenna spacing of the transmitting and receiving ends to be predicted.

[0014] Preferably, the sending end performs Gaussian process model training based on a Gaussian process machine learning method at time t to obtain a kernel function and optimal hyperparameters, including:

[0015] Select the kernel function K=K1+K2 of the Gaussian process, where K1 is a linear kernel function and K2 is a square exponential kernel function;

[0016] Maximize the marginal likelihood function on the training set to obtain the optimal estimate of the hyperparameters in the kernel function where θ GP is a hyperparameter, is the optimal estimate of the hyperparameters.

[0017] Preferably, the transmitting end predicts the first antenna spacing of the transmitting and receiving end at time t+1 according to the data set, the kernel function and the optimal hyperparameter at time t, including:

[0018] According to the maximum a posteriori probability criterion, the first antenna spacing O of the historical transceiver is determined x / y / z The distance between the first antenna of the transmitting and receiving end at the future time t+1 The predicted joint distribution of is:

[0019]

[0020] Among them, K(·) and k(·) are the Gaussian process machine learning kernel matrix and kernel function respectively;

[0021] The posterior probability of the channel capacity at time t+1 is:

[0022]

[0023] in, and are the predicted mean and variance of the first antenna spacing between the transmitter and receiver at the future time t+1 respectively;

[0024] The maximum a posteriori probability of channel capacity is maximized by the gradient descent method, and the predicted value of the coordinates of the first antenna at the receiving end is obtained. t′ is the future time, r is the receiving end; the coordinates of the first antenna of the transmitting end are known to be X s (t′)=(x s (t), y s (t), z s (t)), s is the sender;

[0025] According to the Euclidean distance formula, the predicted distance between the first antenna of the transmitting and receiving end at time t′ is:

[0026]

[0027] Preferably, the transmitting end calculates the corresponding LoS-MIMO channel matrix according to the first antenna spacing of the transmitting and receiving ends at the time t+1, including:

[0028] Using the predicted first antenna spacing between the transmitter and receiver, the channel elements of LoS-MIMO are obtained according to the LoS-MIMO channel modeling principle:

[0029]

[0030] Where r(m, n) is the distance between the transmitting antenna n and the receiving antenna m, λ is the wavelength, j is a complex number, and r(m, n) is expressed by the antenna structure design as

[0031] According to the predicted first antenna spacing r between the transmitting and receiving ends 1，1 (t′), take the direction of the transmitting antenna arrangement as the y-axis, the normal direction of the transmitting antenna as the x-axis, and the first antenna of the transmitting end as the origin to establish a rectangular coordinate system; assume that the first antenna of the receiving end is connected to the first antenna of the transmitting end, and the angle between its direction and the x-axis is θ; the angle between the receiving antenna arrangement direction and the y-axis is The distance between any two adjacent antennas at the receiving end is equal, denoted as d r , the distance between any two adjacent antennas at the transmitting end is equal to d s , then the distance d between the mth antenna at the receiving end and the first antenna at the receiving end m =md r , the distance d between the nth antenna of the transmitter and the first antenna of the transmitter n =nd s .

[0032] Preferably, the MIMO channel capacity is obtained by the elements in the channel matrix H:

[0033]

[0034] Where U is N s and N r The smaller number, N s is the number of transmitting antennas, N r is the number of antennas at the receiving end, I U is the unit matrix of U×U, det() is the determinant operation of the matrix, SNR represents the signal-to-noise ratio, and Q is the channel gain matrix, which is expressed as:

[0035]

[0036] For LoS-MIMO channels, the channel matrix correlation is calculated according to the following formula:

[0037]

[0038] Among them, θ s is the transmitting antenna angle, θ r is the receiving antenna angle, cosθ r is θ r Angular projection on the x-axis, cosθ s is θ s Angular projection on the x-axis, d s is the distance between the transmitting antennas, d r is the receiving antenna spacing, h k ,h l is the channel matrix column vector, r m，k represents the distance between the mth receiving antenna and the kth transmitting antenna, r m，lIndicates the nth r The distance between the first receiving antenna and the first transmitting antenna.

[0039] Preferably, the transmitting end adaptively adjusts the antenna angle, including:

[0040] Using the channel capacity calculation formula and matrix analysis, we can see that the maximum channel capacity condition is when the channel matrix correlation is 0:

[0041]

[0042] Consider the constraint that the channel capacity is maximum when the antenna spacing is the shortest, that is:

[0043]

[0044] The transmission distance that satisfies this condition is d t d r ,θ s ,θ r The transmission distance R that maximizes the channel capacity when fixed opt , then R opt It is expressed as:

[0045]

[0046] Assume N s ≤N r , after the antenna layout is completed, d s d r Fixed, by cosθ r 、cosθ s From the range of values ​​of , it can be seen that the farthest optimal transmission distance that maximizes the channel capacity when the above antenna spacing is fixed is expressed as:

[0047]

[0048] Then, R opt =R ray cosθ s cosθ r

[0049] For any distance R within the optimal transmission distance, to maximize the channel capacity, we need to make R = R opt .

