Hybrid beamforming system design method for single-user multi-relay MIMO Systems

Abstract

The present invention is a design for a hybrid beamforming transceiver in a cooperative communication environment with multiple repeaters. Each node is based on a MIMO system. The present invention designs analog RF and digital baseband signal processing by separating them using the characteristics of unitary conversion to maximize achievable speed. In addition, it designs a transceiver by decomposing complex optimization problems into subproblems using the concept of statistical independence between the first hop and second hop channels. First, the analog RF of the source and destination nodes is designed, and based on this, the analog RF beamforming of the repeater is designed. Next, the digital precoder and combiner are designed. Digital baseband signal processing minimizes interference between different sub-channels, and the digital combiner and precoder are derived using equivalent baseband channels corresponding to different sub-carriers.

Description

Technology Field

[0001] The present invention is a transmit / receive hybrid beamforming design technology for increasing transmission efficiency between multiple repeaters, source terminals, and destination terminals in a broadband MIMO (Multiple-input and multiple-output) cooperative communication system. Background Technology

[0002] Current communication systems are becoming more diverse in terms of services, leading to a demand for high-speed and high-capacity data services. Furthermore, there is a need for low latency and high frequency efficiency. To satisfy these requirements, technologies such as MIMO and beamforming have been developed. Additionally, high frequencies are utilized to achieve high transmission speeds, and increasingly higher frequency bands are being considered. However, while using high frequencies for beamforming offers the advantage of allowing more antennas to be used by narrowing the spacing between multiple antennas, it is disadvantageous for long-distance signal transmission due to the rapid decrease in signal strength, resulting in excessive path loss, atmospheric absorption by rain and various gases, and penetration loss. Overcoming this requires deploying a large number of base stations, but this is costly. Therefore, cooperative communication can be considered to complement this. Cooperative communication is a technology that shares a user's antenna to form a virtual antenna and assists in the transmission of other terminals. Through this, coverage can be expanded and diversity gains can be achieved. The problem to be solved

[0003] This invention utilizes multiple antennas and beamforming to enhance transmission efficiency in broadband cooperative communication systems. However, constructing a beamforming system presents various challenges, one of which is the non-convex optimization problem, where global optimization values ​​are difficult to find. Since this is mathematically complex to solve, it becomes difficult to implement beamforming in practice. Therefore, this invention aims to resolve this issue and provide a method for designing a beamforming system by obtaining a simpler optimization method. means of solving the problem

[0004] The present invention transforms an originally complex optimization problem into a relatively simple form. Then, by utilizing the characteristics of unitary transformation, it separates RF (Radio frequency) and baseband signal processing to reduce design complexity. Additionally, by utilizing statistical independence between the first-hop and second-hop channels, it decomposes the analog beamforming design into two subproblems. Each subproblem obtains array gain through RF precoding and combining using a simplified optimization method. Finally, the beamforming system is designed by determining the digital baseband processing matrix corresponding to each subcarrier in a manner that minimizes interference between transmitted data streams by considering the observed equivalent channel. Effects of the invention

[0005] According to the present invention, by separating the design of components in the analog RF (Radio Frequency) part and the digital baseband processing part and considering them separately, the interrelationships are simplified, and mathematically complex problems in the system can be mitigated by using phase-only precoders and combiners. Furthermore, by transforming the objective function of the decomposed subproblems into an L1-norm maximization problem, the problem is simplified, thereby providing robust performance in situations where noise or distortion, such as channel imperfections and PN (Pilot Noise), is present. Accordingly, the proposed invention enables the implementation of effective beamforming that exhibits relatively good performance compared to theoretical performance in situations where channel estimation errors and PN are present, and demonstrates superior performance compared to other hybrid beamforming methods. Brief explanation of the drawing

[0006] [Fig. 1] This is a block diagram of the proposed hybrid transceiver for the single-user multiple-relay MIMO system of the proposed invention. [Fig. 2] and There are 4 repeaters in the diagram showing the spectral efficiency for SNR when. [Fig. 3] This figure shows the spectral efficiency with respect to SNR when PN is present and when CSI is imperfect. And, am. [Fig. 4] A spectral efficiency graph for SNR showing the case when there are 6 repeater nodes. [Fig. 5] This figure shows the spectral efficiency according to the number of antennas when the SNR is -15, 0, and 15. [Fig. 6] This figure shows the spectral efficiency according to the number of repeater nodes when the SNR is -15, 0, and 15. [Fig. 7] This is a diagram showing the spectral efficiency for SNR compared with the conventional method. Specific details for implementing the invention

