阅读说明：本技术 在Cell-free Massive MIMO下的正交多址与非正交多址的结合方法 (Method for combining orthogonal multiple access and non-orthogonal multiple access under Cell-free Massive MIMO ) 是由 吴少川 隋秋怡 李壮 于 2021-06-21 设计创作，主要内容包括：本发明提出在Cell-freeMassiveMIMO下的正交多址与非正交多址的结合方法,将所有用户分成L簇,每簇内用户数为1或2,每簇内的用户使用相同的导频,不同簇用户使用相互正交的导频,则正交导频的个数为L；将AP随机分为两部分,一部分为OMA用户提供服务,另一部分为NOMA用户提供服务；下行链路的数据传输依赖于共轭波束赋形,得到单用户下行数据传输的频谱效率；使用的功率分配算法来最大化最小SINR的功率控制算法P1,将功率控制转化为凸优化问题P2；对凸优化问题P2使用二分查找方法来找到下行最大最小化SINR的解,根据具体的用户数量和信道相关时间情况,将正交多址接入与非正交多址接入相结合,在不同的用户数量和相关时间条件下都能得到最佳的频谱效率。(The invention provides a method for combining orthogonal multiple access and non-orthogonal multiple access under Cell-freemassiveMIMO, which divides all users into L clusters, the number of users in each cluster is 1 or 2, the users in each cluster use the same pilot frequency, different clusters use mutually orthogonal pilot frequencies, and the number of the orthogonal pilot frequencies is L; the AP is randomly divided into two parts, wherein one part provides services for OMA users, and the other part provides services for NOMA users; the data transmission of the downlink depends on conjugate beam forming to obtain the frequency spectrum efficiency of single-user downlink data transmission; the power control algorithm P1 which uses the power allocation algorithm to maximize the minimum SINR converts the power control into a convex optimization problem P2; a binary search method is used for the convex optimization problem P2 to find a solution of the downlink maximum minimum SINR, orthogonal multiple access and non-orthogonal multiple access are combined according to the specific user number and the channel correlation time condition, and the optimal spectrum efficiency can be obtained under the conditions of different user numbers and correlation time.)

1. The method for combining the orthogonal multiple access and the non-orthogonal multiple access under the Cell-freeMassiveMIMO is characterized in that:

dividing all users into L clusters, wherein the number of users in each cluster is 1 or 2, the users in each cluster use the same pilot frequency, different cluster users use mutually orthogonal pilot frequencies, and the number of the orthogonal pilot frequencies is L;

the AP is randomly divided into two parts, one part provides service for OMA users, namely a cluster with the number of users being 1 in each cluster, namely k_l1, the set of APs serving OMA users is denoted by the symbol Φ_OTo represent; the other part provides services for NOMA users, namely clusters with the number of users being 2 in each cluster, namely k_lSymbol phi for set where AP serving NOMA users is located > 1_NTo represent;

divide all users into 2 parts, some users use OMA, and use symbol omega_OTo represent the set of clusters where users using OMA are located; another part of the users use NOMA, with the symbol omega_NTo represent a set of clusters in which users using NOMA are located;

the combination method comprises the following steps:

the method comprises the following steps: the data transmission of the downlink depends on conjugate beam forming to obtain the frequency spectrum efficiency of single-user downlink data transmission;

step two: the power control algorithm P1 which uses the power allocation algorithm to maximize the minimum SINR converts the power control into a convex optimization problem P2;

step three: a binary search method is used for the convex optimization problem P2 to find the solution of the downlink maximum minimum SINR.