[0050] Preferably, the transmitting end adaptively adjusts the antenna angle, further comprising:

[0051] According to the condition R = R opt cosθ under s and cosθ r Adjust the angle between the transmitting and receiving antennas to meet the requirements and obtain the value for R<R rayThe adjustment angle is: Sure is the antenna rotation angle.

[0052] In a second aspect, an embodiment of the present invention further discloses a LoS-MIMO robust adaptive transmission system with adaptive antenna angle, comprising:

[0053] The transceiver is used to exchange the historical distances between the transceivers at time t and build a data set;

[0054] The transmitting end is used to perform Gaussian process model training based on the Gaussian process machine learning method at time t to obtain the kernel function and the optimal hyperparameters; predict the first antenna spacing of the transmitting and receiving end at time t+1 according to the data set, the kernel function and the optimal hyperparameters at time t; calculate the corresponding LoS-MIMO channel matrix according to the first antenna spacing of the transmitting and receiving end at time t+1; when the correlation of the LoS-MIMO channel matrix exceeds a preset threshold, adaptively adjust the antenna angle so that the LoS-MIMO channel capacity is maximized within the dynamic transmission distance.

[0055] Preferably, the transceiver is further used for:

[0056] Construct the input and output i composed of position coordinates x / y / z and x / y / z , the training set is composed of the first antenna spacing of the historical transceiver The test set is composed of the first antenna spacing of the transmitting and receiving ends to be predicted.

[0057] Compared with the traditional LoS-MIMO communication system, the present invention considers the channel capacity information estimation problem existing in the broadband high-speed LoS-MIMO communication scenario, proposes a LoS-MIMO antenna angle adaptive adaptive transmission scheme based on the Gaussian process distance prediction method, and analyzes the distance prediction scheme based on the Gaussian process and the conditions for maximizing the LoS-MIMO channel capacity and the characteristics of the LoS-MIMO channel. The position of the air-to-ground communication platform is further predicted using a machine learning scheme based on a Gaussian process. After obtaining the first antenna spacing of the transceiver, the corresponding channel matrix and channel capacity information are obtained according to the characteristics of the LoS-MIMO channel matrix. When the channel matrix correlation exceeds the preset threshold, the channel capacity is maximized within a certain transmission distance by designing a LoS-MIMO robust adaptive transmission scheme. The present invention mainly relates to MIMO space division multiplexing technology and adaptive antenna array technology in the field of wireless communications. MIMO space division multiplexing technology uses adaptive array antennas to realize multi-stream transmission, thereby effectively utilizing the spatial degrees of freedom and improving the spectrum efficiency of the wireless communication system. Adaptive antenna array technology adjusts beam coverage, suppresses interference and improves channel capacity by adaptively adjusting array configurations such as antenna angle, spacing and number of activated elements. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A flowchart of a LoS-MIMO robust adaptive transmission method with antenna angle adaptation provided by an embodiment of the present invention;

[0059] Figure 2 Another flow chart of the antenna angle adaptive LoS-MIMO robust adaptive transmission method provided by an embodiment of the present invention;

[0060] Figure 3 A result diagram of the actual and predicted first antenna spacing between the transmitting and receiving ends on the x-axis provided by an embodiment of the present invention;

[0061] Figure 4 A result diagram of the actual and predicted first antenna spacing between the transmitting and receiving ends on the y-axis provided by an embodiment of the present invention;

[0062] Figure 5 A diagram showing the transformation of the LoS-MIMO channel capacity with the first antenna spacing between the transmitting and receiving ends provided in an embodiment of the present invention;

[0063] Figure 6 An adaptive transmission channel capacity transformation diagram based on angle adaptation provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0064] Example embodiments will be described more fully below with reference to the accompanying drawings, but the example embodiments may be embodied in different forms and should not be understood as limited to the embodiments set forth herein. On the contrary, the purpose of providing these embodiments is to make the present invention thorough and complete and to enable those skilled in the art to fully understand the scope of the present invention.

[0065] In the absence of conflict, the embodiments of the present invention and the features thereof may be combined with each other.

[0066] As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0067] The terms used herein are only used to describe specific embodiments and are not intended to limit the present invention. 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.