[0008] Before explaining the specific details for implementing the invention, the notation method of mathematical formulas is explained to facilitate understanding. Uppercase is a matrix, and lowercase represents a vector. The given matrix ; , , , , , , , , are respectively Frobenius, bloke, nth column, from the first j The nth column, trace, determinant, transpose, conjugate transpose, and modulus of each element of the matrix, respectively. Is Identity matrix, , has a mean of 0, and the covariance is It represents a Gaussian distribution. are each An operation to combine a matrix or tensor of the i-th dimension, Represents the product of the mode tensor and the matrix, respectively. is a submatrix Represents a block diagonal matrix composed of is a diagonal matrix represents the expected value.

[0009] Embodiments of the present invention are described in detail with reference to the attached drawings so that those skilled in the art can easily implement the invention.

[0010] The present invention is based on a single-user multiple relay MIMO system.

[0011] To understand the present invention, the system signals are explained first, followed by the simplification of signals. Finally, the beamforming design of the source terminal and destination terminal, and the beamforming design at the relay terminal, are explained.

[0012] [Figure 1] is a diagram illustrating the overall system model. All relay terminals are assumed to operate in half-duplex mode and use the Amplify and Forward (AF) method. Additionally, it is assumed that there is no Line of Sight (LOS) between the destination terminal and the source terminal due to the occurrence of sudden path loss. The number of antennas for the source terminal, repeater terminal, and destination terminal are, respectively , It is defined as. The number of RF chains for each terminal is , It is assumed that... Also, for smooth communication Configure the system assuming that the conditions are satisfied. If is defined as the number of subcarriers of each subcarrier, the transmitted signal is It can be expressed as.

[0013] n The signal at the source node after hybrid beamforming of the i-th subcarrier can be expressed as follows.

[0014]

[0015] n is the subcarrier number, is a radio front-end (RF) beamforming matrix for controlling the radiation pattern, and is the baseband (BB) beamforming matrix for the nth subcarrier, where the first term is the analog beamformer applied to all subcarriers, and the second term is n It refers to the digital base band precoding applied to the nth subcarrier. s[n] refers to the transmission signal on the nth subcarrier.

[0017] The signal in mathematical formula 1 must satisfy the following power limiting formula.

[0018]

[0020] This concerns all parts of Songpa This means that power must be managed and limited to satisfy.

[0021] Since the present invention assumes cooperative communication, the relay node requires a hybrid combiner for processing the received signal and a hybrid precoder for generating the signal to be transmitted to the destination terminal. This is illustrated in FIG. 1.

[0022] As illustrated in Fig. 1, for data streams to reach a destination terminal / destination node from a source node through a relay node, the data streams are finally received and restored by passing through ① the baseband precoder and analog RF precoder of the source node, ② the RF combiner and RF precoder of the relay node, and ③ the RF combiner and baseband combiner of the destination terminal.

[0023] In Fig. 1, the RF combiner of the k-th repeater node to, RF precorder It can be defined as. Here k Is k It refers to the k-th repeater node. Also, the baseband combiner and precoder of the k-th repeater node, respectively , It can be defined as.

[0024] k Hybrid beamforming filter at the nth repeater It is defined as the combination of signals from the combiner and precoder, as in mathematical formula 2.

[0025]

[0027] Meanwhile, regarding the signal that has completed hybrid beamforming at the source node of Equation 1 seen earlierk At the nth repeater n The i-th subcarrier received signal can be written as follows.

[0028]

[0030] Here and this is the received signal from the repeater. And is the channel matrix of the source terminal and the repeater terminal. And is zero mean circularly symmetric complex Gaussian (ZMCSCG) noise.

[0031] Based on mathematical equations 2 and 3, the output signal of the nth subcarrier at the k-th repeater can be expressed as follows.

[0032]

[0033] Equation 4 is the output signal from the repeater is, received signal Repeater hybrid beamforming filter Expresses that it is output through.

[0034] And based on the above formula, the received signal of the nth subcarrier at the destination terminal can be expressed as follows.

