CN113364501A - Power control method based on low-precision ADC (analog to digital converter) de-cellular large-scale MIMO (multiple input multiple output) system under Rice channel - Google Patents

Abstract

The present invention provides a power control method for a cellular massive MIMO system based on a low-precision ADC under a Rice channel. S3, based on the constructed model, perform downlink data transmission; S4, analyze the user's downlink reachable rate; S5, according to the downlink reachable rate , construct an optimization problem of maximizing the total rate of the system; S6, design a power control strategy according to the optimization problem. The technical scheme of the present invention further improves the performance based on low precision under the Rice channel under the premise of ensuring that the service quality of each user is not less than the set minimum reachable rate and the transmit power of each AP is not greater than the set maximum transmit power. ADC to cellular massive MIMO system total rate.

Description

A power control method for decellularized massive MIMO system based on low-precision ADC in Rice channel

technical field

The present invention relates to the technical field of wireless communication, in particular, to a power control method for a cellular massive MIMO system based on a low-precision ADC in a Rice channel.

Background technique

In recent years, with the rapid development of the mobile communication industry, people's demand for data traffic is increasing day by day. In order to meet people's diverse communication needs, the decellularized massive multiple-input multiple-output (MIMO) system, which completely changes the cellular network architecture, has gradually entered people's field of vision, and has become a popular choice in the Beyond 5G and 6G era. One of the core technologies to keep an eye on. The decellularized massive MIMO system is based on distributed massive MIMO, which breaks the division of cell boundaries, that is, a large number of access points (APs) and users are distributed in a wide area at the same time, and different APs pass through the backhaul chain. The road is connected to the central processing unit (CPU), and under the same time-frequency resources, it provides services to all users following the instructions of the CPU. Since the decellularized massive MIMO system greatly reduces the distance between users and APs, it has strong spatial macro-diversity gain and the ability to resist path loss, and can greatly improve the quality of service for edge users.

However, decellularized massive MIMO systems usually need to deploy a large number of APs and users. When APs and users are equipped with full-precision analog-to-digital converters (ADCs) to quantize the received information, it will inevitably bring high hardware costs and high cost. Huge energy consumption. In order to deal with this problem, it is undoubtedly a relatively straightforward solution to configure low-precision ADCs at APs and users to perform quantization processing. In addition, in many future communication scenarios, in order to cope with the increasingly tight spectrum resources, the decellularized massive MIMO system is likely to operate in the millimeter wave frequency band. Due to the short wavelength and strong directivity of millimeter waves, the line-of-sight component dominates the entire channel. Therefore, the research on low-precision ADC-based cellular massive MIMO systems under Rice fading channels has become a current academic topic. research hotspots.

For decellularized massive MIMO systems based on low-precision ADCs in Rice fading channels, the existing research only uses the maximum-minimum power control method to improve the achievable rate of the worst-performing user, in order to make the service quality of all users equal. . For modern communication systems, the total rate of the system is also a very important performance parameter, but the maximum-minimum power control method only focuses on improving the quality of service for the worst user, while still improving the total rate of the system. lack.

SUMMARY OF THE INVENTION

According to the technical problem raised above, the present invention provides a power control method for a cellular massive MIMO system based on a low-precision ADC in a Rice channel. Under the premise of ensuring that the service quality of each user is not less than the set minimum reachable rate and the transmit power of each AP is not greater than the set maximum transmit power, the present invention further improves the low-precision ADC-based decellularization under the Rice channel. Massive MIMO system total rate.

The technical means adopted in the present invention are as follows:

A power control method for a cellular massive MIMO system based on a low-precision ADC under a Rice channel, comprising the following steps:

S1. Establish a model of a cellular massive MIMO system based on a low-precision ADC under the Rice channel;

S2, based on the constructed model, perform uplink pilot training;

S3, based on the constructed model, perform downlink data transmission;

S4, analyze the user's downlink reachable rate;

S5. According to the downlink reachable rate, construct an optimization problem to maximize the total rate of the system;

S6. Design a power control strategy according to the optimization problem.

