A method and system for improving throughput based on power allocation for MIMO-URLLC short code transmission

Abstract

The application provides a throughput improvement method based on power allocation for MIMO-URLLC short code transmission, comprising the following steps: step 1: based on a multi-user MIMO communication scenario, encoding a sending symbol vector and modeling a channel, introducing an additive white Gaussian noise to obtain a receiving signal at a user end; step 2: obtaining a signal-to-interference-and-noise ratio of each user and an achievable rate expression in a limited code length domain through the receiving signal at the user end, taking power allocation as a variable, constructing a target function expression about the throughput of each user and the total throughput of the system, so as to clearly define the optimization problem of the total throughput of the system; step 3: performing convex approximation on the target function constructed in step 2 to obtain an optimization target function, and solving the optimization target function to output a final power allocation configuration. According to various specific parameters of the communication scenario, the power allocation of the sending end is optimized, so that the purpose of effectively improving the total throughput of the system is achieved.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and specifically to a method and system for improving throughput based on power allocation for MIMO-URLLC short code transmission. Background Technology

[0002] High-Reliable Low-Latency Communication (URLLC) is one of the three major services of 5G / B5G and a driving technology for scenarios such as Industrial Internet of Things (IIoT) and Vehicle-to-Everything (V2X). MU-MIMO (Multi-User Multiple-Input Multiple-Output) allows one access point (AP) to communicate with multiple terminals simultaneously, making full use of spatial resources and improving wireless throughput. It is an important multi-user technology in the field of wireless communication and is mainly used in cellular networks and Wi-Fi networks.

[0003] However, the fundamental theories of URLLC are still incomplete both domestically and internationally, especially lacking research on URLLC in multi-antenna scenarios. Furthermore, the reliability, achievable data rate, and inter-communication mechanism of massive MIMO URLLC remain unclear.

[0004] Based on this, we conduct research on URLLC in large-scale MIMO scenarios, study the influence relationship between the total throughput and power allocation strategy of the system under the finite code length field theory, analyze the achievable performance of the system, and propose a high-performance power allocation optimization method. Summary of the Invention

[0005] This invention proposes a power allocation optimization method for MIMO-URLLC scenarios based on FBL, which can optimize the power allocation of the transmitting end according to various specific parameters of the communication scenario, thereby effectively improving the total throughput of the system.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0007] A power allocation-based throughput enhancement method for MIMO-URLLC short code transmission includes the following steps:

[0008] Step 1: Based on a multi-user MIMO communication scenario, the transmitted symbol vector is encoded and the channel is modeled. Additive white Gaussian noise is introduced to obtain the received signal at the user end.

[0009] Step 2: Obtain the signal-to-interference-plus-noise ratio (SINR) and achievable rate expression in the finite code length domain for each user by obtaining the received signal at the user end. Using power allocation as a variable, construct the objective function expression for the throughput of each user and the total system throughput, thereby clarifying the optimization problem for the total system throughput.

[0010] Step 3: Perform a convex approximation on the objective function constructed in Step 2 to obtain the optimized objective function, solve it, and output the final power allocation configuration.

[0011] Furthermore, the MIMO communication scenario in step 1 specifically refers to:

[0012] Transmitter: A base station equipped with a total of t antennas;

[0013] Receiver: A total of m multi-antenna users, (u1, u2, ..., u m ),and

[0014]

[0015] The number of antennas corresponding to each user is (r1, r2, ..., r m ), and the total number of antennas is r.

[0016] Furthermore, in step S1, encoding the transmitted symbol vector specifically involves:

[0017] The system is set as an FDD system, all users are considered through a single time-frequency resource block, and r ≤ t is set. At the same time, the system supports URLLC service, the length of the encoded block during transmission is n, and the maximum decoding error probability is ∈.

[0018] Assuming the antennas at the base station are a uniform linear array, and the spatial correlation coefficient between adjacent antennas at the base station is ρ0, then the transmitter correlation matrix...

[0019]

[0020] Wherein, the spatial correlation coefficient between the i-th antenna and the j-th antenna at the base station. Correlation matrix

[0021] The base station's transmit power is p, and the power allocation matrix is... Where, p i This represents the power allocated to the i-th antenna out of a total of r antennas at the receiver; the symbol vector transmitted at the transmitter is represented as... Where x i This represents the symbol vector after Gaussian codebook encoding, i.e. The superscript indicates the transpose of the matrix; the normalized precoding matrix is The signal vector Qx transmitted by the base station satisfies the power constraint.