2. The method of claim 1, wherein:

let τ be the correlation time length of the channel, i.e., the discretization length, τ be equal to the product of the channel correlation time and the channel correlation bandwidth, τ_pFor the duration of the uplink training, τ_uAnd τ_dRespectively the upstream and downstream rowsThe data transmission time is τ ═ τ -_p+τ_u+τ_d；

Setting the transmission time of uplink and downlink data equal, i.e. tau_u＝τ_dIf τ is τ_p+2τ_d；

In the training phase, all K users in the system automatically send a length tau to the AP_pThe pilot signal of (a); order toWhereinIs the pilot sequence assigned to the kth user, K ═ 1,2, …, K;

let the channel gain model be:

where beta represents a large scale fading, h_mkRepresents small scale fading, M1, 2.. M, K1, 2.. K., K is independently identically distributed;

the signal received by the mth AP is expressed as:

where ρ is_pIs the normalized signal-to-noise ratio, w, of each pilot symbol_mIs additive white Gaussian noise, w, of the mth AP end_mObey independent CN (0,1) distribution;

by means of a received pilot signal y_mThe channel gain g can be obtained_mkIs estimated value ofBy usingDenotes y_mIn thatProjection of (2):

because any two pilots are orthogonal or identical, the MMSE algorithm is usedEstimating the channel gain g_mkThe value of (c):

whereinβ_mkIs a large scale fading between the mth AP and the kth user;

if m APs are arranged in the system, all users are divided into L clusters, and each cluster has K users, then the channel estimation f_mlComprises the following steps:

using MMSE algorithm to obtain f_mlEstimated value of (a):

wherein the content of the first and second substances,β_mlk′is the mth AP and the mth APlarge scale fading for the k' th user in cluster.

3. The method of claim 2, wherein:

the transmission signal of the mth AP is:

wherein eta is_mlkThe power allocated to k users in the ith cluster for the mth AP,is the conjugate of the channel estimate, s_lkSymbols, p, for k users in the ith cluster_dIn order to normalize the downlink transmit power,

then the normalized transmit signal power expression for the mth AP is:

wherein the content of the first and second substances,

each AP needs to satisfy the power control conditions as follows:

wherein, N is the number of antennas on each AP;

equation (10) needs to be satisfied to eliminate the successive interference:

wherein the content of the first and second substances,representing the signal-to-interference-and-noise ratio when the user j in the first cluster decodes the user k in the same cluster when carrying out SIC algorithm; based on this, the achievable rate of user k in the final ith cluster can be written as:

the actual signals for decoding after SIC processing are:

user using NOMA:

users using OMA:

the signal to interference plus noise ratio thus derived is:

user using NOMA:

users using OMA:

the finally obtained spectrum efficiency formula of the single-user downlink data transmission is as follows:

wherein the content of the first and second substances,

4. the method of claim 3, wherein:

the power allocation algorithm used is a power control algorithm which maximizes the minimum SINR, and the power control conditions are as follows:

P1：

s.t.

the optimization problem in formula (19) is a non-convex problem; order toAnd introducing a relaxation variable v_m，λ_lk′j，λ_lk′kConverting P1 into a convex optimization problem P2:

P2：

s.t.

wherein the content of the first and second substances,

v_lj，1＝[λ_l1j … λ_l(k-1)j]^T，v_lj，1＝[λ_l1j … λ_l(k-1j]^T，

5. the method of claim 4, wherein:

considering equation (20) as a second order cone, P2 is a standard second order cone optimization problem, i.e., a convex optimization problem;

the problem P2 is solved using a binary search method to find the solution of the downlink maximum minimum SINR:

setting the maximum value boundary t of the achievable SINR_maxAnd a minimum boundary t_min(ii) a Let initial target SINR t be (t)_max+t_min)/2；

When the problem P2 is solvable for this target SINR t, then the minimum value t is determined_minIs set as t; if not, the maximum value t is determined_maxIs set as t;

the binary search method is continued until the difference between the upper and lower bounds is less than a preset threshold epsilon, i.e. (t)_max-t_min)＜ε。

Technical Field

The invention belongs to the technical field of communication, and particularly relates to a method for combining orthogonal multiple access and non-orthogonal multiple access under Cell-free Massive MIMO.