Further, the model constructed in the step S1 is specifically expressed as follows:

和散射分量h_mn～CN(0,1)共同组成，其中θ_mn～[-π，π]表示到达角且CN(0,1)表示均值为0和方差为1的循环对称复高斯变量；另外，k_mn表示莱斯K-因子，代表视距分量与散射分量之间的比值，定义

和

其中β_mn＝γ_mn/(1+k_mn)，因此，信道系数被重新表示为

并且服从

的统计分布。Among them, _{gmn represents the channel coefficient between the mth access point AP and the nth user; γmn represents} the large-scale fading coefficient, which is used to reflect the influence of path loss and shadow fading on the channel coefficient; the part in brackets represents small-scale fading coefficient, determined by the line-of-sight component

and the scattering components h _mn ～CN(0,1), where θ _mn ～[-π, π] represents the angle of arrival and CN(0,1) represents a cyclic symmetric complex Gaussian variable with mean 0 and variance 1; In addition, k _mn represents the Rice K-factor, which represents the ratio between the line-of-sight component and the scattering component, defined

and

where β _mn =γ _mn /(1+ _{km mn} ), therefore, the channel coefficients are re-expressed as

and obey

statistical distribution.

Further, the specific implementation process of the step S2 is as follows:

S21. Let the length of the correlation interval be T, and let the time occupied by pilot training in each correlation interval be τ _p (τ _p ≥N);

并且假定不同用户间发送的导频序列是相互正交的，即

则第m个接入点AP接收到的上行导频信息表示为：S22. Assume that in the uplink pilot training phase, the nth user can send uplink pilot sequences to all access points APs

And it is assumed that the pilot sequences sent by different users are mutually orthogonal, that is,

Then the uplink pilot information received by the mth access point AP is expressed as:

表示第m个AP接收到的加性高斯白噪声；where _pp represents the normalized signal-to-noise ratio of the uplink pilot signal,

represents the additive white Gaussian noise received by the mth AP;

S23. Use the additive quantization noise model to model the quantization process, then the uplink pilot information received by the mth access point AP is re-expressed as:

的近似关系，

表示与接收信号y_m,p无关的加性量化噪声；Among them, α _m represents the ADC accuracy at the m-th access point AP, and the accuracy is related to the quantization bit number ρ _m used by the m-th access point AP, when the quantization bit number ρ _m =1,2,3, 4 and 5, α _m is precisely given as α _m =0.6366, 0.8825, 0.96546, 0.990503, 0.997501; when ρ _m >5, the relationship between α _m and ρ _m satisfies

approximation relationship,

represents the additive quantization noise independent of the received signal y _m,p ;

的协方差，表示为：S24. Calculate additive quantization noise

The covariance of , expressed as:

与

相乘，得到：S25. In order for the mth access point AP to obtain the channel state information from the nth user, set the

and

Multiply to get:

S26. Since the access point AP can obtain the line-of-sight component information and large-scale fading coefficients from different users, in the channel estimation stage, the system can eliminate the influence of the line-of-sight component, that is, equation (5) can be re-expressed as:

表示消除视距分量影响后余留下的量化噪声；in,

Represents the quantization noise remaining after removing the influence of the line-of-sight component;

的协方差，表示为：S27. Calculate quantization noise

The covariance of , expressed as:

in,

S28. Use the minimum mean square error channel estimation scheme for formula (6) to obtain the channel estimation information of g _mn , as follows:

其中

S29. According to the properties of the MMSE channel estimation scheme, obtain

in

Further, the specific implementation process of the step S3 is as follows:

S31. After the access point AP obtains the CSI from the user, it uses the conjugate transpose method to precode the downlink transmission signal. Therefore, the transmission signal of the mth access point AP is expressed as:

的条件；p_d表示下行发送数据信号的归一化SNR；η_mn表示功率控制系数；Among them, x _n represents the data signal sent to the nth user, satisfying

conditions; p _d represents the normalized SNR of the downlink transmission data signal; η _mn represents the power control coefficient;

的功率约束条件，即：S32. Adjust the power control coefficient η _mn , so that each access point AP satisfies the

The power constraints of , namely:

S33. According to formula (9), after all access points APs have sent downlink data signals, the received signal of the nth user is expressed as:

Among them, w _n ~CN(0,1) represents the additive white Gaussian noise received by the nth user;

S34. Based on the low-precision ADC used by the user terminal, the received signal of the nth user after quantization processing is expressed as:

表示与接收信号r_n不相关的加性量化噪声；Among them, μ _n represents the ADC accuracy of the nth user, which is related to the number of quantization bits σn used by the _nth user,

represents the additive quantization noise uncorrelated with the received signal _rn ;

的协方差，计算公式如下：S35. Calculate additive quantization noise

The covariance is calculated as follows:

Further, the specific implementation process of the step S4 is as follows:

S41. According to formula (12), the received signal of the nth user is re-expressed in the following form:

in,

S42. According to formula (14), the expression of the downlink reachable rate of the nth user is obtained as follows:

S43. According to formula (15), deduce the closed expression of the reachable rate of the nth user as:

的第n列；

和

分别表示

和

的第(n-1)×N+n₁行；矩阵Q、U和V中的对应元素被给定为[Q]_mn＝η_mn，

δ(m,n)表示狄拉克δ函数，即当m＝n时，δ(m,n)＝1；当m≠n时，δ(m,n)＝0。Among them, q _n represents

the nth column of ;

and

Respectively

and

The (n- ₁ )×N+n1th row of ; the corresponding elements in matrices Q, U and V are given as [Q] _mn = _ηmn ,

δ(m,n) represents the Dirac delta function, that is, when m=n, δ(m,n)=1; when m≠n, δ(m,n)=0.

