CN108235425A - Based on the extensive antenna relay system of the optimal pairs of user of efficiency and its resource allocation methods - Google Patents

Abstract

The invention discloses a large-scale antenna relay system for paired users based on optimal energy efficiency and a resource allocation method thereof, belonging to the technical field of wireless communication. The system consists of multiple source nodes, multiple sink nodes and a relay node. The source node and the sink node correspond to each other to form a communication pair. The relay node uses the zero-forcing criterion to amplify and forward the signal of the source node. , the information transmission is completed within two time slots. All source and sink nodes in the system are configured with a single antenna, and the relay nodes are configured with a large-scale antenna array. The method of the present invention aims at maximizing the energy efficiency of the system, uses three system resource parameters of the source node transmission power, the relay node transmission power and the number of relay node antennas as optimization variables to establish a mathematical model, and then uses an alternate iterative optimization algorithm and an objective function The partial convexity and Lagrangian dual method of the method obtained the closed-form solutions of the optimal transmit power of the relay node and the optimal number of antennas of the relay.

Description

Pair-user large-scale antenna relay system and its resource allocation based on optimal energy efficiency method

technical field

The invention belongs to the technical field of wireless communication, and in particular relates to a large-scale antenna relay system for paired users based on optimal energy efficiency and a resource allocation method thereof.

Background technique

In recent years, massive multiple-input multiple-output (referred to as massive MIMO) technology has attracted more and more attention from academia and industry in the field of wireless communication, and is widely regarded as one of the key technologies of the fifth-generation mobile communication system. Massive MIMO technology refers to the centralized deployment of a large number of antenna arrays at the base station to serve a relatively small number of multiple users at the same time. Some studies have pointed out that by using large-scale antenna arrays at the base station to tap the available resources in the air space, many Compared with the new features of the traditional MIMO system, such as fast channel fading and thermal noise are averaged (also known as channel hardening), significantly reducing the transmit power of the base station and user end without affecting the system's achievable rate performance, ultra-high spatial resolution The rate can be used for precise beam alignment, and simple linear processing can almost perfectly eliminate the interference effect between multiple users, etc.

At the same time, as an important part of the future heterogeneous network, the multi-antenna relay communication system has been widely concerned by many research institutions and equipment manufacturers. By introducing multi-antenna relay nodes, many performance improvements can be obtained in terms of cell coverage, edge user throughput, link reliability, and transmission power consumption. However, in multi-user relay systems, inter-user interference has always been the main factor limiting the performance of multi-antenna relay systems. To solve this problem, the industry has proposed two mainstream solutions: one is to use the orthogonality of time-frequency resources to schedule different users on different time-frequency resource blocks, and to suppress inter-user interference by increasing video resource consumption; One is to achieve the purpose of reducing interference between users by jointly designing precoding and receiver modules. However, although the first type of method can obtain a better inter-user interference cancellation effect, it consumes too many system time-frequency resources and has a relatively large negative impact on the overall spectrum efficiency performance of the system. The second method brings greater difficulties in terms of implementation complexity. The jointly designed algorithm is usually too complex, which puts forward higher requirements for the hardware computing resources of the relay node and the sink node. Considering that massive MIMO technology can better eliminate inter-user interference by using low-complexity linear processing, Himal A. Massive MIMO provides the ability to suppress multi-user interference in multi-user communication to solve the problem of inter-user interference in a pair-user multi-antenna relay system, which makes the system performance of large-scale antenna relay have greater growth potential. At the same time, the simple linear processing method of the massive MIMO technology can also be used to greatly reduce the computing resource overhead of the relay node and the sink node.

It is worth noting that when introducing a large-scale antenna array into a relay node, it will inevitably bring about several problems. The use of a large-scale antenna array can reduce the transmit power of the source user and the transmit power of the relay node by multiples without affecting the system spectral efficiency, which is very beneficial for improving the transmit power efficiency of the system. However, the fixed power consumption of the radio frequency link, which is closely related to the large-scale antenna array, will increase exponentially with the number of antennas, and the increase in the total power consumption of the fixed circuit will inevitably affect the overall energy efficiency of the relay communication system. Especially when the total power consumption of the fixed circuit is much greater than the transmission power consumption, continuously increasing the number of antennas will cause the total energy efficiency of the system to decrease rather than increase. It can be seen that the increase in the number of relay antennas can reduce the transmission power consumption, but it will increase the circuit power consumption of the radio frequency link. There is a trade-off between the number of relay antennas and the transmission power. The resource allocation of power, relay node transmit power and relay antenna number has very important practical significance and application background, especially under the concept of green communication, the allocation of transmit power and antenna number will directly affect the energy efficiency level of the system. And this question has yet to be tackled by researchers. In order to solve the problem of resource allocation in large-scale antenna relay systems, we propose a joint resource allocation optimization model based on energy efficiency maximization of source node transmit power, relay node transmit power, and relay node antenna numbers. The function is too complex and there is no precise analytical expression, so the optimization problem solving process is very difficult.

