Transmission method of an IRS-aided OFDMA cloud access network fronthaul link

Abstract

The application discloses a kind of IRS assisted OFDMA cloud access network front link transmission methods.This method includes: the construction intelligent reflecting surface IRS assisted orthogonal frequency division multiple access OFDMA cloud access network C-RAN front link communication system, wherein baseband processing unit BBU pool is communicated with multiple users by multiple remote radio heads RRH.Access link of RRH to user uses OFDMA access technology.For BBU pool to RRH, wireless front link is used, and multiple IRSs are deployed to enhance link transmission capacity;Under the constraint of BBU pool and the transmission power of each RRH, the joint front link and access link resource configuration is configured, the problem of downlink user and rate maximization is expressed;An alternating optimization algorithm for joint optimization of front link and access link resource configuration is designed to maximize downlink user and rate.The application can significantly improve the downlink transmission performance of communication system.

Description

Technical Field

[0001] This invention relates to the field of wireless communication system transmission, and specifically to a transmission method for the fronthaul link of an IRS-assisted OFDMA cloud access network. Background Technology

[0002] With the explosive growth in the number of mobile devices and the amount of communication data, people have placed higher demands on the performance indicators of communication systems, such as latency, coverage, and spectral efficiency, posing an urgent challenge to current mobile communication systems. C-RAN is a new network architecture designed to address these challenges, facilitating centralized information processing and interference coordination. In C-RAN, compared to wired fronthaul links, wireless fronthaul links are more susceptible to weather conditions, blockages, and high path loss, requiring enhanced capacity and reliability. IRS is a low-cost, reprogrammable passive reflective array technology proposed in recent years. By adjusting IRS precoding, the electromagnetic wave propagation environment can be improved, effectively enhancing the coverage and link quality of the communication system. Introducing IRS technology into C-RAN makes it possible to enhance the transmission capabilities of wireless fronthaul links. OFDMA has long been a widely used access technology in various wireless broadband communication networks. Similarly, in C-RAN, OFDMA is an ideal multi-user broadband access transmission scheme for the access link between RRH and users. In summary, the OFDMA cloud access network with IRS-assisted fronthaul link is a cost-effective and scalable network architecture with potential to significantly improve the transmission performance of communication systems. Summary of the Invention

[0003] The purpose of this invention is to improve the transmission performance of a communication system by combining the allocation of resources for the fronthaul link and the access link. This invention deploys the IRS on the wireless fronthaul link of the OFDMA cloud access network and establishes a model of the communication system. At the same time, under the constraint of transmit power, it describes the problem of maximizing downlink users and rate, and uses an alternating optimization algorithm to jointly optimize the resource allocation to maximize downlink users and rate.

[0004] The technical solution of the present invention is as follows:

[0005] A transmission method for an IRS-assisted OFDMA cloud access network fronthaul link involves the BBU pool transmitting user data to the RRH via an IRS-assisted wireless fronthaul link. Then, all RRHs decompress their received data and transmit it to the user using OFDMA cooperative transmission. To maximize downlink users and data rate, the BBU pool transmit beamforming vector, the wireless fronthaul link compressed noise covariance matrix, the IRS phase shift matrix, the RRH sub-channel power allocation vector, and the user-sub-channel allocation vector are jointly optimized, including the following steps:

[0006] 1.1) In the downlink communication system of the OFDMA cloud access network fronthaul link assisted by IRS, with the assistance of L IRSs having R reflection units, a network equipped with N... B The BBU pool of the root antennas first sends user data to M single-antenna RRHs via the wireless fronthaul link. Then, all RRHs decompress their received data and transmit it to K single-antenna users using OFDMA cooperative transmission. The total wireless access transmission bandwidth is B. A Hz, which is equally divided into N orthogonal sub-channels; the wireless fronthaul transmission bandwidth is B. F Hz.

[0007] 1.2) The BBU pool first sends message M to user k on the nth sub-channel. k,n Encoded as baseband signal make Denotes a vector composed of all baseband signals, where k n This represents the user assigned to subchannel n. Then, the BBU pool performs linear precoding on the baseband signal vector s, resulting in the precoded signal:

[0008]

[0009] Among them, L m =diag(L m,1 ,...,L m,N () represents the precoding diagonal matrix of the BBU pool for RRH m. Definition This represents the precoded signal vector for all RRHs.