Further, the specific implementation process of the step S5 is as follows:

和每个AP的发送功率不大于设定的最大发送功率p_d的前提下，将最大化系统总速率的优化问题归纳成如下形式：S51. After ensuring that the service quality of each user is not less than the set minimum reachable rate

On the premise that the transmit power of each AP is not greater than the set maximum transmit power p _d , the optimization problem of maximizing the total rate of the system is summarized into the following form:

S52. Perform mathematical transformation _on the constraints (17b) in the optimization problem P1, and you can obtain:

Among them, the constraint (17b) is equivalent to the second-order conical formula in formula (18), that is, formula (17b) is converted into the convex constraint in formula (18); formulas (17c) and (17d) are both: Convex constraints;

S53. Using the method of continuous convex approximation, the non-convex optimization problem P ₁ is approximately equivalent to the form of the optimization problem P ₂ , and the specific form of the optimization problem P ₂ is:

where l=[l ₁ ,...,l _N ] and t=[t ₁ ,...,t _N ];

S54. For n=1,...,N, the given constraints are as follows:

Among them, the objective function (19a) in the optimization problem P2 is convex, but the constraint ( ₂₀ ) is non-convex;

S55. Convert the constraint (20) into the form of a convex function. For n=1, . . . , N, the constraint (20) is equivalent to:

S56. When n=1,...,N, formula (21) is re-expressed in the following form:

分别表示Q和l_n经过k次SCA处理后的迭代值，基于此，公式(20)已经通过公式(23)成功地转换成了二阶锥形式的约束条件，另外，通过观察优化问题P₂，得到约束条件(19c)是凸的；where ^Qk and

respectively represent the iterative values of Q and _ln after k times of SCA processing. Based on this, formula (20) has been successfully transformed into a second-order conical constraint by formula (23). In addition, by observing the optimization problem P ₂ , the constraint (19c) is convex;

S57, according to the method of continuous convex approximation, use the first-order Taylor expansion in formula (24) to approximate formula (22);

S58. According to the method of continuous convex approximation, and using the property of the inequality ln(x)≥1-x ^-1 , convert the constraint condition (19c) into the following constraint condition with a second-order conical formula:

S59. According to formula (23) and formula (25), at the k-th iteration, express the optimization problem P ₂ as a second-order cone programming problem of the following form:

Further, the specific implementation process of the step S6 is as follows:

用户最小可达速率

初始值Q¹、l¹和t¹；S61, initialize the performance parameters: the initial number of iterations k=1, the maximum number of iterations K, the tolerance value

User minimum reachable rate

initial values Q ¹ , l ¹ and t ¹ ;

S62. Solve the optimization problem in formula (26) through a convex optimization solver, and let Q ^* , l ^* and t ^* be the solutions of this iteration;

或者k＝K时，结束进程；否则，执行步骤S64；S63, when

Or when k=K, end the process; otherwise, execute step S64;

S64, define k= ^k +1, Qk=Q ^* , ^lk =l ^* and ^tk =t ^* , and jump back to step S62.

Compared with the prior art, the present invention has the following advantages:

1. The power control method for the cellular massive MIMO system based on the low-precision ADC under the Rice channel provided by the present invention is compared with the cellular massive MIMO system configured with the full-precision ADC under the Rice fading channel. The low-precision ADC can greatly reduce the high hardware cost and huge energy consumption brought by the full-precision ADC.

By adopting the power control algorithm based on the SCA mode in the present invention, the power control coefficient with the maximum total system rate can be iteratively obtained under the condition of ensuring the service quality of each user and satisfying the transmit power limit of each AP.

2. The power control method of the low-precision ADC-to-cellular massive MIMO system under the Rice channel provided by the present invention is based on the continuous convex approximation method, and the power control strategy is designed, which can ensure the service quality of each user and meet the requirements of each user. Under the condition that the transmit power of the AP is limited, iteratively obtains the power control coefficient with the maximum total system rate.