Background

(1) Non-orthogonal multiple access (nonorthogonal multiple access) is one of the most important core technologies in wireless communication, and the multiple access technology has been a major focus of 5G research. The Multiple Access technology adopted in the 4G system is Orthogonal Frequency Division Multiple Access (OFDMA), which makes the allocation and utilization of time-Frequency resources more reasonable and efficient by dividing subcarriers, thereby improving communication efficiency and obtaining higher communication quality. However, as the number of users increases, non-orthogonal shared resource access between users has become necessary.

The Non-Orthogonal Multiple Access (NOMA) in power domain carries out simple linear superposition to Multiple user signals in power domain, each time-frequency unit in the base station carries Multiple user signals, and the users are distinguished by the transmitting power of the downlink signals of the users. The downlink transmitting power allocated to the user with good channel is weak, and the downlink transmitting power allocated to the user with poor channel is strong. On the terminal side, all user signals are successively detected according to the sequence of strong first and weak second according to the principle of Successive Interference Cancellation (SIC) reception. Compared with the conventional Orthogonal Multiple Access (OMA), the power domain NOMA has significant improvements in both spectral efficiency and throughput.

(2) With the increasing demand for communication and the continuous progress of communication technology, Cell-free Massive MIMO technology is considered as one of the key technologies in the 5G era. However, when this technique is used, the user throughput at the cell edge is low, the reliability is poor, and the communication delay is large, which becomes a limitation to which the massive MIMO technique is applied. Therefore, a distributed, network-based massive MIMO technology, Cell-Free massive MIMO technology, is proposed and receives increasing attention. In the Cell-Free massive MIMO system, a plurality of users occupy the same time-frequency resource in a space multiplexing mode, so that the problems are solved, and the performance of the system is greatly improved. In Cell-Free mMIMO, a CPU is connected with a large number of scattered APs through high-speed error-Free links, and all the APs cooperate with each other to provide service for all users. In this way, each AP only needs to complete simple signal processing, and those complex operations are completed by the CPU, so that the price and power consumption of the AP are relatively low. In CF-mimo, the uplink and downlink operate in TDD mode, and the coherence interval is typically divided into three phases: uplink training, downlink payload data transmission and uplink payload data transmission. The uplink training completes the estimation of the channel, for TDD, the channel gains in the uplink and downlink are the same, and the channel estimation information obtained in the training process is applied to process the data signal in the processes of downlink payload data transmission and uplink payload data transmission.

(3) And the non-orthogonal multiple access of Cell-free Massive MIMO.

Disclosure of Invention

The invention provides a method for combining orthogonal multiple access and non-orthogonal multiple access under Cell-free Massive MIMO, which combines the orthogonal multiple access and the non-orthogonal multiple access, can exert the advantages of the orthogonal multiple access and the non-orthogonal multiple access, and can obtain the optimal frequency spectrum efficiency under the conditions of different user numbers and related time.

The invention is realized by the following scheme:

in the method for combining orthogonal multiple access and non-orthogonal multiple access under Cell-free Massive MIMO,

dividing all users into L clusters, wherein the number of users in each cluster is 1 or 2, the users in each cluster use the same pilot frequency, different cluster users use mutually orthogonal pilot frequencies, and the number of the orthogonal pilot frequencies is L;

the AP is randomly divided into two parts, one part provides service for OMA users, namely a cluster with the number of users being 1 in each cluster, namely k_l1, the set of APs serving OMA users is denoted by the symbol Φ_OTo represent; the other part provides services for NOMA users, namely clusters with the number of users being 2 in each cluster, namely k_lPhi 1, the set in which the AP providing service for NOMA users is located_NTo represent;

divide all users into 2 parts, some users use OMA, and use symbol omega_OTo represent the set of clusters where users using OMA are located; another part of the users use NOMA, with the symbol omega_NTo represent a set of clusters in which users using NOMA are located;

the combination method comprises the following steps:

the method comprises the following steps: the data transmission of the downlink depends on conjugate beam forming to obtain the frequency spectrum efficiency of single-user downlink data transmission;

step two: the power control algorithm P1 which uses the power allocation algorithm to maximize the minimum SINR converts the power control into a convex optimization problem P2;

step three: a binary search method is used for the convex optimization problem P2 to find the solution of the downlink maximum minimum SINR.