[0022]

[0023] The superscript * indicates the conjugate transpose of the matrix;

[0024] The normalized channel vector h between the base station and the i-th receiving antenna i =(h i,1 ,h i,2 ,…,h i,t ) T Then there is

[0025]

[0026] in, E t Represents the t-dimensional identity matrix;

[0027] At this point, the channel matrix consisting of a total of r antennas from the base station to the receiver is expressed as follows:

[0028]

[0029] If we use regularized zero-forcing for precoding, then we have

[0030] Q = H(H) T H+δE r ) -1 (5)

[0031] Where δ is the adjustment coefficient, which is a constant in this model.

[0032] Further, in step S1, the signal vector at the receiving end is y = (y1, y2, ..., y...). i ,…,y r ) T The following formula can be used to give the result:

[0033] y = H T Qx+w (6)

[0034] Among them, y i Let w represent the signal vector received by the i-th antenna at the receiving end, and let w = (w1, w2, ..., w2) be additive white Gaussian noise. r ) T ,and The channel vector in equation (4) is normalized by the noise power, where σ 2 This is the normalized noise power.

[0035] Furthermore, step 2 specifically includes the following processes:

[0036] Channel modeling is based on the spatial correlation of the transmitting antenna and the channel scattering environment.

[0037] The elements in the precoded channel matrix are H T Q is matrix C, and the signal received by the k-th antenna at the receiving end is represented as:

[0038]

[0039] Where 1≤k≤r, the power allocation vector is p=(p1,p2,…,p r ) T The interference signal for the k-th antenna is The interference signal-to-noise ratio of the k-th antenna is expressed as:

[0040]

[0041] The research object was adjusted to the l-th multi-antenna user u. l Where 1≤l≤m, for this user, this communication model is equivalent to the transmitter having t antennas and the receiver having r antennas. l A single-user MIMO system with one antenna, considering r l Since ≤r≤t, the equivalent number of transmitted signal streams is r. l And for the k-th antenna at the receiving end, where k∈u l Its SINR is γ k The throughput of this single-user MIMO system is then expressed as:

[0042]

[0043] Where ∈ represents the maximum decoding error probability during transmission, n represents the FBL coding block length, and Q -1 (·) represents a function The inverse function of V l The number of streams is r l Channel dispersion at time, i.e.

[0044]

[0045] The total throughput of the entire multi-user MIMO system is expressed as the sum of the throughputs of m single-user MIMO systems, i.e.

[0046]

[0047] Furthermore, the objective function for the total system throughput is as follows:

[0048]

[0049] Further, the specific process of step 3 includes the following content:

[0050] For the objective function (12), its constraint condition is a linear inequality constraint. From the objective function and its expansion (11), it can be seen that this function is not convex with respect to its optimization variable p. The following convex approximation scheme is used to optimize the total throughput of the system:

[0051] Lemma 1: For any x>0, y>0, x0>0, y0>0, the following equation holds:

[0052]

[0053] Take

[0054]

[0055] Substituting into formula (13) gives

[0056]

[0057] where the superscript (m) represents the value or function at the m-th iteration, that is, the value or function corresponding to when the power allocation vector p takes the value of and the equality holds when p = p (m) From the above formula, it can be seen that is a concave function with respect to p. Since summation is a convexity-preserving operation, the first summation part in the objective function (11) is approximated as a concave function through a group of summations;

[0058] Process the second summation part in the expansion (11) of the objective function. This summation part is expressed as

[0059]

[0060] Combined with the convexity and concavity of the function and perform its Taylor expansion. For any 0<x≤1, 0<x0≤1, the following equation holds:

[0061]

[0062] Take

[0063]

[0064] Combined with formula (17), we have

[0065]

[0066] In formula (19), the equality holds when p = p (m) ;

[0067] Considering the function It is a convex function on x>0. By performing a similar process to the one described above, we can obtain that when γ... k When >0, there is

[0068]

[0069] In equation (20), when p = p (m) When the equality holds,

[0070] From the basic inequality a 2 +b 2 ≥2ab indicates that

[0071]

[0072] Pick

[0073]

[0074] Substituting into equation (21) yields

[0075]

[0076] Lemma 2: Let Then we have f k (x) is a convex function It is a convex function (24)