Based on the above reasons, the present invention can be widely promoted in the fields of wireless communication and the like.

Description of drawings

In order to illustrate the embodiments of the present invention or the technical solutions in the prior art more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description These are some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can also be obtained from these drawings without any creative effort.

Fig. 1 is the flow chart of the method of the present invention.

FIG. 2 is a model structure of a decellularized massive MIMO system based on a low-precision ADC under a Rice channel to which the solution of the present invention is applicable.

FIG. 3 is a simulation diagram of the system rate performance of the solution of the present invention under different iteration times.

FIG. 4 is a simulation diagram of the cumulative distribution function of the system rate performance when the power control method in the solution of the present invention is used and the power control algorithm is not used.

Detailed ways

In order to make those skilled in the art better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only Embodiments are part of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

It should be noted that the terms "first", "second" and the like in the description and claims of the present invention and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific sequence or sequence. It is to be understood that the data so used may be interchanged under appropriate circumstances such that the embodiments of the invention described herein can be practiced in sequences other than those illustrated or described herein. Furthermore, the terms "comprising" and "having", and any variations thereof, are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those expressly listed Rather, those steps or units may include other steps or units not expressly listed or inherent to these processes, methods, products or devices.

As shown in FIG. 1 , the present invention provides a power control method for a cellular massive MIMO system based on a low-precision ADC under a Rice channel, including the following steps:

S1. Establish a model of a cellular massive MIMO system based on a low-precision ADC under the Rice channel;

S2, based on the constructed model, perform uplink pilot training;

S3, based on the constructed model, perform downlink data transmission;

S4, analyze the user's downlink reachable rate;

S5. According to the downlink reachable rate, construct an optimization problem to maximize the total rate of the system;

S6. Design a power control strategy according to the optimization problem.

During specific implementation, as a preferred embodiment of the present invention, a decellularized massive MIMO system under the Rice fading channel as shown in FIG. 2 is established, wherein M single-antenna APs and N (M>N) single-antenna APs are constructed. Users are randomly distributed in a wide service area, and all APs are connected to the CPU through backhaul links and provide services to all users. In addition, in order to effectively reduce the high hardware cost and huge energy consumption caused by the full-precision ADC quantizer, in this embodiment, it is assumed that a low-precision ADC quantizer is configured at the AP and the user to quantize the received data. The model constructed in the step S1 is specifically expressed as follows:

和散射分量h_mn～CN(0,1)共同组成，其中θ_mn～[-π，π]表示到达角且CN(0,1)表示均值为0和方差为1的循环对称复高斯变量；另外，k_mn表示莱斯K-因子，代表视距分量与散射分量之间的比值，定义

和

其中β_mn＝γ_mn/(1+k_mn)，因此，信道系数被重新表示为

并且服从

的统计分布。Among them, _{gmn represents the channel coefficient between the mth access point AP and the nth user; γmn represents} the large-scale fading coefficient, which is used to reflect the influence of path loss and shadow fading on the channel coefficient; the part in brackets represents small-scale fading coefficient, determined by the line-of-sight component

and the scattering components h _mn ～CN(0,1), where θ _mn ～[-π, π] represents the angle of arrival and CN(0,1) represents a cyclic symmetric complex Gaussian variable with mean 0 and variance 1; In addition, k _mn represents the Rice K-factor, which represents the ratio between the line-of-sight component and the scattering component, defined

and

where β _mn =γ _mn /(1+ _{km mn} ), therefore, the channel coefficients are re-expressed as

and obey

statistical distribution.

During specific implementation, as a preferred embodiment of the present invention, the step S2 performs uplink pilot training based on the constructed model, and the specific implementation process is as follows:

S21. Let the length of the correlation interval be T, and let the time occupied by pilot training in each correlation interval be τ _p (τ _p ≥N);

并且假定不同用户间发送的导频序列是相互正交的，即

则第m个接入点AP接收到的上行导频信息表示为：S22. Assume that in the uplink pilot training phase, the nth user can send uplink pilot sequences to all access points APs

And it is assumed that the pilot sequences sent by different users are mutually orthogonal, that is,

Then the uplink pilot information received by the mth access point AP is expressed as:

表示第m个AP接收到的加性高斯白噪声；where _pp represents the normalized signal-to-noise ratio of the uplink pilot signal,

represents the additive white Gaussian noise received by the mth AP;

S23. Since the AP is equipped with a low-precision ADC, its quantization error will be affected by the received signal strength. In order to explore the impact of low-precision ADC on system performance, in this embodiment, the additive quantization noise model (AQNM) is used to model the quantization process, and the uplink pilot information received by the mth access point AP is re-expressed as :