[0035]

[0036] Here, And, is the total number of repeater nodes, Is k It is a channel matrix between the nth repeater and the destination terminal, and is the noise term between the two nodes. Applying a hybrid combiner to the above formula, it can be written as the following formula.

[0037]

[0038] Here, is an analog RF combiner at the target terminal, and At the target terminal nIt is the nth subcarrier digital processing component. Also is the channel at the first hop, and means the channel of the second hop. silver It is the noise vector of the relay nodes. is a hybrid beamforming filter in the repeater.

[0039] The above formula can be simply summarized as follows.

[0040]

[0041] Here, It can be defined as follows. Based on this, the possible system transmission rates can be calculated as follows.

[0042]

[0043] Here, Nsub is the number of subcarriers, n is the subcarrier index, I Ns is an identity matrix with a size of Ns, which is the number of data streams, and A[n] represents the channel matrix of the n-th subcarrier and the channel matrix of the subcarrier n.

[0044] is the noise covariance matrix, expressed by the following formula, and provides information about the noise in the system.

[0045]

[0046] And 1 / 2 represents that it is divided because cooperative communication consists of two hops.

[0047] Based on this, the design of an efficient beamforming system aims to maximize the transmission rate of Equation 8. Therefore, it can be expressed as the following optimization problem formula.

[0048]

[0049] Equation 9 is an optimization problem to find the values ​​inside [] that maximize the transmission rate R, and the following constraints are given.

[0050] ① Size constraint of RF (radio frequency) and baseband precoder matrices: The sum of the squares of the sizes of the RF and baseband precoder matrices is less than or equal to P_s

[0051]

[0052] ② Power constraint at the repeater: The transmission power at the repeater is less than or equal to P_(r,k).

[0053]

[0054] ③ Normalization conditions for the precoder and weight matrix

[0055] (Each element of the RF precoder matrix must be normalized, N t is the size of the RF precorder matrix)

[0056] (Each element of the baseband precorder matrix must be normalized, N r is the size of the baseband precorder matrix)

[0057] (Each element of the other baseband precorder matrix must be normalized, N r is the size of the baseband precorder matrix)

[0058] (Each element of the weight matrix must be normalized, N d is the size of the weight matrix)

[0059] Mathematical Equation 9 is a matrix as shown in Mathematical Equation 10 below. and its transpose matrix The Moore-Penrose similarity of It is simplified into a problem of maximizing.

[0060]

[0061] The constraints of mathematical formula 10 are as follows.

[0062] ① Transmission power limit: The sum of the squares of the sizes of the RF and baseband precorder matrices is less than or equal to P_s.

[0063]

[0064] ② Power constraint at the repeater: The transmission power at the repeater is less than or equal to P_(r,k).

[0065]

[0066] ③ Normalization condition for precoder and weight matrix: The size of each element of each matrix must satisfy the normalization condition.

[0067]

[0068] In mathematical formula 10, It is defined as such, and since the system of the present invention consists of two independent problems (RF and digital), the beamforming design can be started from Equation 11 below. The goal of the present invention is to maximize the calculated value of Equation 11.

[0069]

[0070] All variables are the same as those explained earlier, and and By using the properties, the right-hand side of mathematical equation 11 can be simplified into mathematical equations such as mathematical equation 12.

[0071]

[0072] Here, if we first break down the problem and simplify the equation by considering only analog beamforming, the following mathematical equation 13 is derived. This is the digital component and It proceeds by setting it to '1'.

[0073]

[0074] Since the purpose of the present invention is to maximize the equation in Equation 13, it can be expressed as two equations, Equation 14 and Equation 15, as follows. Here, Equation 14 represents a precoder and Equation 15 represents a combiner. The design of the beamforming system is optimized through Equations 14 and 15, which are divided into two independent problems. Through this separation of optimization problems, the RF beamforming, the precoder, and the combiner can be optimized individually, thereby improving the beamforming of the entire system.

[0075]

[0076] (Constraints of Equation 14)

[0077]

[0078]

[0079]

[0080] Constraint of (Equation 15)

[0081]

[0082]

[0083] As shown in the formula above, an alternative optimization method is used to find the analog beamforming precoder at the source node and the beamforming combiner at the destination node. Therefore, one value among several is fixed, and values ​​are found one by one. To utilize this method and After finding the value, design the RF beamforming at the k-th repeater.