[0077] Lemma 3: Function In x i The function is convex on the interval [0, 1 ≤ i ≤ n], as can be seen from Lemmas 2 and 3. In p i The function is concave on >0, (1≤i≤r), therefore

[0078]

[0079] Among them, all inequality signs are in p = p (m) The equality sign can be obtained in all cases, and Φ l (m) (p) is a concave function, therefore for the objective function (11) we have

[0080]

[0081] In equation (26), when p = p (m) When the equality holds, it is easy to see that Ψ (m) (p) is a concave function and satisfies R(p) (m) )=Ψ (m) (p (m) );

[0082] A convex optimization objective function was obtained, namely:

[0083]

[0084] Furthermore, the objective function constructed in step 2 is approximated using a convex approximation to obtain an optimized objective function, which is then solved to output the final power allocation configuration, i.e.:

[0085] This pair of convex optimization objective functions has a global optimal solution p. * Let it be the expansion point p in the (m+1)th iteration. (m+1) ,Right now

[0086] p (m+1) =p * (28)

[0087] Then there is

[0088]

[0089] Therefore, the solution of the objective function (12) is gradually approximated by iteratively solving the convex optimization objective function (27). That is, the power allocation strategy p is continuously optimized by iterating the convex optimization objective function (27), so that it quickly converges to the solution of the objective function (12). At this time, the power allocation strategy p is the power allocation strategy that can effectively improve the total throughput of the system under this communication model.

[0090] The present invention also provides a throughput enhancement system based on power allocation for MIMO-URLLC short code transmission, comprising:

[0091] The user-end signal acquisition module, based on the multi-user MIMO communication scenario, encodes the transmitted symbol vector and models the channel, introduces additive white Gaussian noise, and obtains the received signal at the user end.

[0092] The objective function construction module is used to obtain the signal-to-interference-plus-noise ratio (SINR) of each user and the achievable rate expression in the finite code length domain from the received signals at the user end. With power allocation as a variable, it constructs an objective function expression for the throughput of each user and the total system throughput, thereby clarifying the optimization problem for the total system throughput.

[0093] The convex optimization objective function transformation module is used to expand the objective function multiple times at specific points, perform convex approximation on the objective function, and obtain the convex optimization objective function.

[0094] Furthermore, it also includes a convex optimization objective function solution module, which is used to iteratively solve the approximate convex optimization objective and output the final power allocation configuration.

[0095] Compared with the prior art, the present invention has the following beneficial effects:

[0096] 1. In a specific MU-MIMO scenario, an optimization scheme for the original non-convex optimization problem is proposed, making the optimization of the original problem feasible.

[0097] 2. Compared with traversal search, this optimization scheme can achieve almost the same optimization effect as traversal search while greatly reducing the time complexity of the algorithm.

[0098] 3. Compared with the traditional scheme for optimizing power allocation under the Shannon domain condition, this optimization scheme achieves better optimization results than the Shannon domain condition while slightly increasing the time complexity of the algorithm. Attached Figure Description

[0099] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0100] To make the technical problems, technical solutions, and beneficial effects of the embodiments of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the present invention.

[0101] It should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", and "outer" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the embodiments of the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention.

[0102] In the description of this invention, unless otherwise stated, the term "connection" should be interpreted broadly, and may refer to a fixed connection, a detachable connection, or an integral connection. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0103] The implementation process of the present invention will be further described in detail below with reference to specific accompanying drawings and examples.

[0104] Example 1

[0105] To address the shortcomings in the prior art, this embodiment provides a power allocation-based throughput improvement method for MIMO-URLLC short code transmission, which includes the following steps:

[0106] Step 1: Based on the multi-user MIMO communication scenario, the transmitted symbol vector is encoded and the channel is modeled. Additive white Gaussian noise is introduced to obtain the received signal at the user end.

[0107] In this embodiment, a mathematical model is constructed based on a multi-user MIMO communication scenario.

[0108] At the transmitting end, there is a base station equipped with a total of t antennas; at the receiving end, there are a total of m multi-antenna users, (u1, u2, ..., u...). m ),

[0109]

[0110] The number of antennas corresponding to each user is (r1, r2, ..., r m ), and the total number of antennas is r.

[0111] The system is set as an FDD system, all users are considered through a single time-frequency resource block, and r ≤ t is set. At the same time, the system supports URLLC service, the length of the encoded block during transmission is n, and the maximum decoding error probability is ∈.