的近似关系，

表示与接收信号y_m,p无关的加性量化噪声；Among them, α _m represents the ADC accuracy at the m-th access point AP, and the accuracy is related to the quantization bit number ρ _m used by the m-th access point AP, when the quantization bit number ρ _m =1,2,3, 4 and 5, α _m is precisely given as α _m =0.6366, 0.8825, 0.96546, 0.990503, 0.997501; when ρ _m >5, the relationship between α _m and ρ _m satisfies

approximation relationship,

represents the additive quantization noise independent of the received signal y _m,p ;

的协方差，表示为：S24. Calculate additive quantization noise

The covariance of , expressed as:

与

相乘，得到：S25. In order for the m-th access point AP to obtain the channel state information (CSI) from the n-th user, set the

and

Multiply to get:

S26. Since the access point AP can obtain the line-of-sight component information and large-scale fading coefficients from different users, in the channel estimation stage, the system can eliminate the influence of the line-of-sight component, that is, equation (5) can be re-expressed as:

表示消除视距分量影响后余留下的量化噪声；in,

Represents the quantization noise remaining after removing the influence of the line-of-sight component;

的协方差，表示为：S27. Calculate quantization noise

The covariance of , expressed as:

in,

S28. Use the minimum mean square error channel estimation scheme for formula (6) to obtain the channel estimation information of g _mn , as follows:

其中

S29. According to the properties of the MMSE channel estimation scheme, obtain

in

During specific implementation, as a preferred embodiment of the present invention, the step S3 performs downlink data transmission based on the constructed model, and the specific implementation process is as follows:

S31. After the access point AP obtains the CSI from the user, it uses the conjugate transpose (CB) method to precode the downlink transmission signal. Therefore, the transmission signal of the mth access point AP is expressed as:

的条件；p_d表示下行发送数据信号的归一化SNR；η_mn表示功率控制系数；Among them, x _n represents the data signal sent to the nth user, satisfying

conditions; p _d represents the normalized SNR of the downlink transmission data signal; η _mn represents the power control coefficient;

的功率约束条件，即：S32. Adjust the power control coefficient η _mn , so that each access point AP satisfies the

The power constraints of , namely:

S33. According to formula (9), after all access points APs have sent downlink data signals, the received signal of the nth user is expressed as:

Among them, w _n ~CN(0,1) represents the additive white Gaussian noise received by the nth user;

S34. Based on the low-precision ADC used by the user terminal, the received signal of the nth user after quantization processing is expressed as:

表示与接收信号r_n不相关的加性量化噪声；Among them, μ _n represents the ADC accuracy of the n-th user, which is related to the number of quantization bits σ _n used by the n-th user. For the specific correspondence between μ _n and σ _n , please refer to the ADC accuracy α _m of the m-th AP. The numerical correspondence between ρ _m and the number of quantized bits ρ m will not be repeated here.

represents the additive quantization noise uncorrelated with the received signal _rn ;

的协方差，计算公式如下：S35. Calculate additive quantization noise

The covariance is calculated as follows:

During specific implementation, as a preferred embodiment of the present invention, the step S4 analyzes the user's downlink reachable rate, and the specific implementation process is as follows:

S41. According to formula (12), the received signal of the nth user is re-expressed in the following form:

in,

S42. According to formula (14), the expression of the downlink reachable rate of the nth user is obtained as follows:

S43. According to formula (15), deduce the closed expression of the reachable rate of the nth user as:

的第n列；

和

分别表示

和

的第(n-1)×N+n₁行；矩阵Q、U和V中的对应元素被给定为[Q]_mn＝η_mn，

δ(m,n)表示狄拉克δ函数，即当m＝n时，δ(m,n)＝1；当m≠n时，δ(m,n)＝0。Among them, q _n represents

the nth column of ;

and

Respectively

and

The (n- ₁ )×N+n1th row of ; the corresponding elements in matrices Q, U and V are given as [Q] _mn = _ηmn ,

δ(m,n) represents the Dirac delta function, that is, when m=n, δ(m,n)=1; when m≠n, δ(m,n)=0.