Further, the air conditioner is provided with a fan,

let τ be the correlation time length of the channel, i.e., the discretization length, τ be equal to the product of the channel correlation time and the channel correlation bandwidth, τ_pFor the duration of the uplink training, τ_uAnd τ_dFor uplink and downlink data transmission time, respectively, τ ═ τ is obtained_p+τ_u+τ_d；

Setting the transmission time of uplink and downlink data equal, i.e. tau_u＝τ_dIf τ is τ_p+2τ_d；

In the training phase, all K users in the system automatically send a length tau to the AP_pThe pilot signal of (a); order toWhereinIs the pilot sequence assigned to the kth user, K ═ 1,2, …, K;

let the channel gain model be:

where beta represents a large scale fading, h_mkRepresents small scale fading, with M1, 2.. M, K1, 2.. K., K being independently equally distributed;

the signal received by the mth AP is expressed as:

where ρ is_pIs the normalized signal-to-noise ratio, w, of each pilot symbol_mIs additive white Gaussian noise, w, of the mth AP end_mObey independent CN (0,1) distribution;

by means of a received pilot signal y_mThe channel gain g can be obtained_mkIs estimated value ofBy usingDenotes y_mIn thatProjection of (2):

because any two pilots are orthogonal or identical, the MMSE algorithm is usedEstimating the channel gain g_mkThe value of (c):

whereinβ_mkIs a large scale fading between the mth AP and the kth user;

setting m APs in the system, dividing all users into L clusters, and if there are K users in each cluster, then estimating the channel f_mlComprises the following steps:

using MMSE algorithm to obtain f_mlEstimated value of (a):

wherein the content of the first and second substances,β_mlk′large scale fading for the mth AP and the kth user in the ith cluster.

Further, it is characterized in that:

the transmission signal of the mth AP is:

wherein eta is_mlkThe power allocated to k users in the ith cluster for the mth AP,for the co-yoke of channel estimates, s_lkSymbols, p, for k users in the ith cluster_dIn order to normalize the downlink transmit power,

then the normalized transmit signal power expression for the mth AP is:

wherein the content of the first and second substances,

each AP needs to satisfy the power control conditions as follows:

wherein, N is the number of antennas on each AP;

equation (10) needs to be satisfied to eliminate the successive interference:

wherein the content of the first and second substances,representing the signal-to-interference-and-noise ratio when the user j in the first cluster decodes the user k in the same cluster when carrying out SIC algorithm; based on this, the achievable rate of user k in the final ith cluster can be written as:

the actual signals for decoding after SIC processing are:

user using NOMA:

users using OMA:

the signal to interference plus noise ratio thus derived is:

user using NOMA:

users using OMA:

the finally obtained spectrum efficiency formula of the single-user downlink data transmission is as follows:

wherein the content of the first and second substances,

further, the air conditioner is provided with a fan,

the power allocation algorithm used is a power control algorithm which maximizes the minimum SINR, and the power control conditions are as follows:

P1：

the optimization problem in equation (19) is a non-convex problem(ii) a Order toAnd introducing a relaxation variable

υ_m，λ_lk′j，λ_lk′kConverting P1 into a convex optimization problem P2:

P2：

wherein the content of the first and second substances,

v_lj，1＝[λ_l1j…λ_l(k-1)j]^T，vlj，1＝[λ_l1j…λ_l(k-1)j]^T，

further, the air conditioner is provided with a fan,

considering equation (20) as a second order cone, P2 is a standard second order cone optimization problem, i.e., a convex optimization problem;

the problem P2 is solved using a binary search method to find the solution of the downlink maximum minimum SINR:

setting the maximum value boundary t of the achievable SINR_maxAnd a minimum boundary t_min(ii) a Let initial target SINR t be (t)_max+t_min)/2；

When the problem P2 is solvable for this target SINR t, then the minimum value t is determined_minIs set as t; if not, the maximum value t is determined_maxIs set as t;

the binary search method is continued until the difference between the upper and lower bounds is less than a preset threshold epsilon, i.e. (t)_max-t_min)＜ε。

The invention has the beneficial effects

The invention combines the orthogonal multiple access and the non-orthogonal multiple access according to the specific user number and the channel correlation time condition, can exert the advantages of the orthogonal multiple access and the non-orthogonal multiple access, and can obtain the optimal spectrum efficiency under the conditions of different user numbers and correlation time.