[0112] Assuming the antennas at the base station are a uniform linear array, and the spatial correlation coefficient between adjacent antennas at the base station is ρ0, then the transmitter correlation matrix...

[0113]

[0114] Wherein, the spatial correlation coefficient between the i-th antenna and the j-th antenna at the base station. Correlation matrix

[0115] The base station's transmit power is p, and the power allocation matrix is... Where, p i Let represent the power allocated to the i-th antenna out of a total of r antennas at the receiving end; then the symbol vector transmitted at the transmitting end is represented as . Where x i This represents the symbol vector after Gaussian codebook encoding, i.e. The superscript indicates the transpose of the matrix; the normalized precoding matrix is The signal vector Qx transmitted by the base station satisfies the power constraint.

[0116]

[0117] The superscript * indicates the conjugate transpose of the matrix;

[0118] The normalized channel vector h between the base station and the i-th receiving antenna i =(h i,1 ,h i,2 ,…,h i,t ) T Then there is

[0119]

[0120] in, E t Represents the t-dimensional identity matrix;

[0121] At this point, the channel matrix consisting of a total of r antennas from the base station to the receiver is expressed as follows:

[0122] If we use regularized zero-forcing for precoding, then we have

[0123] Q = H(H) T H+δE r ) -1 (5)

[0124] Where δ is the adjustment coefficient, which is a constant in this model.

[0125] At this point, the signal vector at the receiving end is y = (y1, y2, ..., y...). i ,…,y r ) T The following formula can be used to give the result:

[0126] y = H T Qx+w (6)

[0127] Among them, y i Let w represent the signal vector received by the i-th antenna at the receiving end, and let w = (w1, w2, ..., w2) be additive white Gaussian noise. r ) T ,and The channel vector in equation (4) is normalized by the noise power, where σ 2 This is the normalized noise power.

[0128] Step 2: Obtain the signal-to-interference-plus-noise ratio (SINR) and achievable rate expression in the finite code length domain for each user by obtaining the received signal at the user end. Using power allocation as a variable, construct the objective function expression for the throughput of each user and the total system throughput, thereby clarifying the optimization problem for the total system throughput.

[0129] In this embodiment, the total throughput of the system under the condition of limited total power is defined, and a mathematical description of the objective function is given. The elements in the precoded channel matrix are... H T Q is matrix C. At this point, the signal received by the k-th (1≤k≤r) antenna at the receiving end is represented as...

[0130]

[0131] The power allocation vector is p = (p1, p2, ..., p r )T The interference signal for the k-th antenna is The SINR of the k-th antenna is expressed as:

[0132]

[0133] The research object is adjusted to the l-th (1≤l≤m) multi-antenna user u l For this user, this communication model is equivalent to the transmitter having t antennas and the receiver having r antennas. l A single-user MIMO system with one antenna. Considering r l Since ≤r≤t, the equivalent number of transmitted signal streams is r. l And for the k-th (k∈u) receiver l ) antennas, whose SINR is γ k The throughput of this single-user MIMO system is then expressed as:

[0134]

[0135] Where ∈ represents the maximum decoding error probability during transmission, n represents the FBL coding block length, and Q -1 (·) represents a function The inverse function of V l The number of streams is r l Channel dispersion at time, i.e.

[0136]

[0137] At this point, the total throughput of the entire multi-user MIMO system is expressed as the sum of the throughputs of m single-user MIMO systems, i.e.

[0138]

[0139] Therefore, the objective function for the total system throughput is described as follows:

[0140]

[0141] Step 3: Perform a convex approximation on the objective function constructed in Step 2 to obtain the optimized objective function, solve it, and output the final power allocation configuration.

[0142] In this embodiment, the objective function (12) is constrained by a linear inequality constraint. As can be seen from the objective function and its expansion (11), the function is not convex with respect to its optimization variable p. Therefore, we propose an effective convex approximation scheme to optimize the total throughput of the system.

[0143] Lemma 1: For any x>0, y>0, x0>0, y0>0, the following equation holds:

[0144]

[0145] In the above embodiment, the proof of Lemma 1 is as follows:

[0146] Bivariate function A Taylor expansion at the point (x0, y0) yields the following result.

[0147]

[0148] By taking the second-order partial derivatives of f(x,y) respectively, we can obtain...