The invention discloses a method for resource allocation of a large-scale antenna relay system for paired users based on optimal energy efficiency. The system includes multiple source nodes, multiple sink nodes and a relay node. The source nodes correspond to the sink nodes one by one to form a communication pair. The relay node uses the zero-forcing criterion to amplify the signal of the source node. Forwarding, information transmission is completed within two time slots. All source and sink nodes in the system are configured with a single antenna, and the relay nodes are configured with a large-scale antenna array. The method of the invention aims at maximizing the energy efficiency of the system, and establishes a mathematical model with three system resource parameters, the transmission power of the source node, the transmission power of the relay node and the number of antennas of the relay node, as optimization variables. Since there is no exact analytical expression for the objective function in this optimization problem, an exact lower bound analytical expression for the objective function is obtained by means of Jensen's inequality and the statistical properties of the inverse Wishart matrix. Using the quasi-concave characteristics of the analytical expressions about the optimization variables, the mathematical relational expression of the optimal transmit power of the source node and the relay node is obtained, and the three-variable optimization problem is transformed into a two-variable optimization problem. Furthermore, using the property of fractional programming, the original optimization problem is equivalently transformed into a subtractive form, and an alternate iterative optimization algorithm is proposed. Then, according to the partial convexity of the objective function in subtractive form and the Lagrangian dual method, the closed-form solutions of the optimal transmit power of the relay node and the optimal number of antennas of the relay are obtained.

Contents of the invention

The present invention proposes a resource allocation method for a multi-user large-scale antenna relay system based on optimal energy efficiency in order to enable the paired-user large-scale antenna relay system to obtain higher energy efficiency performance, and provides an alternate iterative optimization algorithm to obtain The closed-form solution of relay node transmit power and relay node antenna number.

According to the resource allocation method of the large-scale antenna relay system for paired users based on optimal energy efficiency of the present invention, the relay system includes K source nodes, K sink nodes and a relay node, and the source node and the sink node The nodes correspond one by one to form a communication pair. The relay node adopts the zero-forcing criterion to amplify and forward the signal of the source node, and completes the information transmission within two time slots; the relay system adopts the time-division duplex system. A certain user scheduling strategy is established, and the source node or the sink node with the same large-scale fading is placed in a time-frequency resource block, and the channel is subject to flat block fading. It is characterized in that the method includes the following steps:

S1. Establish a mathematical model with the goal of maximizing the total energy efficiency of the system, and taking the three system resource parameters of source node transmit power, relay node transmit power and relay node antenna number as optimization variables;

S2. Using Jensen's inequality and the characteristics of each step distance of the inverse Wishart matrix, find the lower bound analytical expression of the user's spectral efficiency; and based on the number of large-scale antennas deployed by the relay node is much greater than the number of users for the lower bound analytical expression Approximate simplification, substituting the lower bound analytical expression of the approximate simplification, and substituting the mathematical relationship between the optimal transmission power of the source node and the relay node into the objective function of the original optimization problem;

S3. Using an alternate iterative algorithm to find the optimal transmit power of the relay node and the optimal number of antennas of the relay node; and

S4. The relay node notifies all source nodes of the optimal transmission power value of the source node, and adjusts the transmission power of the relay node and the number of antennas of the relay node to the optimal values obtained in S3.