[0010] 1.3) The limited transmission rate of the wireless fronthaul link is denoted as... The BBU pool further compresses the precoded signal, generating numbers U1,...,U M Each number Corresponding to the compressed signal The compressed signal is modeled as follows:

[0011]

[0012] Among them, compression noise With signal They are independent. The compressed noise covariance matrix Ω m ≥0. The symbol “≥” indicates that the matrix is ​​positive semi-definite. This indicates that the mean is μ and the variance is σ. 2 The signal vector X follows a cyclically symmetric complex Gaussian distribution. R =[X R,1 ;...;X R,M ] is represented as:

[0013]

[0014] Wherein, the compression noise q = [q1; ...; q M The distribution is CN(0,Ω), and the covariance matrix Ω≥0.

[0015] 1.4) In order to make X R,m Transmitted to RRH m, The BBU pool will have each number Encoded as baseband signal And for the baseband signal d m Perform linear beamforming to obtain the transmitted signal:

[0016]

[0017] in, Indicates that for d m The beamforming vector, and is affected by the transmit power P of the BBU pool. B Constraints, i.e. The received signal on RRH m is:

[0018]

[0019] in, and These represent the channel gains from BBU pool to RRH m, IRS l to RRH m, and BBU pool to IRS l, respectively. Let Φ be the phase shift matrix of IRS l, and define Φ = diag(Φ1,...,Φ1). L ). n R,m This represents the received noise on RRH m.

[0020] 1.5) Define Q m =H m,B +θ H G m,B , where θ=diag(Φ H )and They represent the components of Φ H The column vector consisting of the diagonal elements and the cascaded channel between the BBU pool-IRS-RRH m. and y represents the channel gain vector from all IRSs to RRHm and from the BBU pool to all IRSs, respectively. R,m This can be further expressed as:

[0021]

[0022] Among them, Q m Fm d m The term represents the received signal that RRH m expects to decode.

[0023] 1.6) Indicates in sub-channel From RRH To users The channel gain. The signal that RRH m sends to user k, who is assigned to that channel, on subchannel n can be written as:

[0024]

[0025] Among them, X R,m,n For X R,m The nth element. The RRH transmit power constraint is set to The received signal of user k on subchannel n can be expressed as:

[0026]

[0027] in, Let represent additive white Gaussian noise at user k, assumed to be equal at all users. To ensure that all RRH signals transmitted to user k on subchannel n can coherently superimpose at the user, let ...

[0028] Furthermore, for the downlink communication system of the IRS-assisted OFDMA cloud access network fronthaul link: under the transmit power constraints of the BBU pool and each RRH, the transmit beamforming vector F of the BBU pool is jointly optimized. m Wireless fronthaul link compressed noise covariance matrix Ω, IRS phase shift matrix Φ, RRH subchannel power allocation vector and user-subchannel allocation vector ν n This maximizes the weighted sum rate for all users. The specific steps are as follows:

[0029] 2.1) Before joint optimization.

[0030] 2.1.1) Wireless fronthaul link rate C m , It is constrained as follows:

[0031] C m ≤I(y R,m ;d m ),

[0032] Wherein I(y) R,m ;d m ) represents y R,m With d m Mutual information between them, that is:

[0033]

[0034] RRH m was able to successfully recover number U m , The following conditions must be met:

[0035]

[0036] 2.1.2) Based on the received signal y k,n Decoding message M k,n The transmission rate of user k on subchannel n is:

[0037]

[0038] in, Let represent the gain vector from all RRHs to user k on subchannel n. Let This represents the power allocation vector for all RRHs on subchannel n. Let the compressed noise vector of the BBU pool be the sum of the values ​​of all RRH subchannels n, and its covariance be...

[0039] 2.1.3) Let v κ,n This indicates whether user k is assigned to subchannel n, i.e.:

[0040]

[0041] Define ν n =[v 1,n ,...,v K,n ] T ∈{0,1} K×1 This represents the allocation vector for a user on subchannel n. Radio access link transmission uses OFDMA, meaning each subchannel n can be allocated to at most one user.