[0068] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be understood to have a meaning consistent with their meaning in the context of the relevant art and the present invention, and will not be understood to have an idealized or overly formal meaning unless explicitly defined as such herein.

[0069] In order to enable those skilled in the art to better understand the technical solution, a LoS-MIMO robust adaptive transmission method and system with adaptive antenna angle provided by the present invention are described in detail below with reference to the accompanying drawings.

[0070] First, as Figure 1 , 2 As shown, an embodiment of the present invention provides a LoS-MIMO robust adaptive transmission method with antenna angle adaptation, including:

[0071] Step S101, the transmitter and receiver exchange historical distance information between the transmitter and receiver at time t to construct a data set;

[0072] Step S102, the sending end performs Gaussian process model training based on the Gaussian process machine learning method at time t to obtain a kernel function and optimal hyperparameters;

[0073] Step S103, the transmitting end predicts the first antenna spacing of the transmitting and receiving end at time t+1 according to the data set, the kernel function and the optimal hyperparameters at time t;

[0074] Step S104, the transmitting end calculates the corresponding LoS-MIMO channel matrix according to the first antenna spacing of the transmitting and receiving ends at time t+1;

[0075] Step S105, when the correlation of the LoS-MIMO channel matrix exceeds a preset threshold, the transmitting end adaptively adjusts the antenna angle so that the LoS-MIMO channel capacity is maximized within the dynamic transmission distance;

[0076] Return to step S103 and continue execution.

[0077] In an embodiment of the present invention, the transceiver exchanges historical distance information between the transceivers in a lower frequency band to establish a training data set. Then, based on the Gaussian process model training, the optimal hyperparameters are obtained to obtain the predicted position distribution of the aerial communication platform, that is, the first antenna spacing of the transceiver at time t+1 is predicted. Further, the LoS-MIMO channel matrix and channel capacity at future times are estimated based on the predicted first antenna spacing of the transceiver. Finally, when the channel matrix correlation is higher than a preset threshold, the antenna angle adaptive adjustment mechanism is triggered to maximize the LoS-MIMO channel capacity. Subsequently, the above distance prediction and angle adaptive adjustment process are repeated at the next moment to ensure efficient and robust information transmission throughout the flight process.

[0078] Compared with the traditional LoS-MIMO communication system, the present invention considers the channel capacity information estimation problem existing in the broadband high-speed LoS-MIMO communication scenario, proposes a LoS-MIMO antenna angle adaptive adaptive transmission scheme based on the Gaussian process distance prediction method, and analyzes the distance prediction scheme based on the Gaussian process and the conditions for maximizing the LoS-MIMO channel capacity and the characteristics of the LoS-MIMO channel. The position of the air-to-ground communication platform is further predicted using a machine learning scheme based on a Gaussian process. After obtaining the first antenna spacing of the transceiver, the corresponding channel matrix and channel capacity information are obtained according to the characteristics of the LoS-MIMO channel matrix. When the channel matrix correlation exceeds the preset threshold, the channel capacity is maximized within a certain transmission distance by designing a LoS-MIMO robust adaptive transmission scheme. The present invention mainly relates to MIMO space division multiplexing technology and adaptive antenna array technology in the field of wireless communications. MIMO space division multiplexing technology uses adaptive array antennas to realize multi-stream transmission, thereby effectively utilizing the spatial degrees of freedom and improving the spectrum efficiency of the wireless communication system. Adaptive antenna array technology adjusts beam coverage, suppresses interference and improves channel capacity by adaptively adjusting array configurations such as antenna angle, spacing and number of activated elements.

[0079] Preferably, the transceiver exchanges historical distance information between the transceivers at time t to construct a data set, including:

[0080] Construct the input and output i composed of position coordinates x / y / z and x / y / z , the training set is composed of the first antenna spacing of the historical transceiver The test set is composed of the first antenna spacing of the transmitting and receiving ends to be predicted.

[0081] In a preferred implementation, taking the transmitter predicting the receiver position as an example, assuming that the first antenna of the receiver is used as its position information prediction, the historical distance information between the transmitter and receiver may only include the position of the first antenna of the receiver.

[0082] Preferably, the sending end performs Gaussian process model training based on a Gaussian process machine learning method at time t to obtain a kernel function and optimal hyperparameters, including:

[0083] Select the kernel function K=K1+K2 of the Gaussian process, where K1 is a linear kernel function and K2 is a square exponential kernel function;

[0084] Maximize the marginal likelihood function on the training set to obtain the optimal estimate of the hyperparameters in the kernel function where θ GP is a hyperparameter, is the optimal estimate of the hyperparameters.