Compared with the non-orthogonal multiple access in which the number of users per cluster must be 2, the method is more flexible and can combine the advantages of the two access modes when the relevant time is not long or short: some users have a shorter pilot length by non-orthogonal multiple access, and some users have a higher sum rate by orthogonal multiple access.

Drawings

FIG. 1 is a system diagram of the combination of NOMA and OMA in Cell-Free massive MIMO according to the present invention;

FIG. 2 is a comparison of the combination of NOMA and OMA of the present invention with the spectral efficiencies of OMA and NOMA.

Detailed Description

The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments; all other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, are within the scope of the present invention.

The current methods all discuss the case of K ═ 2. Obviously, it is not reasonable nor flexible to forcibly specify that the number of users in each cluster is equal, and the maximum value K of the number of users in a cluster should be set according to the specific number of users and the channel correlation time condition_maxAnd specify k_l≤K_maxWherein k is_lThe number of users in the ith cluster.

All users are divided into L clusters, the number of users in each cluster is 1 or 2, the users in each cluster use the same pilot frequency, different cluster users use mutually orthogonal pilot frequencies, and therefore the number of the orthogonal pilot frequencies is L. In order to be able to serve all users, we also randomly split the AP into two parts, one part serving OMA users (i.e. a cluster of 1 user per cluster) and one part serving NOMA users (a cluster of more than 1 user per cluster). The system block diagram is shown in FIG. 1:

as shown in FIG. 1, all APs and UEs are divided into 2 parts, one part of users uses non-orthogonal multiple access (light UE), and each cluster has more than 1 user, i.e. k_l> 1, all k_lThe same pilot is used by each user, and the cluster set in which this part of users is located is denoted by the symbol Ω_NTo indicate. Serving users non-orthogonally multiple-accessed by a portion of APs (light-colored APs), the set in which this portion of APs is located being marked by the symbol Φ_NTo indicate. Another part of the users use orthogonal multiple access (dark UE), and there are only 1 user in each cluster, i.e. k_l1, the cluster set where this part of users is located is denoted by the symbol Ω_oTo indicate. And served by another part of the APs (dark colored APs), the set in which this part of the APs is located being denoted by the symbol Φ_OTo indicate.

The invention is realized by the following scheme:

in the method for combining orthogonal multiple access and non-orthogonal multiple access under Cell-free Massive MIMO,

dividing all users into L clusters, wherein the number of users in each cluster is 1 or 2, the users in each cluster use the same pilot frequency, different cluster users use mutually orthogonal pilot frequencies, and the number of the orthogonal pilot frequencies is L;

the AP is randomly divided into two parts, one part provides service for OMA users, namely a cluster with the number of users being 1 in each cluster, namely k_l1, the set of APs serving OMA users is denoted by the symbol Φ_OTo represent; the other part provides services for NOMA users, namely clusters with the number of users being 2 in each cluster, namely k_lPhi 1, the set in which the AP providing service for NOMA users is located_NTo represent;

divide all users into 2 parts, some users use OMA, and use symbol omega_OTo represent the set of clusters where users using OMA are located; another part of the users use NOMA, with the symbol omega_NTo represent a set of clusters in which users using NOMA are located;

the combination method comprises the following steps:

the method comprises the following steps: the data transmission of the downlink depends on Conjugate beam forming (CB), and the spectrum efficiency of single-user downlink data transmission is obtained;

step two: the power control algorithm P1 which uses the power allocation algorithm to maximize the minimum SINR converts the power control into a convex optimization problem P2;

step three: a binary search method is used for the convex optimization problem P2 to find the solution of the downlink maximum minimum SINR.