[0084] first To find the value, the channel of the first hop as follows Assuming, the channel of the second hop is Assume that. Matrix The size of and K is the number of repeaters, is the number of relay nodes, is the number of antennas. Matrix The size of is, is the number of antennas of the target node.

[0085] By tensor-decomposing the source nodes and relay nodes of this channel matrix, a new matrix is ​​obtained as shown in Equations 16 and 17 below. and Acquires.

[0086]

[0087] Here, It is a low-core tensor, and is, It is a unitary matrix with an orthonormal basis.

[0088] especially, It consists of an orthonormal basis for the column space generated by the mathematical equation 17 below.

[0089]

[0090] If we perform SVD decomposition on this , It can be expressed as and This can be explained by the following mathematical formula 18.

[0091]

[0092] While accurate values ​​can be obtained through this SVD decomposition, the order is too high, making practical implementation difficult. Therefore, a simpler It can be derived through the eigenvalue decomposition of. Furthermore, the beamforming matrix of Equation 18 is difficult to apply in practice because there is no limit on the elemental amplitude. Therefore, a beamforming matrix for which the elemental amplitude is 1 is expressed as in Equation 19, and an optimization problem as in Equation 20 is set up to obtain the beamforming vector that satisfies it. Find .

[0093]

[0094] Although transforming into Equation 19 is more suitable for the source, it is very difficult to complete this transformation without avoiding errors due to the associated constant modulus constraint. Therefore, we transform this into an optimization problem that minimizes the objective function of Equation 20.

[0095]

[0096] Constraints:

[0097] Simplifying mathematical formula 20 yields mathematical formula 21, and then mathematical formula 22.

[0098]

[0099] Constraints:

[0100] To summarize this simply, it can be written as follows.

[0101]

[0102] Constraints:

[0104] As is evident from the objective function, the optimal solution for minimizing Equation 22 is to design the required RF beamformer at the source. It is obtained when. Furthermore, the obtained result indicates the conditions for deriving analog beamforming components from an unrestricted beamforming matrix under the Minimum Mean Squared Error (MMSE). In addition, the phase transformation value describes the direction in which the beamforming vector forms maximum energy. Therefore, the components of the analog beamformer at the source can be obtained as follows.

[0106]

[0107] Here means argument.

[0109] Using a similar method, the channels of the repeater node and the destination node are tensor-decomposed to form an unconstrained combiner Get ) and at the target terminal , that is, the analog beamforming combining matrix can be obtained as follows.

[0110]

[0111] Meanwhile, beamformers for each repeater node can be designed using the two previously designed ones. Expanding Equations 14 and 15 yields the following expression.

[0112]

[0113] Constraints:

[0114]

[0115]

[0116] Constraints:

[0117]

[0118]

[0119] As mentioned earlier, in the repeater and Since they are independent of each other, each repeater node can also be designed as described above. Therefore, it can be simplified as follows.

[0120]

[0121] Constraints:

[0122]

[0123]

[0124] Constraints:

[0125]

[0126]

[0127] The component that processes only the phase at the terminal in the k-th repeater, i.e., and Is and It can be obtained by maximizing each. Therefore, it can be expressed more simply with the following two mathematical equations.

[0128]

[0129] Constraints:

[0131]

[0132] Constraints:

[0133]

[0134]

[0135] The solutions to mathematical equations 29 and 30 are, respectively and This can lead to the design of. Before explaining this, the essential characteristics of optimization techniques are as follows. Optimization techniques must be robust even when data errors, noise, and other irregularities are present. The higher the ability to avoid unwanted values, the more robust the design will be. Also, Gambling is It is more robust to expected and irregular values ​​of measurements than the norm. Frobenius norm and matrix Norms are closely related to one another. Considering this perspective, the task of designing RF processing components involves the objective function of each optimization problem. This can be achieved by transforming it into a norm maximization problem. Therefore, the new objective function yields better results in channels with outliers or noise and can be expressed by the following two mathematical equations.