[0149]

[0150] Therefore, its Hessian matrix

[0151]

[0152] Find the principal minors D of order i (i = 1, 2) respectively. i ,

[0153] When i = 1

[0154]

[0155] When i = 2

[0156]

[0157] Therefore, its Hessian matrix is ​​a positive definite matrix. n ≥0 and the original function is a convex function. Combining the Taylor expansion, we know that equation (13) holds.

[0158] Based on Lemma 1, this embodiment takes...

[0159]

[0160] Substituting into equation (13), we can obtain

[0161]

[0162] Where the superscript (m) represents the value or function at the m-th iteration, that is, when the power allocation vector p takes the value of The corresponding value or function, and p = p (m) The equality holds true. As can be seen from the above equation, Let p be a concave function. Since summation is a convexity-preserving operation, the first summation part in the objective function (11) is obtained through a set of... Summing them up approximates it as a concave function.

[0163] Process the second summation part in the objective function (11). This summation part is expressed as

[0164]

[0165] Combined with the function Based on its convexity and concavity and performing Taylor expansion on it, it is easy to know that for any 0 < x ≤ 1, 0 < x0 ≤ 1, the following formula holds:

[0166]

[0167] Take

[0168]

[0169] Combined with equation (17), we have

[0170]

[0171] In equation (19), the equal sign holds when p = p (m) at this time.

[0172] Considering that the function is a convex function when x > 0, performing a similar treatment to the above text, we can obtain that when γ k > 0, there is

[0173]

[0174] In equation (20), the equal sign holds when p = p (m) at this time.

[0175] From the basic inequality a 2 + b 2 ≥ 2ab, it can be known that

[0176]

[0177] Take

[0178]

[0179] Substitute it into equation (21) to obtain

[0180]

[0181] Next, introduce two lemmas.

[0182] Lemma 2: Let Then there is

[0183] f k (x) is a convex function is a convex function (24)

[0184] Lemma 3: The function In x i >0, and is a convex function on (1≤i≤n).

[0185] The proof of Lemma 2 is as follows:

[0186] matrix Vector b = (b1, b2, ..., b n ) T ,in

[0187]

[0188] Then there is

[0189]

[0190] That is, f k (x) can be derived from g k (x) is obtained through affine and scaling transformations. Since both affine and scaling operations are convexity-preserving operations, equation (24) holds true.

[0191] The proof of Lemma 3 is as follows:

[0192] Without loss of generality, we rearrange the components of vector x such that x k Let k be the first element in the vector. This is equivalent to setting k = 1. At this point, the original function becomes... Find its second-order partial derivatives.

[0193]

[0194] Therefore, its Hessian matrix is

[0195]

[0196] Examine the principal minors D of order i (i = 1, 2, ..., n) respectively. i ,

[0197] When i = 1

[0198]

[0199] When i = 2

[0200]

[0201] When 3 ≤ i ≤ n, it is easy to see that the matrices corresponding to each principal minor are non-full-rank matrices, hence we have

[0202] D i =0, (3≤i≤n)

[0203] Therefore, its Hessian matrix is ​​a positive semi-definite matrix, gk (x) in x i >0, and is a convex function on (1≤i≤n).

[0204] From Lemmas 2 and 3, we know that In p i The function is concave on the interval [0, 1 ≤ i ≤ r]. Therefore...

[0205]

[0206] Among them, all inequality signs are in p = p (m) The equality sign can be obtained in all cases, and It is a concave function. Therefore, for the objective function (11), we have:

[0207]

[0208] In equation (26), when p = p (m) The equality holds true. It is easy to see that Ψ (m) (p) is a concave function and satisfies R(p) (m) )=Ψ (m) (p (m) ).

[0209] Thus, we have obtained a convex optimization objective function, namely

[0210]

[0211] The convex optimization objective function has a global optimal solution p. * Let it be the expansion point p in the (m+1)th iteration. (m+1) ,Right now

[0212] p (m+1) =p * (28)

[0213] Then there is

[0214]

[0215] Therefore, the solution to the objective function (12) is gradually approximated by iteratively solving the convex optimization objective problem (27). That is, the power allocation strategy p is continuously optimized by iterating the convex optimization objective problem (27), so that it quickly converges to the solution of the objective function (12), and this method can theoretically achieve the optimal solution of the objective function (12). At this time, the power allocation strategy p is the power allocation strategy that can effectively improve the total throughput of the system under this communication model.