Let τ be the correlation time length of the channel, i.e. the discretization lengthτ is equal to the product of the channel correlation time and the channel correlation bandwidth, τ_pFor the duration of the uplink training, τ_uAnd τ_dFor uplink and downlink data transmission time, respectively, τ ═ τ is obtained_p+τ_u+τ_d；

Setting the transmission time of uplink and downlink data equal, i.e. tau_u＝τ_dIf τ is τ_p+2τ_d；

In the training phase, all K users in the system automatically send a length tau to the AP_pThe pilot signal of (a); order toWhereinIs the pilot sequence assigned to the kth user, K ═ 1,2, …, K;

let the channel gain model be:

where beta represents a large scale fading, h_mkRepresents small scale fading, with M1, 2.. M, K1, 2.. K., K being independently equally distributed;

since the APs and the users are discretely distributed over a large area, h can be considered as_mkM1, 2,. M, K1, 2,. K, K are independently and identically distributed;

the signal received by the mth AP is expressed as:

where ρ is_pIs the normalized signal-to-noise ratio, w, of each pilot symbol_mIs additive white Gaussian noise, w, of the mth AP end_mObey independent CN (0,1) distribution;

by means of a received pilot signal y_mCan be used forObtain the channel gain g_mkIs estimated value ofBy usingDenotes y_mIn thatProjection of (2):

because any two pilots are orthogonal or identical, the MMSE algorithm is usedEstimating the channel gain g_mkThe value of (c):

whereinβ_mkIs a large scale fading between the mth AP and the kth user;

setting m APs in the system, dividing all users into L clusters, and if there are K users in each cluster, then estimating the channel f_mlComprises the following steps:

using MMSE algorithm to obtain f_mlEstimated value of (a):

wherein the content of the first and second substances,β_mlk′large-scale fading for the mth AP and the kth user in the ith cluster;

the transmission signal of the mth AP is:

wherein eta is_mlkThe power allocated to k users in the ith cluster for the mth AP,for the co-yoke of channel estimates, s_lkSymbols, p, for k users in the ith cluster_dIn order to normalize the downlink transmit power,

then the normalized transmit signal power expression for the mth AP is:

wherein the content of the first and second substances,

each AP needs to satisfy the power control conditions as follows:

wherein, N is the number of antennas on each AP;

equation (10) needs to be satisfied to eliminate the successive interference:

wherein the content of the first and second substances,representing the signal-to-interference-and-noise ratio when the user j in the first cluster decodes the user k in the same cluster when carrying out SIC algorithm; based on this, the achievable rate of user k in the final ith cluster can be written as:

the actual signals for decoding after SIC processing are:

user using NOMA:

users using OMA:

the signal to interference plus noise ratio thus derived is:

user using NOMA:

users using OMA:

the finally obtained spectrum efficiency formula of the single-user downlink data transmission is as follows:

wherein the content of the first and second substances,

the power allocation algorithm used is a power control algorithm which maximizes the minimum SINR, and the power control conditions are as follows:

P1：

the optimization problem in formula (19) is a non-convex problem; order toAnd introducing a relaxation variable upsilon_m，λ_lk′j，λ_lk′kConverting P1 into a convex optimization problem P2:

P2：

wherein the content of the first and second substances,

v_lj，1＝[λ_l1j…λ_l(k-1)j]^T，v_lj，1＝[λ_l1j…λ_l(k-1)j]^T，

considering equation (20) as a Second Order Cone (SOC), P2 is a standard Second Order Cone optimization problem, i.e., a convex optimization problem;

the problem P2 is solved using a binary search method to find the solution of the downlink maximum minimum SINR:

setting the maximum value boundary t of the achievable SINR_maxAnd a minimum boundary t_min(ii) a Let initial target SINR t be (t)_max+t_min)/2；