[0085] Preferably, the transmitting end predicts the first antenna spacing of the transmitting and receiving end at time t+1 according to the data set, the kernel function and the optimal hyperparameter at time t, including:

[0086] According to the maximum a posteriori probability criterion, the first antenna spacing O of the historical transceiver is determined x / y / z The distance O between the first antenna of the transmitting and receiving end at the future time t+1 x / y / z，* The predicted joint distribution of is:

[0087]

[0088] Among them, K(·) and k(·) are the Gaussian process machine learning kernel matrix and kernel function respectively;

[0089] The posterior probability of the channel capacity at time t+1 is:

[0090]

[0091] in, and are the predicted mean and variance of the first antenna spacing between the transmitter and receiver at the future time t+1 respectively;

[0092] The maximum a posteriori probability of channel capacity is maximized by the gradient descent method, and the predicted value of the coordinates of the first antenna at the receiving end is obtained. t′ is the future time, r is the receiving end; the coordinates of the first antenna of the transmitting end are known to be X s (t′)=(x s (t), y s (t), z s (t)), s is the sender;

[0093] According to the Euclidean distance formula, the predicted distance between the first antenna of the transmitting and receiving end at time t′ is:

[0094]

[0095] In this step, the distance prediction process based on Gaussian process is introduced. As the LoS-MIMO channel matrix information is highly correlated with the position, the first antenna spacing of the corresponding transceiver is calculated by high-precision distance prediction of the aerial communication platform, thereby predicting the LoS-MIMO channel capacity information. Therefore, this step is a position prediction method suitable for the movement of aerial flying communication platforms.

[0096] In a preferred implementation, taking the transmitter predicting the receiver position as an example, assuming that the first antenna of the receiver is used as its position information prediction, the historical distance information between the transmitter and receiver may only include the position of the first antenna of the receiver. x / y / z can be the first antenna position of the historical receiving end, O x / y / z，* It can be the position of the first antenna of the receiving end at the future time t+1, and They are respectively the predicted mean and variance of the first antenna position of the receiving end at the future time t+1.

[0097] It should be noted that in the above formula, “~” means obedience. represents normal distribution and p represents probability.

[0098] In a preferred embodiment, the distance between the first antenna of the transmitting and receiving end is as shown in the following Figure 3 and Figure 4 As shown, Figure 3 and Figure 4 The upper middle graph is the average predicted value, and the lower graph is the actual value. It can be seen that the Gaussian random process can accurately predict the first antenna spacing of the corresponding transceiver.

[0099] Preferably, the transmitting end calculates the corresponding LoS-MIMO channel matrix according to the first antenna spacing of the transmitting and receiving ends at the time t+1, including:

[0100] Using the predicted first antenna spacing between the transmitter and receiver, the channel elements of LoS-MIMO are obtained according to the LoS-MIMO channel modeling principle:

[0101]

[0102] Where r(m, n) is the distance between the transmitting antenna n and the receiving antenna m, λ is the wavelength, j is a complex number, and r(m, n) is expressed by the antenna structure design as

[0103] According to the predicted first antenna spacing r between the transmitting and receiving ends 1，1 (t′), take the direction of the transmitting antenna arrangement as the y-axis, the normal direction of the transmitting antenna as the x-axis, and the first antenna of the transmitting end as the origin to establish a rectangular coordinate system; assume that the first antenna of the receiving end is connected to the first antenna of the transmitting end, and the angle between its direction and the x-axis is θ; the angle between the receiving antenna arrangement direction and the y-axis is The distance between any two adjacent antennas at the receiving end is equal, denoted as d r , the distance between any two adjacent antennas at the transmitting end is equal to d s , then the distance d between the mth antenna at the receiving end and the first antenna at the receiving end m =md r , the distance d between the nth antenna of the transmitter and the first antenna of the transmitter n =nd s .

[0104] It should be noted that the channel elements h based on LoS-MIMO m，n The LoS-MIMO channel matrix H can be obtained.

[0105] The following is the process of maximizing channel capacity. The realization of this process depends on the prediction result of the first antenna spacing of the transceiver. Specifically, the LoS-MIMO channel matrix and channel capacity are closely related to the first antenna spacing of the transceiver (the position of the aerial communication platform). The channel capacity that can be achieved by the antenna array setting under the condition of fixed antenna spacing is calculated through position feedback.