[0042] 2.2) The joint optimization is expressed as:

[0043]

[0044]

[0045]

[0046]

[0047]

[0048]

[0049]

[0050]

[0051]

[0052] Ω≥0

[0053] In the above optimization problem, This represents the rate weight assigned to user k. The symbol "≥" indicates that all elements in the vector are non-negative or the matrix is ​​positive semi-definite.

[0054] 2.3) For the joint optimization described in 2.2), the problem is non-convex. The original optimization problem is transformed.

[0055] 2.3.1) Let u k,n Let represent the linear receiver on user k, and be the receiver from y. k,n The signal s was recovered. k,n The objective function in section 2.2) is transformed using the Continuous Convex Approximation (SCA) method:

[0056]

[0057] Where, λ k,n It is an auxiliary variable. The mean square error is:

[0058]

[0059] in, Let R represent the estimated signal of the linear receiver at user k on subchannel n. k,n The optimal solution is:

[0060]

[0061]

[0062] Similar to 2.3.1), the right side of the forward link constraint inequality in 2.2) is transformed into:

[0063]

[0064] Where, λ R,m It is an auxiliary variable. The mean square error is:

[0065]

[0066] in, This is the estimated signal of the linear receiver on RRH m. k,nDenotes a linear receiver on RRH m, for the receiver from y R,m The recovered signal d m The above formula I(y) R,m ;d m The optimal solution is:

[0067]

[0068]

[0069] 2.3.3) For the compressibility constraint, let Using the SCA method again, we get:

[0070]

[0071] The above expression retains the equality if and only if:

[0072]

[0073] This holds true. The compressibility constraint in 2.2) is transformed into:

[0074]

[0075] 2.3.4) In each iteration, according to 2.3.1) to 2.3.3), first update the auxiliary variables respectively. and The new optimization problem is as follows:

[0076]

[0077]

[0078]

[0079]

[0080]

[0081]

[0082]

[0083]

[0084]

[0085] Ω≥0

[0086] 2.4) For the new optimization problem in 2.3.4), it is broken down into three subproblems and solved alternately.

[0087] The specific steps are as follows:

[0088] 2.4.1) Subproblem 1: Fixed F m Φ and Ω, optimize the power allocation vector of RRH and user-subchannel allocation vector ν n The first subproblem is expressed as:

[0089]

[0090]

[0091]

[0092]

[0093]

[0094]

[0095]

[0096] Ω≥0

[0097] Among them, a m,n and Ω m,n Each is a matrix and Ω m The diagonal elements.

[0098] make and These are the dual variables of the compression constraint and the RRH power constraint, respectively. Define α = [α1; ...; α...]. M ],β=[β1;...;β M Therefore, the Lagrange dual function is expressed as:

[0099]

[0100]

[0101]

[0102]

[0103]

[0104]

[0105] Ω≥0

[0106] Wherein, objective function for:

[0107]

[0108] in, for:

[0109]

[0110] 2.4.2) Divide the dual function into N subproblems, each with the same structure as follows:

[0111]

[0112]

[0113] 1 T ν n ≤1

[0114]

[0115] For users Assigned to sub-channel The situation, namely and The above problem can then be simplified to:

[0116]

[0117] The optimal RRH power allocation vector for this problem for:

[0118]

[0119] in Symbol [x] + This means taking the larger of x and 0.

[0120] 2.4.3) For the optimization problem in 2.4.1), the dual problem is as follows:

[0121]

[0122] The problem is convex, and the optimal dual variable α can be obtained using the ellipsoid method. * and β * .

[0123] 2.5) Subproblem 2: With the optimization variable v fixed n , and F m Next, optimize the compression noise covariance matrix Ω and the phase shift matrix Φ of the IRS in the fronthaul link. The second subproblem can be written as:

[0124]

[0125]

[0126]

[0127]

[0128]

[0129] Ω≥0

[0130] According to 2.3.2), in the constraints of the forward link... The item expands to:

[0131]

[0132] definition:

[0133]

[0134]

[0135]

[0136] The forward link constraint can be written as B F [J m +θ H f m K m θ-2Re{θ H E m}]+C m ≤0, make and Therefore, the forward link constraint condition can be further written as:

[0137]

[0138] Subproblem 2 is transformed into:

[0139]

[0140]

[0141]

[0142]

[0143]

[0144]

[0145] Ω≥0

[0146] in, express The nth diagonal element. In the above problem, the rank-1 constraint is non-convex. Applying the positive semi-definite relaxation (SDR) method to ignore the rank-1 constraint yields a convex positive semi-definite problem. Then, the CVX convex optimization tool is used to solve it. If the optimal value is obtained... If the rank is not 1, then randomization techniques are used to obtain a feasible suboptimal solution.