When the problem P2 is solvable for this target SINR t, then the minimum value t is determined_minIs set as t; if not, the maximum value t is determined_maxIs set as t;

the binary search method is continued until the difference between the upper and lower bounds is less than a preset threshold epsilon, i.e. (t)_max-t_min)＜ε。

The numerical simulation result of the method is given as follows:

the APs are uniformly distributed in a square area of 1km x 1km, and all users are randomly distributed. The path loss model is used here as follows:

wherein d is_mlkDenotes the distance between user k and base station m in the ith cluster, and PL_mlkRepresenting the path loss between the user k in the ith cluster and the base station;

wherein the content of the first and second substances,

L ＝46.3+33.9log₁₀(f)-13.82log₁₀(h_AP)-(1.1log₁₀(f)-0.7)h_u+(1.56log₁₀(f)- 0.8)

the rest simulation parameters are shown in Table 1

Parameter name	Parameter value
		Carrier frequency	1.9GHz
Bandwidth of	20MHz
		Noise figure	9dB
AP antenna height	15m
		Number of antennas per AP (N)	10
Height of user antenna	1.65m
		Pilot transmission power	100mW
Downlink data transmission power	200mW
		Shadow fading variance	8dB

TABLE 1

The power allocation problem for users using NOMA and OMA is optimized separately using the power allocation algorithm described above. A total of 40 APs are provided, wherein 20 of the APs provide services for NOMA users and 20 of the APs provide services for OMA users; there are 20 users, 10 of which use NOMA technology, divided into 5 clusters, 2 users per cluster, and 10 users use OMA technology, divided into 10 clusters, and 1 user per cluster. The final results are shown in FIG. 2:

spectral efficiency results the plots show that the spectral efficiency of cell-free massive MIMO combining NOMA and OMA is higher than that of NOMA or OMA alone under certain relevant time conditions.

This is because the advantage of orthogonal multiple access is that there is no interference from users using the same pilot and therefore the sum rate (R in the spectral efficiency formula) is higher than that of non-orthogonal multiple access at the same downlink transmission time. However, the orthogonal multiple access requires a large number of orthogonal pilots, and the length of the pilots is longer, so that the time occupied for transmitting the pilots is longer, which results in shorter time for uplink and downlink data transmission, which may seriously affect the spectrum efficiency when the correlation time is shorter and the number of users is larger. The advantages of non-orthogonal multiple access are that the number of needed orthogonal pilot frequencies is less, the length of the pilot frequency is shorter, and more time can be used for uplink and downlink data transmission. However, non-orthogonal multiple access has intra-cluster interference, influence and rate, when the correlation time is longer, the influence of the pilot frequency length on the spectrum efficiency is smaller, and the influence of the intra-cluster interference causes the spectrum efficiency to be lower than that of orthogonal multiple access.

The combination of the orthogonal multiple access and the non-orthogonal multiple access is a compromise implementation mode, one part of users in the system use the orthogonal access (the number of users per cluster is 1), one part of users use the non-orthogonal access (the number of users per cluster is 2), compared with the non-orthogonal multiple access, the system is more flexible in that the number of users per cluster is required to be 2, and the advantages of the two access modes can be combined when the relevant time is not long or short (the pilot frequency length is shortened by the non-orthogonal multiple access of one part of users, and the rate is higher by the non-orthogonal multiple access of one part of users). But when the correlation time is small, because a part of the users use the orthogonal multiple access, the length of the needed orthogonal pilot frequency is larger than that of all the users using the non-orthogonal multiple access; when the correlation time is large, the intra-cluster interference causes the sum rate to be reduced and the spectrum efficiency is influenced because a part of users use the non-orthogonal multiple access.

The method for combining orthogonal multiple access and non-orthogonal multiple access under Cell-free Massive MIMO proposed by the present invention is described in detail above, and the principles and embodiments of the present invention are explained in the present document by applying numerical simulation examples, and the description of the above examples is only used to help understanding the method and core ideas of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and as described above, the content of the present specification should not be construed as a limitation to the present invention.