[0106] The MIMO channel capacity is obtained by the elements in the channel matrix H:

[0107]

[0108] Where U is N s and N r The smaller number, N s is the number of transmitting antennas, N r is the number of antennas at the receiving end, I U is the unit matrix of U×U, det() is the determinant operation of the matrix, SNR represents the signal-to-noise ratio, and Q is the channel gain matrix, which is expressed as:

[0109]

[0110] For LoS-MIMO channels, the channel matrix correlation is calculated according to the following formula:

[0111]

[0112] Among them, θ s is the transmitting antenna angle, θ r is the receiving antenna angle, cosθ r is θ r Angular projection on the x-axis, cosθ s is θ s Angular projection on the x-axis, d s is the distance between the transmitting antennas, d r is the receiving antenna spacing, h k ,h l is the channel matrix column vector, r m，k represents the distance between the mth receiving antenna and the kth transmitting antenna, r m，l Indicates the nth r The distance between the first receiving antenna and the first transmitting antenna.

[0113] It should be noted that the kth transmitting antenna corresponds to the kth column vector in the channel matrix, and the 1st transmitting antenna corresponds to the 1st column vector in the channel matrix.

[0114] Preferably, the transmitting end adaptively adjusts the antenna angle, including:

[0115] Using the channel capacity calculation formula and matrix analysis, we can see that the maximum channel capacity condition is when the channel matrix correlation is 0:

[0116]

[0117] The greater the correlation of the channel matrix, the smaller the channel capacity. When the correlation of the channel matrix is ​​0, the channel capacity is the largest. From the above formula, we can see that there are many optimal solutions to maximize the channel capacity, but in practical applications, the constraint of the maximum channel capacity when the antenna spacing is the shortest is often considered. Consider the constraint of the maximum channel capacity when the antenna spacing is the shortest, that is:

[0118]

[0119] The transmission distance that satisfies this condition is d t d r ,θ s ,θ r The transmission distance R that maximizes the channel capacity when fixed opt , then R opt It is expressed as:

[0120]

[0121] Assume N s ≤N r , after the antenna layout is completed, d s d r Fixed, by cosθ r 、cosθ s From the range of values ​​of , it can be seen that the farthest optimal transmission distance (Rayleigh distance) that maximizes the channel capacity when the above antenna spacing is fixed is expressed as:

[0122]

[0123] Then, R opt =R ray cosθ s cosθ r

[0124] For any distance R within the optimal transmission distance, to maximize the channel capacity, we need to make R = R opt .

[0125] It should be noted that, considering the actual transmission environment, the antenna spacing is often limited by physical factors, and in the millimeter wave air-to-air and air-to-ground ultra-long distance transmission scenarios, the ground transmitting antenna often needs to be realized by arranging an antenna array. After the antenna array is set up, its spacing is determined and cannot be moved. In addition, the antenna position of the aerial communication platform cannot be moved at will, so it is usually impossible to change the channel capacity by directly changing the antenna spacing. Therefore, when the antenna spacing is constant, the relationship between the transmission distance and the channel capacity is the key to whether the LoS-MIMO system can ensure the channel quality and realize multi-stream data transmission, that is, R is derived. opt =R ray cosθ s cosθ r .

[0126] It can be seen from this that the optimal transmission distance is the longest transmission distance under a certain distance between the transmitting and receiving antennas. Within the optimal transmission distance, through R opt =R ray cosθ s cosθ r It can be seen that arbitrarily changing the antenna angle at one end of the transceiver can make it meet the corresponding conditions of the optimal transmission distance.

[0127] Therefore, after the antenna is arranged, the transmission distance can be determined. By adjusting the antenna spacing or antenna angle, the channel can achieve favorable transmission conditions and improve the channel capacity. When the transmission distance meets the optimal transmission distance, the eigenvalues ​​of the channel matrix are equal, and the channel capacity is maximized. At the same time, it can be seen from the above formula that when the antenna spacing, wavelength, and number of transceiver antennas are given, the optimal transmission distance is determined. Within the optimal transmission distance, it can be seen from the above formula that by adjusting the antenna angle of the transceiver, a suitable angle can always be found to change the projection distance on the x-axis, so that the channel capacity can be maximized within the optimal transmission distance. At the same time, the antenna angle rotation only needs to be performed at one end of the transceiver. As long as the appropriate transmission angle is calculated through position feedback, the channel capacity within the optimal transmission distance under the antenna setting can be maximized when a certain antenna spacing is determined.

[0128] Furthermore, the following is based on R opt =R ray cosθ s cosθ r The robust transmission scheme with antenna angle adaptation is obtained. The implementation of this part depends on the prediction of the first antenna spacing at the transceiver end and the results of the derivation of the channel capacity maximization condition. After completing the analysis and calculation of the prediction of the first antenna spacing at the transceiver end and the channel capacity maximization condition, this part conducts an empirical evaluation of the LoS-MIMO channel quality. When the channel matrix correlation is higher than the preset threshold, the poor channel quality cannot support multi-stream data transmission. For this reason, a robust adaptive transmission scheme based on antenna angle adaptation is designed to maximize the channel capacity within the optimal transmission distance.