[0147] 2.6) Subproblem 3: When the optimization variable ν n , Optimize the transmit beamforming vector F of the BBU pool when Φ and Ω are both given. m And the covariance matrix Ω of the fronthaul link compression noise. Since the objective function of the optimization problem in 2.3.4) does not include F... m Therefore, Ω is optimized again. Introducing information about Ω... m scaling factor η m , The third subproblem can be expressed as:

[0148]

[0149]

[0150]

[0151]

[0152]

[0153]

[0154] For subproblem three, the convex optimization tool CVX is used to find the optimal solution.

[0155] 2.7) Repeat steps 2.3.4) to 2.6) to perform alternating optimization until convergence.

[0156] The beneficial effects of this invention are as follows: By jointly optimizing the resource configuration of the wireless fronthaul link and the access link, the transmission performance of the communication system is improved. This invention constructs a communication system for the fronthaul link of the IRS-assisted OFDMA cloud access network and designs an alternating optimization algorithm for jointly optimizing resource configuration to maximize downlink users and rates. Attached Figure Description

[0157] Figure 1 This is a model diagram of the downlink communication system of the fronthaul link of the IRS-assisted OFDMA cloud access network of the present invention;

[0158] Figure 2This is a graph showing the relationship between the downlink users and rate of the joint optimization scheme of the present invention and other benchmark schemes as a function of the number of reflection units R per IRS. Detailed Implementation

[0159] The present invention will be further described below with reference to the accompanying drawings.

[0160] This invention deploys the IRS on the wireless fronthaul link of the OFDMA cloud access network, which can effectively improve the capacity of the wireless fronthaul link, meet the high data volume requirements between the BBU pool and RRH, and improve the transmission performance of the C-RAN system.

[0161] A transmission method for an IRS-assisted OFDMA cloud access network fronthaul link involves the BBU pool transmitting user data to the RRH via an IRS-assisted wireless fronthaul link. Then, all RRHs decompress their received data and transmit it to the user using OFDMA cooperative transmission. To maximize downlink users and data rate, the BBU pool transmit beamforming vector, the wireless fronthaul link compressed noise covariance matrix, the IRS phase shift matrix, the RRH sub-channel power allocation vector, and the user-sub-channel allocation vector are jointly optimized, including the following steps:

[0162] 1.1) In the downlink communication system of the OFDMA cloud access network fronthaul link assisted by IRS, with the assistance of L IRSs having R reflection units, a network equipped with N... B The BBU pool of the root antennas first sends user data to M single-antenna RRHs via the wireless fronthaul link. Then, all RRHs decompress their received data and transmit it to K single-antenna users using OFDMA cooperative transmission. The total wireless access transmission bandwidth is B. A Hz, which is equally divided into N orthogonal sub-channels; the wireless fronthaul transmission bandwidth is B. F Hz.

[0163] 1.2) The BBU pool first sends message M to user k on the nth sub-channel. k,n Encoded as baseband signal make Denotes a vector composed of all baseband signals, where k n This represents the user assigned to subchannel n. Then, the BBU pool performs linear precoding on the baseband signal vector s, resulting in the precoded signal:

[0164]

[0165] Among them, L m =diag(L m,1 ,...,L m,N () represents the precoding diagonal matrix of the BBU pool for RRH m. Definition This represents the precoded signal vector for all RRHs.

[0166] 1.3) The limited transmission rate of the wireless fronthaul link is denoted as... The BBU pool further compresses the precoded signal, generating numbers U1,...,U M Each number Corresponding to the compressed signal The compressed signal is modeled as follows:

[0167]

[0168] Among them, compression noise With signal They are independent. The compressed noise covariance matrix Ω m ≥0. The symbol “≥” indicates that the matrix is ​​positive semi-definite. This indicates that the mean is μ and the variance is σ. 2 The signal vector X follows a cyclically symmetric complex Gaussian distribution. R =[X R,1 ;…;X R,M ] is represented as:

[0169]

[0170] Wherein, the compression noise q = [q1; ...; q M Distribution is And the covariance matrix Ω ≥ 0.