[0136]

[0137]

[0138]

[0139]

[0140] Constraints:

[0141]

[0142]

[0143] However, the two mathematical equations above have the problem that all elements are still non-convex. Therefore, to avoid mathematically difficult situations while resolving this issue, the constant amplitude constraint is temporarily relaxed. This is to extract the phase shift value that is offset by the constant amplitude constraint in order to better design the RF beamforming solution required at the relay node. The equation excluding the amplitude constraint can be obtained as follows.

[0144]

[0145] Constraints:

[0146]

[0147] Constraints:

[0148]

[0149] Since it can be calculated similarly as a theoretical upper bound value, it can be considered as an alternative method. This is replaced by the following mathematical formula.

[0150]

[0151] It can be seen that this is equivalent to the EGT (equal gain transmission) form. Therefore k RF combiner of the nth relay ( ) can be expressed as follows.

[0152]

[0153] Here is the normalized power value. Likewise, The RF precoding matrix values ​​are as follows.

[0154]

[0155] So far, we have looked at analog beamforming, and to complete the design of the hybrid transceiver, we must design digital beamforming, which must derive the corresponding baseband processing components for different frequency-carriers.

[0156] To achieve this goal, the equivalent baseband channel of the first hop is and the equivalent baseband channel of the second hop SVD (Singular Value Decomposition) is performed on . Through this process k At the nth repeater node n We can find the orthogonal digital precoder and combiner corresponding to the i-th frequency subcarrier. As a result of singular value decomposition, and since, k Optimal baseband precorder for each sub-carrier at the nth relay station and combiner are each The largest singular value corresponding to and column vector of It can be obtained by selecting dogs. Therefore, the following mathematical formula 37 is derived.

[0157]

[0159] Now, define the equivalent baseband channel from source to destination to determine the optimal digital precorder and combiner It is possible to design it. To do this, an equivalent baseband channel can be defined as follows.

[0160]

[0162] From mathematical equations 37 and 38, the equivalent channel is It can be expressed as, and the necessary baseband processing components are and of It can be obtained by selecting column vectors. Therefore, as shown below, This is derived.

[0163]

[0165] This design minimizes indirect interference between transmitted data streams and can derive low-dimensional digital precorders and combiners.

[0167] The beamformer configuration method according to the present invention has been described above. Although a description of the specific hardware configuration has been omitted, it is obvious to a person skilled in the art that a beamforming system can be implemented by applying the beamforming algorithm according to the present invention to the hardware configuration of a conventional signal processing device in a hybrid beamforming system.

[0169] <Example>

[0170] Figures 2, 3, 4, 5, 6, and 7 are the results of a simulation using the present invention. For the simulation, a system having 64 subcarriers is considered. Channel interference between the source node and the repeater node is ignored, and it is assumed that the distance between the source node and the repeater node is approximately the same. n The mm-Wave channel in the frequency domain at the i-th subcarrier is It can be expressed as follows in the mathematical formula.

[0171]

[0172] Here, is the number of clusters represents the number of transmission signal paths per cluster. Is l Represents the complex gain of the nth propagation path, ~ am. and represents the array response vectors of transmission and reception. And and represents the angle of departure (AoD) and angles of arrival (AoA) in terms of bearing and altitude, respectively.

[0173] In Figure 2, the transmission rate achievable by increasing the number of relay nodes can be seen. In this case, the system parameters are It is initialized to and transmitted an N_s={6,8} data stream by placing 4 relay nodes between the source and the destination. Figure 2 shows the transmission rates achievable by increasing the number of relay nodes. In this case, the system parameters are It is initialized to, and a data stream N_s={6,8} was transmitted by placing four relay nodes between the source and destination. The mm-Wave channel matrix corresponding to each frequency-carrier is these channel parameters , , It was set to [value]. It can be seen that the performance of the proposed method approaches the performance under the assumption of perfect CSI (Channel state information) when the channel estimation error is 10%. As shown in Figure 2, performance gradually decreases as the channel estimation error increases to 20% and 30%, respectively. Figure 3 shows the performance of the proposed invention in the case of PN (Phase noise). The system parameters and mm-Wave channel parameters are the same as in the previous case and were used to generate the simulation results. The simulation was performed assuming an imperfect channel as shown in the mathematical equation below.