[0129] After the antenna arrangement is determined, the optimal transmission distance can be determined. The optimal transmission distance is actually the longest transmission distance that can maximize the channel capacity under the antenna setting. For the distance within the optimal transmission, the channel capacity can be maximized by rotating the angle of any end of the receiving end. and (The channel matrix correlation is 0) It can be seen that considering only cosθ t Make adjustments.

[0130] Preferably, according to the condition R = R opt cosθ under s and cosθ r Adjust the angle between the transmitting and receiving antennas to meet the requirements and obtain the value for R<R ray The adjustment angle is: Sure is the antenna rotation angle.

[0131] It should be noted that within the optimal transmission distance, the channel capacity can be maximized by adjusting the transmitting end antenna angle. However, when the transmission distance is greater than the optimal transmission distance, because the value of cosθ is 0-1, it is impossible to find an angle that maximizes the channel capacity when transmitting at a super-optimal transmission distance.

[0132] The channel capacity of the LoS-MIMO channel matrix and the adaptive transmission channel capacity based on angle adaptation are shown in the attached figure. Figure 5 and attached Figure 6 As shown. Figure 5 When transmission is performed directly without adaptive antenna angle modulation, the channel capacity changes dramatically within the optimal transmission distance (horizontal axis), and the channel capacity jumps between the maximum and minimum values ​​multiple times. Figure 6 By adaptively adjusting the antenna angle, the channel capacity within the optimal transmission distance (horizontal axis) can be maximized.

[0133] In the present invention, the channel capacity of the wireless communication system in the air-to-ground and air-to-air dynamic long-distance transmission scenarios is predicted, and a distance prediction method based on Gaussian process is proposed to obtain high-precision position prediction information of the communication platform, and then a LoS-MIMO channel matrix estimation method based on the predicted distance is proposed. On this basis, an antenna angle adaptive adjustment method based on LoS-MIMO channel correlation is proposed, and the LoS-MIMO channel capacity maximization condition is derived according to the predicted position information. According to the estimated LoS-MIMO channel correlation, the antenna angle is adaptively adjusted so that the channel capacity is always maximized within the dynamic transmission distance. The distance prediction technology based on Gaussian process is applied to the design of the robust adaptive transmission scheme of the LoS-MIMO channel. The LoS-MIMO channel matrix and the corresponding channel capacity at the future moment are estimated by predicting the distance information. When the channel correlation exceeds the preset threshold, the antenna array angle is adaptively adjusted in advance to keep the channel capacity maximized within the optimal transmission distance. In order to obtain the maximum space division multiplexing gain by rotating the antenna angle under the condition of fixed antenna spacing, it is necessary to predict the future channel capacity information, derive the constraint conditions for the maximum channel capacity, and design the antenna angle adaptive robust transmission scheme.

[0134] Compared with the prior art, the present invention introduces the channel capacity information acquisition method into the LoS-MIMO antenna angle adaptive transmission scheme, and expands the antenna array design from a linear array to a two-dimensional planar array, thereby improving the flexibility of antenna deployment and being more suitable for air-to-ground and air-to-air high-dynamic communication scenarios. Compared with the prior art, the present invention introduces the UAV position prediction method based on Gaussian process machine learning into the channel capacity information prediction, thereby calculating the future channel capacity of LoS-MIMO in real time, and then maximizing the channel capacity within the optimal transmission distance by pre-adjusting the antenna angle. Considering that in an actual transmission environment, the antenna array is often already set up and its antenna spacing cannot be changed, a LoS-MIMO robust adaptive transmission scheme based on antenna angle adaptation is proposed. Without changing the antenna layout, the channel capacity of the actual system is optimized by adaptively adjusting the antenna angle.

[0135] In a second aspect, an embodiment of the present invention further discloses a LoS-MIMO robust adaptive transmission system with adaptive antenna angle, comprising:

[0136] The transceiver is used to exchange the historical distances between the transceivers at time t and build a data set;

[0137] The transmitting end is used to perform Gaussian process model training based on the Gaussian process machine learning method at time t to obtain the kernel function and the optimal hyperparameters; predict the distance between the transmitting and receiving ends at time t+1 according to the data set, the kernel function and the optimal hyperparameters at time t; calculate the corresponding LoS-MIMO channel matrix according to the first antenna spacing of the transmitting and receiving ends at time t+1; when the correlation of the LoS-MIMO channel matrix exceeds a preset threshold, adaptively adjust the antenna angle so that the LoS-MIMO channel capacity is maximized within the dynamic transmission distance.