[0171] 1.4) In order to make X R,m Transmitted to RRH m, The BBU pool will have each numbered U m , Encoded as baseband signal And for the baseband signal d m Perform linear beamforming to obtain the transmitted signal:

[0172]

[0173] in, Indicates that for d m The beamforming vector, and is affected by the transmit power P of the BBU pool. B Constraints, i.e. The received signal on RRH m is:

[0174]

[0175] in, and These represent the channel gains from BBU pool to RRH m, IRS l to RRH m, and BBU pool to IRS l, respectively. Let Φ be the phase shift matrix of IRS l, and define Φ = diag(Φ1,...,Φ1). L ). n R,m This represents the received noise on RRH m.

[0176] 1.5) Define Q m =H m,B +θ H G m,B , where θ=diag(Φ H )and They represent the components of Φ H The column vector consisting of the diagonal elements and the cascaded channel between the BBU pool-IRS-RRH m. and Let y represent the channel gain vectors from all IRSs to RRH m and from the BBU pool to all IRSs, respectively. R,m This can be further expressed as:

[0177]

[0178] Among them, Q m F m d m The term represents the received signal that RRH m expects to decode.

[0179] 1.6) Indicates in sub-channel From RRH To users The channel gain. The signal that RRH m sends to user k, who is assigned to that channel, on subchannel n can be written as:

[0180]

[0181] Among them, X R,m,n For X R,m The nth element. The RRH transmit power constraint is set to The received signal of user k on subchannel n can be expressed as:

[0182]

[0183] in, Let represent additive white Gaussian noise at user k, assumed to be equal at all users. To ensure that all RRH signals transmitted to user k on subchannel n can coherently superimpose at the user, let ...

[0184] Among them, the downlink communication system of the IRS-assisted OFDMA cloud access network fronthaul link: under the transmit power constraints of the BBU pool and each RRH, jointly optimize the BBU pool transmit beamforming vector F. m Wireless fronthaul link compressed noise covariance matrix Ω, IRS phase shift matrix Φ, RRH subchannel power allocation vector and user-subchannel allocation vector ν n This maximizes the weighted sum rate for all users. The specific steps are as follows:

[0185] 2.1) Before joint optimization,

[0186] 2.1.1) Wireless fronthaul link rate C m , It is constrained as follows:

[0187] C m ≤I(y R,m ;d m )

[0188] Wherein I(y) R,m ;d m ) represents y R,m With d m Mutual information between them, that is:

[0189]

[0190] RRH m was able to successfully recover number U m , The following conditions must be met:

[0191]

[0192] 2.1.2) Based on the received signal y k,n Decoding message M k,n The transmission rate of user k on subchannel n is:

[0193]

[0194] in, Let represent the gain vector from all RRHs to user k on subchannel n. Let This represents the power allocation vector for all RRHs on subchannel n. Let the compressed noise vector of the BBU pool be the sum of the values ​​of all RRH subchannels n, and its covariance be...

[0195] 2.1.3) Let v κ,n This indicates whether user k is assigned to subchannel n, i.e.:

[0196]

[0197] Define ν n =[v 1,n ,...,v K,n ] T ∈{0,1} K×1 This represents the allocation vector for a user on subchannel n. Radio access link transmission uses OFDMA, meaning each subchannel n can be allocated to at most one user.

[0198] 2.2) The joint optimization is expressed as:

[0199]

[0200]

[0201]

[0202]

[0203]

[0204]

[0205]

[0206]

[0207]

[0208] Ω≥0

[0209] In the above optimization problem, This represents the rate weight assigned to user k. The symbol "≥" indicates that all elements in the vector are non-negative or the matrix is ​​positive semi-definite.

[0210] 2.3) For the joint optimization described in 2.2), the problem is non-convex. The original optimization problem is transformed.