[0216] Example 2

[0217] In Example 2, step 3 of the above examples is explained as follows:

[0218] The specific implementation method is as follows:

[0219] 1. Let the iteration number m = 1, and give the initial power allocation p. (m) =p0;

[0220] 2. According to p (m) And R(p) determines the corresponding Ψ (m) (p);

[0221] 3. With Ψ (m) (p) is the objective function for solving a convex optimization problem, and its global optimal solution is p. * ;

[0222] 4. Let p (m+1) =p * ;

[0223] 5. If m has reached the specified number of iterations, proceed to step 6; otherwise, m = m + 1, and return to step 2.

[0224] 6. Obtain the optimized power allocation p * At this point, the system's total throughput reaches its maximum, which is R(p * ).

[0225] Example 3

[0226] Furthermore, embodiments of the present invention also provide a throughput enhancement system based on power allocation for MIMO-URLLC short code transmission, wherein the enhancement system is applied to the enhancement method described above, and the enhancement system includes:

[0227] The user-end signal acquisition module, based on the multi-user MIMO communication scenario, encodes the transmitted symbol vector and models the channel, introduces additive white Gaussian noise, and obtains the received signal at the user end.

[0228] The objective function construction module is used to obtain the signal-to-interference-plus-noise ratio (SINR) of each user and the achievable rate expression in the finite code length domain from the received signals at the user end. With power allocation as a variable, it constructs an objective function expression for the throughput of each user and the total system throughput, thereby clarifying the optimization problem for the total system throughput.

[0229] The convex optimization objective function transformation module is used to expand the objective function multiple times at specific points, perform convex approximation on the objective function, and obtain the convex optimization objective function.

[0230] The convex optimization objective function solution module is used to iteratively solve the approximate convex optimization objective and output the final power allocation configuration.

[0231] The above are merely preferred embodiments of the present invention and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should recognize that any equivalent substitutions and obvious changes made based on the content of this specification should be included within the protection scope of the present invention.