[0174]

[0175] Here refers to the accuracy of the channel information. is an error matrix composed of independent and identically distributed (iid) Gaussian random variables. That is, It can be expressed as. From the obtained results, the spectral efficiency obtained with the proposed invention when PN is present is It can be clearly seen that using this, the performance is nearly identical to that of the incomplete channel matrix generated through Equation 41. Figure 4 is a graph showing spectral efficiency by further increasing the number of relay nodes with a relatively small number of antennas at the communication nodes. In this case, the number of antennas at the source, each repeater station, and the destination is It is configured as such, and six repeater nodes are deployed to assist in source node signal transmission. Transmits a data stream. The obtained result is the proposed invention It shows that the spectral efficiency achieved approaches the performance under the assumption of knowing perfect channel information. Furthermore, the performance of the proposed invention is In this case, performance degradation occurs gradually as the channel estimation error increases. This means that the performance curve follows a pattern similar to that of other figures, as shown in Fig. 4. This aspect demonstrates the performance consistency and utility of the proposed invention. Fig. 5 is a graph plotting the spectral efficiency performance of the proposed hybrid beamforming design according to the number of antennas at the communication node. To demonstrate the validity of the proposed method for small, medium, and large antenna counts, the number of antennas at the source, each relay station, and the destination It was gradually changed over a wide range as follows. The number of data streams It was set to [ ], and the simulation results fixed the SNR at -15, 0, and 15 [dB ]. In addition, four relay nodes were deployed to support signal transmission from the source node to the destination node.

[0176] Figure 5 shows that spectral efficiency increases rapidly when the number of antennas is small, while it exhibits a relatively slow growth rate for an intermediate number of antennas. When the number of antennas is large, the increase in spectral efficiency drops significantly, and this trend can be seen as reaching a saturation level. Along with consistent performance at various SNR levels, the spectral efficiency when the channel estimation error is 15% approaches the performance under the CSI assumption. Furthermore, PN and It can be confirmed that there is a negligible difference in performance between the two. Figure 6 is a graph showing the transmission rate according to the number of repeater nodes. All nodes use 16 antennas. It is assumed that there are 10 data streams. The remaining mm-Wave channel variables are the same as in the previous example. From the results obtained, it can be seen that the transmission rate increases significantly when a second repeater node is placed between the source and the destination. Additionally, the performance curve shows that the data transmission rate increases with the increase in repeater nodes, but the increase in the data transmission rate gradually decreases. When the change in the data transmission rate becomes negligibly small due to the addition of relay stations, the system enters a saturation level, and this behavior can help determine the maximum number of relay nodes that can be deployed to improve performance. The final figure compares the performance of the proposed figure with that of two OMP (Orthogonal matching pursuit)-based AF hybrid beamforming cooperative communication systems.

[0177] In Figure 7, [1] and [2] show the performance according to the method proposed in the paper below, respectively.

[0178] [1] J. Lee and YH Lee, "AF relaying for millimeter wave communication systems with hybrid RF / baseband MIMO processing," 2014 IEEE International Conference on Communications (ICC), Sydney, NSW, Australia, 2014, pp. 5838-5842, doi: 10.1109 / ICC.2014.6884253.

[0179] [2] H. Abbas and K. Hamdi, “Hybrid Beamformers Design for MIMO Relay Networks in Millimeter Wave,” 2017 IEEE Wireless Communications and Networking Conference (WCNC), San Francisco, CA, USA, 2017, pp. 1-6, doi: 10.1109 / WCNC.2017.7925884.

[0180] The number of antennas deployed at the communication node It is set to, A data stream is transmitted. Four relay nodes were set between the source and destination to improve transmission range, link quality, coverage, and spectrum efficiency. The mm-Wave channel parameters are the same as in the previous case. Additionally, it was assumed that the number of RF chains is equal to the number of data streams while generating the simulation results. From the results obtained, the proposed method demonstrates superior performance compared to the hybrid precoding techniques of [1], [2].