Further, the step S1 includes the following steps:

1). The relay node obtains the ideal channel state information between it and all source nodes and sink nodes through channel estimation, that is, the channel matrix and Among them, h _k represents the channel vector from the kth source node to the relay node and obeys the complex Gaussian distribution Indicates the large-scale fading factor from the source node to the relay node, Represents the channel vector from the relay node to the kth sink node and obeys the complex Gaussian distribution Indicates the large-scale fading factor from the relay node to the sink node; assuming that the system adopts a time-division duplex system, and the channel obeys flat block fading, that is, the channel coefficient remains unchanged during the channel coherence time; assuming that the system adopts a certain user scheduling strategy , put the source nodes (sink nodes) with the same large-scale fading in one time-frequency resource block;

2). In the first time slot, K source nodes send information symbols to the relay node at the same time, then the received signal vector r at the relay node can be expressed as the following form,

Among them, s=[s ₁ ,s ₂ ,...,s _K ] ^T , s _k (k=1,2,...,K) represents the transmitted symbol of the kth source node and n _r represents the additive white noise at the relay node in the first time slot and satisfies the complex Gaussian distribution ρ _s represents the average transmit power variable of each source node;

3). In the second time slot, the relay node uses zero-forcing reception and zero-forcing precoding to amplify and forward the received signal r, and its processing matrix is The resulting forwarded signal vector y _t is shown below,

y _t =Dy _r

where ψ is the power normalization factor to satisfy the average total transmit power constraint ρ _r at the relay node, namely,

but, Then, the relay node forwards the signal y _t to all sink nodes, and the signal received by the kth sink node can be expressed as follows,

Among them, n _k represents the additive white noise at the kth sink node and satisfies the complex Gaussian distribution

4). Based on the received signal expression of the sink node in step 3), the received SINR expression of the kth sink node can be obtained as follows,

Thus, the average spectral efficiency of the kth sink node can be obtained as shown in the following formula,

in, Indicates the spectral efficiency loss caused by taking the occupied two time slot resources into account;

5). Based on the average spectral efficiency expression in step 4), establish at the relay node to maximize the total system energy efficiency function η(ρ _s , ρ _r ) as the goal, take the source node transmit power ρ _s and the relay node The mathematical optimization model where the transmit power ρ _r is a variable is as follows,

Among them, EE(p, N) represents the energy efficiency function, S(p, N) represents the total spectral efficiency of the system, P _tot (p, N) represents the total power consumption of the system, μ _s ≥ 1 represents that each source node transmits The efficiency loss constant factor of the transmitter power amplifier device, μ _r ≥ 1 represents the efficiency loss constant factor of the relay node transmitter power amplifier device, P _s represents the constant fixed power consumption of each source node transmitter, P _r represents the relay node transceiver constant fixed power consumption on each antenna of the machine;

Described step S2 comprises the following steps:

6). Due to the inclusion of S _k in the objective function in step 5), its precise analytical expression is difficult to obtain, which is not conducive to the solution of subsequent optimization problems. Here, use Regarding the convexity of x and Jensen's inequality, the lower bound of the average spectral efficiency S _k in step 4) can be obtained as follows,

7).Using the moment properties of each order of the complex inverse Wishart matrix, it can be directly calculated to get The analytical expression for is as follows,

in,

8). Considering that the number of large-scale antennas deployed by relay nodes is usually much larger than the number of users, that is, N>>K>1, and using the condition of high signal-to-noise ratio, that is, γ _s ＞>1 and γ _r ＞>1, the The analytical expression obtained in step 7) Approximate simplification to the following form,

9). Based on the analytical expression in step 8) The objective function of the optimization problem in step 5) is approximately replaced, and the objective function of the optimization problem in step 5) is replaced, and transformed into an optimization problem of the following form,

10). In step 9), the objective function EE(p, N) is strictly quasi-concave with respect to the power vector under the condition of a given antenna number variable N, and there is a unique optimal transmit power combination The maximum energy efficiency is satisfied, and the relationship between the optimal transmit power of the source node and the optimal transmit power of the relay node is as follows,

In step 9), the objective function EE(p, N) is strictly quasi-concave with regard to the number of relay antennas under the condition of a given power variable, and it increases first and then decreases;

11). Substituting the optimal transmit power relationship in step 10) into the objective function in step 9), it can be simplified as a two-variable optimization problem, as shown below,

in,

Described step S3 comprises the following steps:

12). According to the nature of fractional programming, the optimization problem in step 11) can be equivalently transformed into a subtraction problem with parameter η, as shown below,

Wherein, η represents the energy efficiency parameter, solves relay optimal transmitting power and antenna number parameter and optimal energy efficiency value η ^* ;

The meanings of the symbols and some variables in the above steps are as follows: (·) ^H represents the conjugate transpose operation of the matrix, ε{·} represents the mathematical expectation operation for random quantities (vectors), tr{·} represents the trace of the matrix, Indicates a circular symmetric complex Gaussian random distribution with mean value μ and variance ^σ2 , ||·||indicates vector 2 norm operation, |·|indicates real number absolute value operation or complex number modulo operation, and N is the number of relay node antennas , K is the total number of user pairs.