During specific implementation, as a preferred embodiment of the present invention, the step S5 constructs an optimization problem of maximizing the total system rate according to the downlink reachable rate, and the specific implementation process is as follows:

和每个AP的发送功率不大于设定的最大发送功率p_d的前提下，将最大化系统总速率的优化问题归纳成如下形式：S51. For modern communication systems, the total rate of the system and the achievable rate of each user are important performance indicators for evaluating the quality of a communication system. Therefore, after ensuring that the quality of service for each user is not less than the set minimum reachable rate

On the premise that the transmit power of each AP is not greater than the set maximum transmit power p _d , the optimization problem of maximizing the total rate of the system is summarized into the following form:

S52. Perform mathematical transformation _on the constraints (17b) in the optimization problem P1, and you can obtain:

Among them, the constraint (17b) is equivalent to the second-order conical formula in formula (18), that is, formula (17b) is converted into the convex constraint in formula (18); formulas (17c) and (17d) are both: Convex constraints;

S53. Since the objective function in the optimization problem P1 is non _- convex, there is no way to directly obtain the global optimal solution for the optimization problem through a convex optimization solver. Therefore, in order to solve the above optimization problem, in this embodiment, the method of continuous convex approximation is used to approximate the non-convex optimization problem P ₁ into the form of the optimization problem P ₂ . If the optimization problem P ₂ can be solved smoothly, then _A suboptimal solution that can solve the optimization problem P1 is obtained, and this suboptimal solution is close to the global optimal solution. In this embodiment, the specific form of the optimization problem P ₂ is given as:

where l=[l ₁ ,...,l _N ] and t=[t ₁ ,...,t _N ];

S54. For n=1,...,N, the given constraints are as follows:

Among them, the objective function (19a) in the optimization problem P2 is convex, but the constraint ( ₂₀ ) is non-convex;

S55. Convert the constraint (20) into the form of a convex function. For n=1, . . . , N, the constraint (20) is equivalent to:

S56. When n=1,...,N, formula (21) is re-expressed in the following form:

分别表示Q和l_n经过k次SCA处理后的迭代值，基于此，公式(20)已经通过公式(23)成功地转换成了二阶锥形式的约束条件，另外，通过观察优化问题P₂，得到约束条件(19c)是凸的；where ^Qk and

respectively represent the iterative values of Q and _ln after k times of SCA processing. Based on this, formula (20) has been successfully transformed into a second-order conical constraint by formula (23). In addition, by observing the optimization problem P ₂ , the constraint (19c) is convex;

S57, according to the method of continuous convex approximation, use the first-order Taylor expansion in formula (24) to approximate formula (22);

S58. According to the method of continuous convex approximation, and using the property of the inequality ln(x)≥1-x ^-1 , convert the constraint condition (19c) into the following constraint condition with a second-order conical formula:

S59. According to formula (23) and formula (25), at the k-th iteration, the optimization problem P ₂ is expressed as a second-order cone programming problem (convex optimization problem) of the following form:

During specific implementation, as a preferred embodiment of the present invention, the step S6 designs a power control strategy according to the optimization problem, and the specific implementation process is as follows:

需要保证的用户最小可达速率

初始值Q¹、l¹和t¹；S61, initialize the performance parameters: the initial number of iterations k=1, the maximum number of iterations K, the tolerance value

The minimum user reachable rate that needs to be guaranteed

initial values Q ¹ , l ¹ and t ¹ ;

S62. Solve the optimization problem in formula (26) through a convex optimization solver, and let Q ^* , l ^* and t ^* be the solutions of this iteration;

或者k＝K时，结束进程；否则，执行步骤S64；S63, when

Or when k=K, end the process; otherwise, execute step S64;

S64, define k= ^k +1, Qk=Q ^* , ^lk =l ^* and ^tk =t ^* , and jump back to step S62.

Example

In order to verify the effectiveness of the scheme of the present invention, the following simulation experiments were carried out:

表示阴影衰落。系数

和

是相互独立的，并且给定δ_sf＝0.5和σ_sf＝8。莱斯K-因子描述了视距分量与散射分量之间的比值，给定为

Scenario setting: Assume that all users and access point APs are evenly and randomly distributed in a square area of 1×1 km, and define the distance between the mth access point AP and the nth user as _dmn . The large-scale fading coefficients describe path loss and shadow fading and are given as: γ _mn = -30.18-26log ₁₀ (d _mn )+F _mn (dB), where

Indicates shadow fading. coefficient

and

are independent of each other and given δ _sf =0.5 and σ _sf =8. The Rice K-factor describes the ratio between the line-of-sight component and the scattering component, given as

Other parameter values required in the simulation of the solution of the present invention are set as shown in Table 1.

Table 1 Parameter settings

As shown in FIG. 3 , the simulation results of the total rate of the system changing with the number of iterations are given when the scheme of the present invention is used for power control. It can be seen from Figure 2 that as the number of iterations continues to increase, the overall rate performance of the system continues to increase, and after the number of iterations exceeds 8, it gradually approaches a fixed value, which indicates that the method in the solution of the present invention will It gradually converges after 8 iterations.