Claims

Claim 1 A method for designing beamforming for a multiplexer and a single user, comprising: a transmission signal analysis step of analyzing a signal at a source node using an analog beamformer and a digital precoder applied to the n-th subcarrier; a hybrid combiner application step of expressing the transmission signal by applying a hybrid combiner to a received signal at a destination terminal; and an optimization step of calculating a precoder and a combiner that maximize the system transmission rate for the transmission signal based on an effective channel between the source node, the multiplexer, and the destination terminal; wherein, in the optimization step, the system transmission rate is the R value calculated by the following (Equation 1). ( , : Received signal at the target terminal, :Number of source terminal antennas, Number of repeater antennas, Number of target terminal antennas, Total number of subcarriers :Transmission signal, : Analog beamformer at the source node, : Digital precoder applied to the nth subcarrier at the source terminal, RF combiner at the repeater terminal, :RF precoder at repeater terminal, k: k-th repeater, K: total number of repeaters, : Base band combiner at the repeater node, Base band precorder at the repeater node, : Channel matrix of source terminal and repeater terminal, : Noise of the signal received from the k-th repeater node, : Channel matrix between the k-th repeater and the destination terminal, : Noise generated at the target node, : Analog RF combiner at the destination terminal, : nth subcarrier digital processing component at the destination terminal, : First hop channel, : Second hop channel, : Noise vector, (Hybrid beamforming filter in repeater) Claim 2 delete Claim 3 A method for configuring a hybrid beamforming system according to claim 1, wherein, in the optimization step, maximizing the system transmission rate is converted into a first optimization problem of (Equation 2) below. (Equation 2) Claim 4 A method for configuring a hybrid beamforming system according to claim 3, wherein the above-mentioned first optimization problem is converted into the second optimization problem of (Equation 3) below. (Equation 3) Claim 5 A method for configuring a hybrid beamforming system, wherein, in paragraph 4, the second optimization problem is given as a third optimization problem that maximizes the calculated value of (Equation 4) below. (Equation 4) Claim 6 A method for configuring a hybrid beamforming system according to claim 5, comprising an optimization problem separation step of separating the third optimization problem of (Equation 4) above into a problem of optimizing a precorder and a problem of optimizing a combiner; wherein the problem of optimizing the precorder is given by (Equation 5) below and the problem of optimizing the combiner is given by (Equation 6) below. (Equation 5) (Constraints: ) (Equation 6) (Constraints: ) Claim 7 In claim 6, the step of decomposing the channel tensors of the source and relay nodes as in (Equation 7) in order to design analog beamforming at the source node and the destination node based on the above Equations 5 and 6; among these, of (Equation 7). Unconstrained beamformer given by (Equation 8) through the first column of Step of generating; a beamforming matrix with an elemental amplitude of 1, such as (Equation 9). A method for configuring a hybrid beamforming system, further comprising the step of setting and setting a fourth optimization problem set by Equation 10. (Equation 7) ( : First hop channel, : Core Tensor, , , : unitary matrix formed by orthonormal basis) (Equation 8) (Equation 9) (Equation 10) Constraints: Claim 8 A method for configuring a hybrid beamforming system according to claim 7, further comprising the step of obtaining an analog beamforming vector of a source node by simplifying the fourth optimization problem set by the above (Equation 10) into the fifth optimization problem of the following (Equation 11). (Equation 11) ( :argument value) Claim 9 In Clause 6, the channels of the repeater node and the destination node are tensor-decomposed to form an unconstrained combiner Calculate ) and the RF combining matrix from Equation 12 A step of obtaining; a method for configuring a hybrid beamforming system additionally including (Equation 12) ( :argument value) Claim 10 In Clause 9, at the relay node, excluding the amplitude constraint of the L1 norm through Equations 13 and 14 below , Step to calculate; RF combiner (through Equations 15 and 16 below) ) and precoder( A method for configuring a hybrid beamforming system that additionally includes the step of calculating the ) matrix. (Equation 13) Constraints: (Equation 14) Constraints: (Equation 15) ( : Normalized Power Value)(Formula 16) Claim 11 In paragraph 10, the equivalent baseband channel of the first hop and the equivalent baseband channel of the second hop Step of performing SVD (Singular Value Decomposition) on; optimal baseband precorder for each subcarrier at the k-th relay station and combiner are each The largest singular value corresponding to and column vector of A method for configuring a hybrid beamforming system additionally comprising a step obtained by selecting a dog; Claim 12 In paragraph 11, the step of defining an equivalent baseband channel from source to destination and performing SVD; and of Select the optimal digital precorder given by Equations 17 and 18 and combiner A method for configuring a hybrid beamforming system, additionally including a step of designing; (Equation 17) (Equation 18)

Images