Further, in order to solve the parameters of the optimal transmission power and the number of antennas of the relay and the optimal energy efficiency value η ^* , it can be realized by using the Dinkelbach method to perform alternate iterations. Specific steps are as follows:

12.1). Set the iteration termination ε, the iteration number variable m=0, and give the initial value of the energy efficiency parameter η ⁽⁰⁾ ≥ 0;

12.2). Use η ^(m) to solve the optimization problem about the relay transmission power and the number of relay antennas, as shown below,

12.3). When (NK)ρ _r ≥ (1+ξ)K/ν, the objective function in step 12.2) is convex. Taking advantage of the partial convexity of the objective function, let it calculate the first-order partial derivatives for _ρr and N respectively and make them 0, and then obtain the closed-form solution of the optimal transmit power of the relay node and the optimal number of antennas of the relay node, As follows,

in,

12.4). Update energy efficiency parameters And the number of iterations variable m=m+1;

12.5). Judgment When , the iterative operation is terminated and the optimal resource variable combination is output And the optimal energy efficiency value ^ηopt =η ^(m) ; if the termination condition is not met, return to step 12.2) to re-iterate;

12.6). The optimal number of antennas in step 12.5) is usually a decimal number. Take the two closest integer values to N ^opt and calculate the corresponding energy efficiency value. The antenna value with a larger energy efficiency value is the optimal antenna value.

The present invention also proposes a large-scale antenna relay system for paired users based on optimal energy efficiency, including K source nodes, K sink nodes and a relay node, and the source nodes correspond to the sink nodes one by one , forming a communication pair, the relay node adopts the zero-forcing criterion to amplify and forward the signal of the source node, and completes the information transmission within two time slots; the relay system adopts the time division duplex system, and adopts a certain Scheduling strategy, put the source node or sink node with the same large-scale fading in a time-frequency resource block, and the channel obeys flat block fading, characterized in that the relay system uses the above resource allocation method to configure the source The transmit power of the node, the transmit power of the relay node, and the number of antennas of the relay node.

The present invention proposes a method for resource allocation of a large-scale antenna relay system for paired users based on optimal energy efficiency, deduces the mathematical relationship between the optimal transmission power of the source node and the relay node, and uses an alternate iterative algorithm to directly obtain The closed-form optimal solution of the transmit power of the relay node and the number of antennas of the relay node is obtained. By allocating the optimal transmission power to the source node and the relay node, and configuring the optimal number of antennas on the relay node, the large-scale antenna relay system can obtain the benefits of large-scale antenna arrays while avoiding the need for large-scale antennas. The impact of more circuit power consumption generated by the number, so that the overall energy efficiency of the system reaches the highest level. This patent algorithm adopts the mature and efficient Dinkelbach iterative optimization method, and uses the standard convex optimization scheme to obtain the closed-form optimal solution of the transmit power of the relay node and the number of antennas of the relay node. Therefore, after a small number of iterative operations, it can converge to The optimal energy efficiency value, the algorithm complexity is relatively low.

Description of drawings

Fig. 1 is the system model of the inventive method;

Fig. 2 is the basic flowchart of algorithm of the present invention;

Figure 3 is a comparison diagram of the spectral efficiency lower bound analytical expression proposed in this patent and the Monte Carlo simulation results under different source-sink node pairs K scenarios;

Fig. 4 shows the variation of the energy efficiency performance of the alternate iterative resource allocation algorithm proposed in this patent under different circuit power consumption conditions with the number of source-sink node pairs;

Fig. 5 is a graph of the convergence trajectory of the alternate iterative algorithm proposed in this patent.