[0211] 2.3.1) Let u k,n Let represent the linear receiver on user k, and be the receiver from y. k,n The signal s was recovered. k,n The objective function in 2.2) is transformed into the following by using the continuous convex approximation SCA method:

[0212]

[0213] Where, λ k,n It is an auxiliary variable. The mean square error is:

[0214]

[0215] in, Let R represent the estimated signal of the linear receiver at user k on subchannel n. k,n The optimal solution is:

[0216]

[0217]

[0218] Similar to 2.3.1), the right side of the forward link constraint inequality in 2.2) is transformed into:

[0219]

[0220] Where, λ R,m It is an auxiliary variable. The mean square error is:

[0221]

[0222] in, This is the estimated signal of the linear receiver on RRH m. k,n Denotes a linear receiver on RRH m, for the receiver from y R,m The recovered signal d m The above formula I(y) R,m ;d m The optimal solution is:

[0223]

[0224]

[0225] 2.3.3) For the compressibility constraint, let Using the SCA method again, we get:

[0226]

[0227] The above expression retains the equality if and only if:

[0228]

[0229] This holds true. The compressibility constraint in 2.2) is transformed into:

[0230]

[0231] 2.3.4) In each iteration, according to 2.3.1) to 2.3.3), first update the auxiliary variables respectively. and The new optimization problem is as follows:

[0232]

[0233]

[0234]

[0235]

[0236]

[0237]

[0238]

[0239]

[0240]

[0241] Ω≥0

[0242] 2.4) For the new optimization problem in 2.3.4), it is broken down into three subproblems and solved alternately.

[0243] The specific steps are as follows:

[0244] 2.4.1) Subproblem 1: Fixed F m Φ and Ω, optimize the power allocation vector of RRH and user-subchannel allocation vector ν n The first subproblem is expressed as:

[0245]

[0246]

[0247]

[0248]

[0249]

[0250]

[0251]

[0252] Ω≥0

[0253] Among them, a m,n and Ω m,n Each is a matrix and Ω m The diagonal elements.

[0254] make and These are the dual variables of the compression constraint and the RRH power constraint, respectively. Define α = [α1; ...; α...]. M ],β=[β1;...;β M Therefore, the Lagrange dual function is expressed as:

[0255]

[0256]

[0257]

[0258]

[0259]

[0260]

[0261] Ω≥0

[0262] Wherein, objective function for:

[0263]

[0264] in, for:

[0265]

[0266] 2.4.2) Divide the dual function into N subproblems, each with the same structure as follows:

[0267]

[0268]

[0269] 1 T ν n ≤1

[0270]

[0271] For users Assigned to sub-channel The situation, namely and The above problem can then be simplified to:

[0272]

[0273] The optimal RRH power allocation vector for this problem for:

[0274]

[0275] in Symbol [x] + This means taking the larger of x and 0.

[0276] 2.4.3) For the optimization problem in 2.4.1), the dual problem is as follows:

[0277]

[0278] The problem is convex, and the optimal dual variable α can be obtained using the ellipsoid method. * and β * .

[0279] 2.5) Subproblem 2: With the optimization variable v fixed n , and F m Next, optimize the compression noise covariance matrix Ω and the phase shift matrix Φ of the IRS in the fronthaul link. The second subproblem can be written as:

[0280]

[0281]

[0282]

[0283]

[0284]

[0285] Ω≥0

[0286] According to 2.3.2), in the constraints of the forward link... The item expands to:

[0287]

[0288] definition:

[0289]

[0290] The forward link constraint can be written as B F [J m +θ H f m K m θ-2Re{θ H E m}]+C m≤0, make and Therefore, the forward link constraint condition can be further written as:

[0291]

[0292] Subproblem 2 is transformed into:

[0293]

[0294]

[0295]

[0296]

[0297]

[0298]

[0299] Ω≥0

[0300] in, express The nth diagonal element. In the above problem, the rank-1 constraint is non-convex. Applying the positive semi-definite relaxation SDR method to ignore the rank-1 constraint yields a convex positive semi-definite problem. Then, the CVX convex optimization tool is used to solve it. If the optimal value is obtained... If the rank is not 1, then randomization techniques are used to obtain a feasible suboptimal solution.

[0301] 2.6) Subproblem 3: When the optimization variable ν n , Optimize the transmit beamforming vector F of the BBU pool when Φ and Ω are both given. m And the covariance matrix Ω of the fronthaul link compression noise. Since the objective function of the optimization problem in 2.3.4) does not include F... m Therefore, Ω is optimized again. Introducing information about Ω... m scaling factor η m , The third subproblem can be expressed as:

[0302]

[0303]

[0304]

[0305]

[0306]

[0307]

[0308] For subproblem three, the convex optimization tool CVX is used to find the optimal solution.