As shown in FIG. 4 , when the number of iterations K=10, the cumulative distribution function of the total rate of the system when using the power control algorithm in the solution of the present invention and not using the power control algorithm is given. Through observation, it can be found that when the system does not use the power control algorithm, the total rate of the system has a 95% probability of being greater than 31.8bits/s/Hz; but when the system uses the scheme of the present invention for power control, the system There is a 95% probability that the total rate is greater than 42.9 bits/s/Hz, which is 1.35 times that of a system without a power control algorithm. The above simulation results verify the effectiveness of the solution of the present invention in improving the overall rate performance of the system.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, but not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: The technical solutions described in the foregoing embodiments can still be modified, or some or all of the technical features thereof can be equivalently replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the technical solutions of the embodiments of the present invention. scope.

Claims (7)

1. a power control method for removing cellular massive MIMO system based on low-precision ADC under a Rice channel, is characterized in that, comprises the steps: S1. Establish a model of a cellular massive MIMO system based on a low-precision ADC under the Rice channel; S2, based on the constructed model, perform uplink pilot training; S3, based on the constructed model, perform downlink data transmission; S4, analyze the user's downlink reachable rate; S5. According to the downlink reachable rate, construct an optimization problem to maximize the total rate of the system; S6. Design a power control strategy according to the optimization problem. 2. the power control method of removing cellular massive MIMO system based on low-precision ADC under the Rice channel according to claim 1, it is characterized in that, the model constructed in described step S1 is specifically expressed as follows:

Among them, _{gmn represents the channel coefficient between the mth access point AP and the nth user; γmn represents} the large-scale fading coefficient, which is used to reflect the influence of path loss and shadow fading on the channel coefficient; the part in brackets represents small-scale fading coefficient, determined by the line-of-sight component

and the scattering components h _mn ～CN(0,1), where θ _mn ～[-π, π] represents the angle of arrival and CN(0,1) represents a cyclic symmetric complex Gaussian variable with mean 0 and variance 1; In addition, k _mn represents the Rice K-factor, which represents the ratio between the line-of-sight component and the scattering component, defined

and

where β _mn =γ _mn /(1+ _{km mn} ), therefore, the channel coefficients are re-expressed as

and obey

statistical distribution. 3. the power control method of removing cellular massive MIMO system based on low-precision ADC under the Rice channel according to claim 1, is characterized in that, the concrete realization process of described step S2 is as follows: S21. Let the length of the correlation interval be T, and let the time occupied by pilot training in each correlation interval be τ _p (τ _p ≥N); S22. Assume that in the uplink pilot training phase, the nth user can send uplink pilot sequences to all access points APs

And it is assumed that the pilot sequences sent by different users are mutually orthogonal, that is,

Then the uplink pilot information received by the mth access point AP is expressed as:

where _pp represents the normalized signal-to-noise ratio of the uplink pilot signal,

represents the additive white Gaussian noise received by the mth AP; S23. Use the additive quantization noise model to model the quantization process, then the uplink pilot information received by the mth access point AP is re-expressed as:

Among them, α _m represents the ADC accuracy at the m-th access point AP, and the accuracy is related to the quantization bit number ρ _m used by the m-th access point AP, when the quantization bit number ρ _m =1,2,3, 4 and 5, α _m is precisely given as α _m =0.6366, 0.8825, 0.96546, 0.990503, 0.997501; when ρ _m >5, the relationship between α _m and ρ _m satisfies

approximation relationship,

represents the additive quantization noise independent of the received signal y _m,p ; S24. Calculate additive quantization noise

The covariance of , expressed as:

S25. In order for the mth access point AP to obtain the channel state information from the nth user, set the

and

Multiply to get:

S26. Since the access point AP can obtain the line-of-sight component information and large-scale fading coefficients from different users, in the channel estimation stage, the system can eliminate the influence of the line-of-sight component, that is, equation (5) can be re-expressed as:

in,

Represents the quantization noise remaining after removing the influence of the line-of-sight component; S27. Calculate quantization noise

The covariance of , expressed as:

in,

S28. Use the minimum mean square error channel estimation scheme for formula (6) to obtain the channel estimation information of g _mn , as follows:

S29. According to the properties of the MMSE channel estimation scheme, obtain

in

4. the power control method of removing cellular massive MIMO system based on low-precision ADC under the Rice channel according to claim 1, is characterized in that, the concrete realization process of described step S3 is as follows: S31. After the access point AP obtains the CSI from the user, it uses the conjugate transpose method to precode the downlink transmission signal. Therefore, the transmission signal of the mth access point AP is expressed as:

Wherein, x _n represents the data signal sent to the nth user, which satisfies the condition of E{|x _n | ² }=1; p _d represents the normalized SNR of the downlink transmitted data signal; η _mn represents the power control coefficient; S32. Adjust the power control coefficient η _mn , so that each access point AP satisfies the power constraint condition of E{|s _m | ² }≤p _d , that is:

S33. According to formula (9), after all access points APs have sent downlink data signals, the received signal of the nth user is expressed as:

Among them, w _n ~CN(0, 1) represents the additive white Gaussian noise received by the nth user; S34. Based on the low-precision ADC used by the user terminal, the received signal of the nth user after quantization processing is expressed as:

Among them, μ _n represents the ADC accuracy of the nth user, which is related to the number of quantization bits σn used by the _nth user,

represents the additive quantization noise uncorrelated with the received signal _rn ; S35. Calculate additive quantization noise

The covariance is calculated as follows:

5. the power control method of removing cellular massive MIMO system based on low-precision ADC under Rice channel according to claim 1, is characterized in that, the concrete realization process of described step S4 is as follows: S41. According to formula (12), the received signal of the nth user is re-expressed in the following form:

in,

S42. According to formula (14), the expression of the downlink reachable rate of the nth user is obtained as follows:

S43. According to formula (15), deduce the closed expression of the reachable rate of the nth user as:

Among them, q _n represents

the nth column of ;

and

Respectively

and

The (n- ₁ )×N+n1th row of ; the corresponding elements in matrices Q, U and V are given as [Q] _mn = _ηmn ,

δ(m,n) represents the Dirac delta function, that is, when m=n, δ(m,n)=1; when m≠n, δ(m,n)=0. 6. the power control method of removing cellular massive MIMO system based on low-precision ADC under the Rice channel according to claim 1, is characterized in that, the concrete realization process of described step S5 is as follows: S51. After ensuring that the service quality of each user is not less than the set minimum reachable rate

And under the premise that the transmit power of each AP is not greater than the set maximum transmit power p _d , the optimization problem of maximizing the total system rate is summarized into the following form:

S52. Perform mathematical transformation _on the constraint condition (17b) in the optimization problem P1, to obtain:

Among them, the constraint (17b) is equivalent to the second-order conical formula in formula (18), that is, formula (17b) is converted into a convex constraint in formula (18); formulas (17c) and (17d) are also is a convex constraint; S53. Using the method of continuous convex approximation, the non-convex optimization problem P ₁ is approximately equivalent to the form of the optimization problem P ₂ , and the specific form of the optimization problem P ₂ is:

where l=[l ₁ ,...,l _N ] and t=[t ₁ ,...,t _N ]; S54. For n=1,...,N, the given constraints are as follows:

Among them, the objective function (19a) in the optimization problem P2 is convex, but the constraint ( ₂₀ ) is non-convex; S55. Convert the constraint (20) into the form of a convex function. For n=1, . . . , N, the constraint (20) is equivalent to:

S56. When n=1,...,N, formula (21) is re-expressed in the following form:

where ^Qk and

Respectively represent the iterative values of Q and l _n after k times of SCA processing. Based on this, formula (20) has been successfully converted into a second-order conical constraint by formula (23). In addition, by observing the optimization problem P ₂ , the constraint (19c) is convex; S57, according to the method of continuous convex approximation, use the first-order Taylor expansion in formula (24) to approximate formula (22); S58. According to the method of continuous convex approximation, and using the property of the inequality ln(x)≥1-x ^-1 , convert the constraint condition (19c) into the following constraint condition with a second-order conical formula:

S59. According to formula (23) and formula (25), at the k-th iteration, express the optimization problem P ₂ as a second-order cone programming problem of the following form:

7. the power control method of removing cellular massive MIMO system based on low-precision ADC under Rice channel according to claim 1, is characterized in that, the concrete realization process of described step S6 is as follows: S61. Initialize the performance parameters: the initial number of iterations k=1, the maximum number of iterations K, the tolerance value θ=0.01, and the minimum user achievable rate

initial values Q ¹ , l ¹ and t ¹ ; S62. Solve the optimization problem in formula (26) through a convex optimization solver, and let Q ^* , l ^* and t ^* be the solutions of this iteration; S63, when

Or when k=K, end the process; otherwise, execute step S64; S64, define k= ^k +1, Qk=Q ^* , ^lk =l ^* and ^tk =t ^* , and jump back to step S62.

Images