[0138] Preferably, the transceiver is further used for:

[0139] Construct the input and output i composed of position coordinates x / y / z and x / y / z , the training set is composed of the first antenna spacing of the historical transceiver The test set is composed of the first antenna spacing of the transmitting and receiving ends to be predicted.

[0140] Preferably, the sending end is further used for:

[0141] Select the kernel function K=K1+K2 of the Gaussian process, where K1 is a linear kernel function and K2 is a square exponential kernel function;

[0142] Maximize the marginal likelihood function on the training set to obtain the optimal estimate of the hyperparameters in the kernel function where θ GP is a hyperparameter, is the optimal estimate of the hyperparameters.

[0143] Example embodiments have been disclosed herein, and although specific terms are employed, they are used and should be interpreted only in a general illustrative sense and not for limiting purposes. In some embodiments, it will be apparent to those skilled in the art that, unless otherwise expressly noted, features, characteristics, and / or elements described in conjunction with a particular embodiment may be used alone or in combination with features, characteristics, and / or elements described in conjunction with other embodiments. Therefore, it will be understood by those skilled in the art that various changes in form and detail may be made without departing from the scope of the invention as set forth in the appended claims.

Claims

1. A LoS-MIMO robust adaptive transmission method with antenna angle adaptation, characterized in that: include: The sender and receiver exchange the historical distance information between the sender and receiver at time t to build a data set; At time t, the sender performs Gaussian process model training based on Gaussian process machine learning method to obtain the kernel function and optimal hyperparameters; At time t, the transmitter predicts the first antenna spacing of the transmitter and receiver at time t+1 based on the data set, kernel function and optimal hyperparameters; The transmitting end calculates the corresponding LoS-MIMO line-of-sight multi-antenna channel matrix according to the first antenna spacing of the transmitting and receiving ends at the time t+1; When the correlation of the LoS-MIMO channel matrix exceeds the preset threshold, the transmitter adaptively adjusts the antenna angle to maximize the LoS-MIMO channel capacity within the dynamic transmission distance; Return to the sending end and continue to execute the step of predicting the distance between the sending and receiving ends at time t according to the data set, the kernel function and the optimal hyperparameters.

2. The antenna angle adaptive LoS-MIMO robust adaptive transmission method according to claim 1, characterized in that: The sender and receiver exchange the historical distance information between the sender and receiver at time t to construct a data set, including: Construct the input and output i composed of position coordinates x / y / z and x / y / z , the training set is composed of the first antenna spacing of the historical transceiver The test set is composed of the first antenna spacing of the transmitting and receiving ends to be predicted.

3. The antenna angle adaptive LoS-MIMO robust adaptive transmission method according to claim 2, characterized in that: At time t, the sender performs Gaussian process model training based on the Gaussian process machine learning method to obtain the kernel function and optimal hyperparameters, including: Select the kernel function K=K1+K2 of the Gaussian process, where K1 is a linear kernel function and K2 is a square exponential kernel function; Maximize the marginal likelihood function on the training set to obtain the optimal estimate of the hyperparameters in the kernel function where θ GP is a hyperparameter, is the optimal estimate of the hyperparameters.

4. The antenna angle adaptive LoS-MIMO robust adaptive transmission method according to claim 3, characterized in that: At time t, the transmitter predicts the first antenna spacing of the transmitter and receiver at time t+1 based on the data set, kernel function, and optimal hyperparameters, including: According to the maximum a posteriori probability criterion, the first antenna spacing O of the historical transceiver is determined x / y / z The distance O between the first antenna of the transmitting and receiving end at the future time t+1 x / y / z，* The predicted joint distribution of is: Among them, K(·) and k(·) are the Gaussian process machine learning kernel matrix and kernel function respectively; The posterior probability of the channel capacity at time t+1 is: in, and are the predicted mean and variance of the first antenna spacing between the transmitter and receiver at the future time t+1 respectively; The maximum a posteriori probability of channel capacity is maximized by the gradient descent method, and the predicted value of the coordinates of the first antenna at the receiving end is obtained. t′ is the future time, r is the receiving end; the coordinates of the first antenna of the transmitting end are known to be x s (t′)=(x s (t), y s (t), z s (t)), s is the sender; According to the Euclidean distance formula, the predicted distance between the first antenna of the transmitting and receiving end at time t′ is:

5. The antenna angle adaptive LoS-MIMO robust adaptive transmission method according to claim 1, characterized in that: The transmitting end calculates the corresponding LoS-MIMO channel matrix according to the first antenna spacing of the transmitting and receiving ends at time t+1, including: Using the predicted first antenna spacing between the transmitter and receiver, the channel elements of LoS-MIMO are obtained according to the LoS-MIMO channel modeling principle: Where r(m, n) is the distance between the transmitting antenna n and the receiving antenna m, λ is the wavelength, j is a complex number, and r(m, n) is expressed by the antenna structure design as According to the predicted first antenna spacing r between the transmitting and receiving ends 1，1 (t′), take the direction of the transmitting antenna arrangement as the y-axis, the normal direction of the transmitting antenna as the x-axis, and the first antenna of the transmitting end as the origin to establish a rectangular coordinate system; assume that the first antenna of the receiving end is connected to the first antenna of the transmitting end, and the angle between its direction and the x-axis is θ; the angle between the receiving antenna arrangement direction and the y-axis is The distance between any two adjacent antennas at the receiving end is equal, denoted as d r , the distance between any two adjacent antennas at the transmitting end is equal to d s , then the distance d between the mth antenna at the receiving end and the first antenna at the receiving end m =md r , the distance d between the nth antenna of the transmitter and the first antenna of the transmitter n =nd s .

6. The antenna angle adaptive LoS-MIMO robust adaptive transmission method according to claim 1, characterized in that: The MIMO channel capacity is obtained by the elements in the channel matrix H: Where U is N s and N r The smaller number, N s is the number of transmitting antennas, N r is the number of antennas at the receiving end, I U is the unit matrix of U×U, det() is the determinant operation of the matrix, SNR represents the signal-to-noise ratio, and Q is the channel gain matrix, which is expressed as: For LoS-MIMO channels, the channel matrix correlation is calculated according to the following formula: Among them, θ s is the transmitting antenna angle, θ r is the receiving antenna angle, cosθ r is θ r Angular projection on the x-axis, cosθ s is θ s Angular projection on the x-axis, d s is the distance between the transmitting antennas, d r is the receiving antenna spacing, h k ,h l is the channel matrix column vector, r m，k represents the distance between the mth receiving antenna and the kth transmitting antenna, r m，l Indicates the nth r The distance between the first receiving antenna and the first transmitting antenna.

7. The antenna angle adaptive LoS-MIMO robust adaptive transmission method according to claim 6, characterized in that: The transmitter adaptively adjusts the antenna angle, including: Using the channel capacity calculation formula and matrix analysis, we can see that the maximum channel capacity condition is when the channel matrix correlation is 0: Consider the constraint that the channel capacity is maximum when the antenna spacing is the shortest, that is: The transmission distance that satisfies this condition is d t ,d r ,θ s ,θ r The transmission distance R that maximizes the channel capacity when fixed opt , then R opt It is expressed as: Assume N s ≤N r , after the antenna layout is completed, d s d r Fixed, by cosθ r 、cosθ s From the range of values ​​of , it can be seen that the farthest optimal transmission distance that maximizes the channel capacity when the above antenna spacing is fixed is expressed as: Then, R opt = R ray cosθ s cosθ r For any distance R within the optimal transmission distance, to maximize the channel capacity, we need to make R = R opt .

8. The antenna angle adaptive LoS-MIMO robust adaptive transmission method according to claim 7, characterized in that: The transmitter adaptively adjusts the antenna angle, which also includes: According to the condition R = R opt cosθ under s and cosθ r Adjust the angle between the transmitting and receiving antennas to meet the requirements and obtain the value for R<R ray The adjustment angle is: Sure is the antenna rotation angle.

9. A LoS-MIM0 robust adaptive transmission system with antenna angle adaptation, characterized in that: include: The transceiver is used to exchange the historical distances between the transceivers at time t and build a data set; The sending end is used to train the Gaussian process model based on the Gaussian process machine learning method at time t to obtain the kernel function and the optimal hyperparameters; at time t, the distance between the sending and receiving ends at time t+1 is predicted based on the data set, the kernel function and the optimal hyperparameters; The corresponding LoS-MIM0 channel matrix is ​​calculated according to the first antenna spacing of the transceiver at the time t+1; When the LoS-MIMO channel matrix correlation exceeds a preset threshold, the antenna angle is adaptively adjusted to maximize the LoS-MIMO channel capacity within the dynamic transmission distance.

10. The antenna angle adaptive LoS-MIMO robust adaptive transmission system according to claim 9, characterized in that: The transceiver is also used for: Construct the input and output i composed of position coordinates x / y / z and x / y / z , the training set is composed of the first antenna spacing of the historical transceiver The test set is composed of the first antenna spacing of the transmitting and receiving ends to be predicted.

Images