Claims

1. A method for throughput improvement based on power allocation for MIMO-URLLC short code transmission, characterized in that, The specific process comprises the following steps: Step 1: based on a multi-user MIMO communication scenario, encoding a sending symbol vector and modeling a channel, introducing additive white Gaussian noise to obtain a receiving signal at a user end; the encoding of the sending symbol vector is specifically: The setting system is an FDD system, all users are considered through a single time-frequency resource block, and At the same time, the system supports URLLC service, the code block length in the transmission process is , and the maximum decoding error probability is ; The antennas at the base station are set as a uniform linear array, where the spatial correlation coefficient of adjacent antennas at the base station is The correlation matrix of the transmitting end is （2） wherein the spatial correlation coefficient between the first root antenna and the second root antenna at the base station , the correlation matrix ; The transmit power of the base station is , the power allocation matrix , wherein , the total power received by the receiving end is , the power allocated to the th antenna; the symbol vector transmitted by the transmitting end is represented as , wherein , the symbol vector after Gaussian codebook encoding, i.e. , the superscript represents the transpose of the matrix; the normalized precoding matrix is , the signal vector transmitted by the base station satisfies the power constraint （3） Wherein, the superscript * represents a conjugate transpose of a matrix; Base station to first Normalized channel vector between receive antennas Then, there is （4） wherein , denotes the identity matrix of dimension At this time, the total number of channels from the base station to the receiving end is The channel matrix of the root antenna is expressed as ; A regularization zero-forcing method is used for pre-coding, and there is （5） wherein is a modulation coefficient; Signal vector of the receiving end This can be given by （6） wherein, represents the signal vector received by the receiving end root antenna, and additive white Gaussian noise , and The channel vector in formula (4) is normalized by noise power, where is the normalized noise power; Step 2: obtaining a signal-to-interference-and-noise ratio of each user and an achievable rate expression in a finite code length domain through the receiving signal at the user end, taking power allocation as a variable, constructing an objective function expression about a throughput of each user and a total throughput of the system, and thus determining an optimization problem for the total throughput of the system; including: Based on spatial correlation of a sending end antenna and channel scattering environment, a channel is modeled, The elements in the precoded channel matrix are , For matrix The first one located at the receiving end The signal received by the antenna is represented as follows: （7） wherein, , the power allocation vector is , the interference signal of the jthantenna is , the interference signal of the jthantenna is , the interference signal of the jthantenna is , the interference signal of the jthantenna is （8） The research object is adjusted to the multi-antenna user , For this user, the communication model is equivalent to a single-user MIMO system with root antennas at the sending end and root antennas at the receiving end, considering , so the number of streams of the transmission signal is , and for the root antennas at the receiving end, wherein The SINR is The throughput of the single-user MIMO system is represented as （9） wherein denotes the maximum decoding error probability during transmission, denotes the FBL code block length, denotes the inverse function of denotes the channel dispersion for a stream number of i.e.​ （10） The total throughput of the entire multi-user MIMO system is expressed as the sum of the throughputs of the individual single-user MIMO systems, i.e. （11） The objective function of the total throughput of the system is as follows: （12） Step 3: performing convex approximation on the objective function constructed in step 2 to obtain an optimization objective function, and solving the optimization objective function to output a final power allocation configuration; including the following contents: For the objective function (12), its constraints are linear inequality constraints, while from the objective function and its expansion (11), the function is Not being convex, the following convex approximation scheme is used to optimize the total throughput of the system: Lemma 1: For any the following holds: （13） Substituting formula (13) can obtain （14） The second summation part in the objective function expansion formula (11) is processed, and the summation part is represented as （15） where the superscript denotes the value or function at the th iteration, i.e., the value or function corresponding to the power allocation vector when it takes the value , and the equality holds when , as can be seen from the above equation, is a concave function in , and since summation is a convex preserving operation, the first summation part in the objective function (11) is approximated by a set of summations to a concave function. Taking （16） Combination function the concavity and convexity of the function f and Taylor expanding it, for any the following holds: （17） Combined with formula (17), there is （18） Taking （19） In formula (19), the equality holds when when Consider the function In is convex, when there is （20） In formula (20), the equality holds when ​ From the basic inequality it is known that （21） Substituting formula (21) can obtain （22） A convex optimization objective function is obtained, that is: （23） Lemma 2: Let , then there exists （24） Lemma 3: The function is convex on is convex on From Lemmas 2 and 3, we have In is a concave function, so （25） where all the inequalities can be taken as equalities at and is a concave function, so for the objective function (11) we have （26） In formula (26), the equality holds when It is easy to know that is a concave function, and satisfies ; The objective function constructed in step 2 is approximated to obtain an optimization objective function, and the optimization objective function is solved to output a final power allocation configuration, that is: （27） There is The pair of convex optimization objective functions has a global optimal solution as the first iteration expansion point that is （28） The MIMO communication scenario in step 1 is specifically: （29） Therefore, the solution of the objective function (12) is gradually approached by iteratively solving the convex optimization objective function (27), i.e., the power allocation strategy is constantly optimized by iteratively solving the convex optimization objective function (27) so as to quickly converge to the solution of the objective function (12), and at this time, the power allocation strategy is the power allocation strategy that can effectively improve the total system throughput under the communication model. 2.The method of claim 1, wherein, Including: Transmitting end: equipped with a total of Base stations with root antennas; Receiving end: total multiple antenna users, and （1） wherein the number of antennas corresponding to each user is , and the total number of antennas is . 3.The method of claim 1, wherein, In the step 1, the signal vector of the receiving end This can be given by: （6） wherein represents the received signal vector at the receiving end of the root antenna, and is the channel vector received by the root antenna, and the channel vector in equation (4) is normalized by the noise power, where is the normalized noise power.

4. A system for throughput improvement based on power allocation for MIMO-URLLC short code transmission, characterized in that, A receiving signal acquisition module at a user end, which encodes a sending symbol vector and models a channel based on a multi-user MIMO communication scenario, introduces additive white Gaussian noise to obtain a receiving signal at a user end; An objective function construction module, which is used to obtain a signal-to-interference-and-noise ratio of each user and an achievable rate expression in a finite code length domain through the receiving signal at the user end, take power allocation as a variable, construct an objective function expression about a throughput of each user and a total throughput of the system, and thus determine an optimization problem for the total throughput of the system; A convex optimization objective function conversion module, which is used to expand the objective function at a specific point multiple times, perform convex approximation on the objective function, and obtain a convex optimization objective function; The MIMO-URLLC short code transmission power allocation oriented throughput improvement system is used to execute the steps in the MIMO-URLLC short code transmission power allocation oriented throughput improvement method in any one of claims 1-3. Further comprising a convex optimization objective function solving module, which is used to iteratively solve the convex optimization objective obtained after approximation, and output a final power allocation configuration.

5. The MIMO-URLLC short code transmission based throughput improvement system facing power allocation according to claim 4, characterized in that, ​

Images