Detailed ways:

In conjunction with the algorithm flow chart shown in Figure 2, a method for resource allocation of a large-scale antenna relay system for paired users based on optimal energy efficiency of the present invention is specifically described, including the following steps:

1). Establish at the relay node to maximize the total system energy efficiency function EE(p, N) as the goal, with the source-relay node transmit power vector p = [ρ _s , ρ _r ] and the number of relay node antennas N is the mathematical optimization model of the variable, as shown below,

Among them, EE(p, N) represents the energy efficiency function, S(p, N) represents the total spectral efficiency of the system, P _tot (p, N) represents the total power consumption of the system, μ _s ≥ 1 represents that each source node transmits The efficiency loss constant factor of the transmitter power amplifier device, μ _r ≥ 1 represents the efficiency loss constant factor of the relay node transmitter power amplifier device, P _s represents the constant fixed power consumption of each source node transmitter, P _r represents the relay node transceiver The constant fixed power consumption on each antenna of the machine, S _k (p,N) represents the average spectral efficiency of the kth sink node, as shown below,

Among them, γ _k represents the received signal-to-interference-noise ratio of the kth sink node.

3).Using Jensen's inequality and the moment characteristics of each order of the complex inverse Wishart matrix, and considering the large-scale number of antennas and the high signal-to-noise ratio interval, that is, N>>K, and The simplified form of the lower bound closed expression of the spectral efficiency S _k in step 2) can be obtained, as shown below,

4). Based on the approximate expression of spectral efficiency in step 3) Replace the objective function of the optimization problem in step 2), and approximately convert it into an optimization problem of the following form,

5). Substitute the mathematical relationship between the source node and the relay node's optimal transmit power into the objective function in step 4), and simplify the original problem into a two-variable optimization problem, as shown below,

in,

6). Based on the optimization problem in step 5), an alternate iterative method is used to solve the optimal transmit power value. The specific steps are as follows:

6.1). Set the iteration termination ε, the iteration number variable m=0, and give the initial value of the energy efficiency parameter η ⁽⁰⁾ ≥ 0;

6.2). Use η ^(m) to solve the optimization problem about the relay transmission power and the number of relay antennas, as shown below,

6.3). When (NK)ρ _r ≥ (1+ξ)K/ν, the objective function in step 12.2) is convex. Taking advantage of the partial convexity of the objective function, let it calculate the first-order partial derivatives for _ρr and N respectively and make them 0, and then obtain the closed-form solution of the optimal transmit power of the relay node and the optimal number of antennas of the relay node, As follows,

in,

6.4). Update energy efficiency parameters And the number of iterations variable m=m+1;

6.5). Judgment When , the iterative operation is terminated and the optimal resource variable combination is output And the optimal energy efficiency value ^ηopt =η ^(m) ; if the termination condition is not met, return to step 12.2) to re-iterate;

6.6). The optimal number of antennas in step 6.5) is usually a decimal number, and the two nearest integer values to N ^opt are used to calculate the corresponding energy efficiency value, and the antenna value with a larger energy efficiency value is the optimal antenna value;

Among them, (·) ^H —represents the conjugate transposition operation of the matrix, ε{·}—mathematical expectation operation for random quantities (vectors), tr{·}—the trace of the matrix, —Represents a circular symmetric complex Gaussian random distribution with a mean of μ and a variance of ^σ2 , ||·||—represents a vector 2 norm operation, |·|—represents a real number absolute value operation or a complex number modulus operation, N—relay The number of node antennas, K—the total number of user pairs.

Fig. 3 shows different scenarios of the number of user pairs, when the transmission power ρ _r = ρ _s = 20dB, with the increase of the number of relay node antennas, the analytical expression of the lower bound of the spectrum efficiency given in this patent and the Monte Carlo The comparison curve of the numerical simulation results. It can be seen from the figure that the analytical approximate expression proposed by this patent has a very good approximation effect, and the difference between it and the Monte Carlo numerical simulation curve is almost negligible, which shows that the approximate analytical expression proposed by this patent Has very good effect. Figure 4 shows that when the fixed power consumption of the source node antenna and the fixed power consumption of the relay node antenna satisfy P _s =P _r =-20dB and -15dB, the energy efficiency value obtained by the iterative resource allocation algorithm given in this patent varies with Changes in the number of source-sink node pairs. Fig. 5 shows the convergence speed and the convergence track of the alternate iterative optimization algorithm proposed in this patent converging to the optimal energy efficiency value. It can be seen from the figure that the optimal energy efficiency value can be converged after about 6 iterations, and the convergence process is fast.