[0309] 2.7) Repeat steps 2.3.4) to 2.6) to perform alternating optimization until convergence.

Claims

1. A transmission method for the fronthaul link of an IRS-assisted OFDMA cloud access network, characterized in that: The BBU pool first sends user data to the RRH via the IRS-assisted wireless fronthaul link; then, all RRHs decompress their received data and transmit it to the user using OFDMA cooperative transmission; to maximize downlink users and rate, the BBU pool transmit beamforming vector, the wireless fronthaul link compressed noise covariance matrix, the IRS phase shift matrix, the RRH subchannel power allocation vector, and the user-subchannel allocation vector are jointly optimized, including the following steps: 1.1) In the downlink communication system of the OFDMA cloud access network fronthaul link assisted by IRS, with the assistance of L IRSs having R reflection units, a network equipped with N... B The BBU pool of the root antennas first sends user data to M single-antenna RRHs via the wireless fronthaul link; then, all RRHs decompress their received data and transmit it to K single-antenna users using OFDMA cooperative transmission; the total wireless access transmission bandwidth is B. A Hz, which is equally divided into N orthogonal sub-channels; the wireless fronthaul transmission bandwidth is B. F Hz; 1.2) The BBU pool first sends message M to user k on the nth sub-channel. k,n Encoded as baseband signal ;make Denotes a vector composed of all baseband signals, where , This represents the user assigned to subchannel n; then, the BBU pool performs linear precoding on the baseband signal vector s to obtain the precoded signal: ； in, Represents the precoding diagonal matrix of the BBU pool for RRH m; Define This represents the precoded signal vector for all RRHs; 1.3) The limited transmission rate of the wireless fronthaul link is denoted as... The BBU pool further compresses the precoded signal to generate a number. Each number Corresponding to the compressed signal The compressed signal is modeled as follows: ； Among them, compression noise With signal Independent of each other, compress the noise covariance matrix ;symbol This indicates that the matrix is ​​positive semi-definite. The mean is The variance is Cyclic symmetric complex Gaussian distribution; compressed signal vector Represented as: ； Among them, compression noise Distribution And the covariance matrix ; 1.4) In order to Transmitted to RRH m, The BBU pool will assign each number Encoded as baseband signal and the baseband signal d m Perform linear beamforming to obtain the transmitted signal: ； in, Indicates that for d m The beamforming vector, and is affected by the transmit power P of the BBU pool. B Constraints, i.e. The received signal on RRH m is: ； in, , and These represent the channel gains from the BBU pool to RRH m, from IRS l to RRH m, and from the BBU pool to IRS l, respectively. Let IRS l be the phase shift matrix, and define... , This represents the received noise on RRH m. ; 1.5) Definition ,in and They represent respectively by The column vector composed of the diagonal elements and the cascaded channel between the BBU pool-IRS-RRH m, and These represent the channel gain vectors from all IRSs to RRH m and from the BBU pool to all IRSs, respectively. This can be further expressed as: ； in, The term represents the received signal that RRH m expects to decode; 1.6) Indicates in sub-channel From RRH To users The channel gain; the signal RRH m transmitted on subchannel n to user k assigned to that channel is written as: ； in, for The nth element, the transmit power constraint of RRH is set to The received signal of user k on subchannel n is expressed as: ； in, Let the additive white Gaussian noise at user k be equal at all users; in order for all RRH signals transmitted to user k on subchannel n to be coherently superimposed at the user, let ; The downlink communication system of the IRS-assisted OFDMA cloud access network fronthaul link: under the transmit power constraints of the BBU pool and each RRH, jointly optimize the transmit beamforming vector of the BBU pool. Wireless fronthaul link compressed noise covariance matrix IRS phase shift matrix RRH subchannel power allocation vector and user-subchannel allocation vector This maximizes the weighted sum rate for all users; The specific steps are as follows: 2.1) Before joint optimization; 2.1.1) Wireless fronthaul link rate C m , It is constrained as follows: ； in, express and Mutual information between them, that is: ； RRH m was able to successfully recover number U m , The following conditions must be met: ； 2.1.2) Based on the received signal Decoding messages The transmission rate of user k on subchannel n is: ； in, Let represent the gain vector from all RRHs to user k on subchannel n; let This represents the power allocation vector for all RRHs on subchannel n. Let the compressed noise vector of the BBU pool be the sum of the values ​​of all RRH subchannels n, and its covariance be... ; 2.1.3) Order This indicates whether user k is assigned to subchannel n, i.e.: ； definition This represents the allocation vector for a user on subchannel n. The radio access link uses OFDMA for transmission, meaning each subchannel n can be allocated to at most one user. ; 2.2) The joint optimization is expressed as: ； s.t. ; ； ； ； ； ； ； ； ； In the above optimization problem, This represents the rate weight assigned to user k, symbol [symbol missing]. This indicates that the matrix is ​​positive semi-definite; 2.3) For the joint optimization described in 2.2), the problem is non-convex, so the original optimization problem is transformed; 2.3.1) Let Let K denote the linear receiver on user k, which is from Signal recovered The objective function in section 2.2) is transformed into the following using the continuous convex approximation SCA method: ； in, It is an auxiliary variable. The mean square error is: ； in, The above equation represents the estimated signal of the linear receiver on subchannel n at user k; The optimal solution is: ； ； Similar to 2.3.1), the right-hand side of the forward link constraint inequality in 2.2) is transformed into: ； in, It is an auxiliary variable. The mean square error is: ； in, It is the estimated signal of the linear receiver on RRH m. Describes a linear receiver on RRH m, for from Signal recovered The above formula The optimal solution is: ； ； 2.3.3) For the compressibility constraint, let Using the SCA method again, we get: ； The above expression retains the equality if and only if: ； When it is true; the compressibility constraint condition in 2.2) is transformed into: ； 2.3.4) In each iteration, according to 2.3.1)~2.3.3), first update the auxiliary variables respectively. and The new optimization problem is as follows: ； s.t. ; ； ； ； ； ； ； ； ； 2.4) For the new optimization problem in 2.3.4), it is broken down into three sub-problems and solved alternately; the specific steps are as follows: 2.4.1) Subproblem 1: Fixed F m , and Optimize the power allocation vector of RRH and user-subchannel allocation vector The first subproblem is expressed as: ； s.t. ; ； ； ； ； ； ； in, and Each is a matrix and The diagonal elements; make and Let be the dual variables of the compression constraint and the RRH power constraint, respectively. Therefore, the Lagrange dual function is expressed as: ； s.t. ; ； ； ； ； ； Wherein, objective function for: ； in, for: ； 2.4.2) Divide the dual function into N subproblems, each with the same structure as follows: ； s.t. ; ； ； For users Assigned to sub-channel The situation, namely and The above problem can then be simplified to: The optimal RRH power allocation vector for this problem for: ； in , , symbol [x] + This means taking the larger of x and 0; 2.4.3) For the optimization problem in 2.4.1), the dual problem is as follows: ； The problem is convex, and the optimal dual variable can be obtained using the ellipsoid method. and ; 2.5) Subproblem 2: With fixed optimization variables , and Below, optimize the compression noise covariance matrix of the fronthaul link. Phase shift matrix of IRS The second subproblem is written as: ； s.t. ; ； ； ； ； According to 2.3.2), in the constraints of the forward link... The item expands to: ； definition: ； The forward link constraint is written as ,make and Therefore, the forward link constraint condition can be further written as: ； Subproblem 2 is transformed into: ； s.t. ; ； ； ； ； ， ； in, express Given the nth diagonal element, the rank-1 constraint in the above problem is non-convex. Applying the positive semi-definite relaxation SDR method to ignore the rank-1 constraint yields a convex positive semi-definite problem. Then, the CVX convex optimization tool is used to solve it. If the optimal value is obtained... If the rank is not 1, then randomization techniques are used to obtain a feasible suboptimal solution; 2.6) Sub-problem 3: When optimizing variables , , and Given all parameters, optimize the transmit beamforming vector F of the BBU pool. m and fronthaul link compression noise covariance matrix Since the objective function of the optimization problem in 2.3.4) does not contain F... m Therefore, further optimization is needed. ; Introducing information about scaling factor The third subproblem can be expressed as: ； s.t. ; ； ； ； ； For subproblem three, the optimal solution is obtained using the convex optimization tool CVX; 2.7) Repeat steps 2.3.4) to 2.6) to perform alternating optimization until convergence.

Images