Claims (5)

1. A resource allocation method for a paired user large-scale antenna relay system based on optimal energy efficiency is disclosed, wherein the relay system comprises K information source nodes, K information sink nodes and a relay node, the information source nodes and the information sink nodes are in one-to-one correspondence to form a communication pair, the relay node adopts a zero forcing criterion to amplify and forward signals of the information source nodes, and information transmission is completed in two time slots; the relay system adopts a time division duplex system, adopts a certain user scheduling strategy, places information source nodes or information sink nodes with the same large-scale fading in a time frequency resource block, and the channel obeys flat block fading, and is characterized in that the method comprises the following steps:

s1, establishing a mathematical model which takes the total energy efficiency of a maximized system as a target and three system resource parameters of information source node transmitting power, relay node transmitting power and relay node antenna number as optimization variables;

s2, solving a lower bound analytical expression of the user spectrum efficiency by using a Jensen inequality and each step distance characteristic of an inverse Wishart matrix; the lower bound analytic expression is simplified approximately based on the fact that the number of large-scale antennas deployed by the relay node is far larger than the number of users, the lower bound analytic expression of the approximate simplification and the mathematical relation of the optimal transmitting power of the information source node and the relay node are substituted into the objective function of the original optimization problem;

s3, solving the optimal transmitting power of the relay node and the optimal number of the antennas of the relay node by using an alternative iterative algorithm; and

and S4, the relay node feeds back the optimal transmitting power value of the information source node to all the information source nodes, and adjusts the transmitting power of the relay node and the number of the relay node antennas to the optimal value obtained in S3.

2. The paired subscriber large scale antenna relay system resource allocation method according to claim 1, wherein said step S1 includes the steps of:

1) the relay node obtains ideal channel state information from the relay node to all source nodes and sink nodes through channel estimation, namely a channel matrixAndwherein h is_kRepresenting the channel vector from the kth source node to the relay node and obeying a complex Gaussian distribution Representing a large scale fading factor from the source node to the relay node,representing the channel vector of the relay node to the kth sink node and obeying a complex Gaussian distribution Representing a large-scale fading factor from the relay node to the sink node;

2) k source nodes transmit information symbols simultaneously to the relay node in the first slot, the received signal vector r at the relay node can be represented in the form,

wherein s ═ s₁,s₂,...,s_K]^T，s_k(K ═ 1, 2.. times.k) denotes a transmission symbol of the kth source node andn_rrepresents additive white noise of the first time slot at the relay node and satisfies a complex Gaussian distributionρ_sRepresenting an average transmission power variable of each source node;

3) in the second time slot, the relay node adopts zero-forcing receiving and zero-forcing precoding to amplify and forward the received signal r, and the processing matrix isFormed forwarded signal vector y_tAs will be shown below, in the following,

y_t＝Dy_r

where ψ is a power normalization factor to satisfy an average total transmit power constraint ρ at the relay node_rThat is to say that,

then the process of the first step is carried out,the relay node then transmits the signal y_tTo all the sink nodes, the signal received by the k-th sink node can be represented in the form,

wherein n is_kRepresents additive white noise at the kth sink node and satisfies a complex Gaussian distribution

4) Based on the received signal expression of the sink node in step 3), the received signal to interference plus noise ratio expression of the kth sink node can be obtained as shown below,

the average spectral efficiency of the kth sink node can thus be obtained:

wherein, represents the spectral efficiency loss generated by taking the occupied two time slot resources into account;

5) based on the average spectral efficiency expression in step 4), establishing at the relay node to maximize the total system energy efficiency function η (ρ_s,ρ_r) Targeting the source node transmit power ρ_sAnd relay node transmit power ρ_rFor a mathematical optimization model of the variables, as shown below,

wherein EE (P, N) represents an energy efficiency function, S (P, N) represents the total spectral efficiency of the system, P_tot(p, N) represents the total power consumption of the system, μ_sMore than or equal to 1 represents the constant factor of the efficiency loss of each source node transmitter power amplifier, mu_rMore than or equal to 1 represents the efficiency loss constant factor, P, of the power amplifier device of the relay node transmitter_sRepresenting a constant fixed power consumption, P, of each source node transmitter_rRepresents a constant fixed power consumption on each antenna of the relay node transceiver;

the step S2 includes the steps of:

6) the average spectral efficiency S in step 4) can be obtained according to the Jensen inequality_kThe lower bound of (a), as shown below,

7) directly calculating by using the property of each order moment of the complex inverse Wishart matrix to obtain the characteristic of each order moment in the step 6)The analytical expression of (a) is as follows,

wherein,

8) consider that the number of large-scale antennas deployed at a relay node is usually much larger than the number of users, i.e., N > K > 1, and utilize the condition of high signal-to-noise ratio, i.e., γ_s> 1 and γ_r> 1, using the analytical expression obtained in step 7)The approximation is simplified to the form that,

9) based on the analytical expression in step 8)Approximately replacing the objective function of the optimization problem in the step 5), converting the objective function into the following form of optimization problem,

10) step 9) the objective function EE (p, N) is strictly pseudo-concave with respect to the power vector and there is a unique optimum transmit power combination, given the variable N of the number of antennasThe energy efficiency maximization is satisfied, and the relation formula satisfied by the optimal transmitting power of the information source node and the optimal transmitting power of the relay node is shown as follows,

in the step 9), the objective function EE (p, N) is strictly pseudo-concave with respect to the number of relay antennas and is increased and then decreased under the condition of a given power variable;

11) substituting the optimal transmit power relation in step 10) into the objective function in step 9) can be simplified into a two-variable optimization problem, as shown below,

wherein,

the step S3 includes the steps of:

12) the optimization problem in step 11) is equivalently converted into a subtractive form problem with parameters η, according to the fractional programming nature, as shown below,

η represents energy efficiency parameters, and the relay optimal transmitting power, antenna number parameters and optimal energy value η are solved^＊；

The symbols and some variables in the above steps have the following meanings: (.)^HRepresents the conjugate transpose operation of the matrix,. epsilon. {. cndot.represents the mathematical expectation operation on the random quantity (vector), tr {. cndot.represents the trace of the matrix,mean is μ and variance is σ²The cyclic symmetry complex Gaussian is randomly distributed, | | | represents vector 2 norm operation, | represents real absolute value operation or complex modulo operation, and N is the number of the relay node antennas.

3. The resource allocation method for paired-user large-scale antenna relay system of claim 2, wherein in the step 12), the method of performing alternate iteration by using the Dinkelbach method is used to solve the optimal transmission power and the antenna number parameter of the relay and the optimal energy value η^＊Then (c) is performed.

4. The paired-user large-scale antenna relay system resource allocation method according to claim 2 or 3, wherein the step 12) includes the steps of:

12.1), setting iteration termination epsilon, setting an iteration number variable m to be 0, and setting an energy efficiency parameter initial value η⁽⁰⁾≥0；

12.2) utilization of η^(m)Solving the optimization problem with respect to the relay transmit power and the number of relay antennas, as shown below,

12.3) when (N-K) ρ_rWhen the value is more than or equal to (1+ ξ) K/v, the target function is convex in the step 12.2), and the partial convexity of the target function is utilized to make the target function to rho_rAnd N respectively find the first order partial derivatives and make them be 0, so that a closed form solution of the optimal transmission power of the relay node and the optimal number of antennas of the relay node can be obtained, as shown below,

wherein,

12.4) updating energy efficiency parametersAnd an iteration number variable m is m + 1;

12.5) judgmentThen stopping iterative operation and outputting optimum resource variable combinationAnd an optimum energy value η^opt＝η^(m)(ii) a If the termination condition is not met, returning to the step 12.2) to repeat iteration again;

12.6) the optimum number of antennas in step 12.5) is usually a fraction taken as N^optAnd calculating corresponding energy efficiency values of the two nearest integer values, wherein the antenna value with the larger energy efficiency value is the optimal antenna value.

5. A paired user large-scale antenna relay system based on optimal energy efficiency comprises K information source nodes, K information sink nodes and a relay node, wherein the information source nodes and the information sink nodes are in one-to-one correspondence to form a communication pair, the relay node adopts a zero forcing criterion to amplify and forward signals of the information source nodes, and information transfer is completed in two time slots; the relay system adopts a time division duplex system, adopts a certain user scheduling strategy, places information source nodes or information sink nodes with the same large-scale fading in a time frequency resource block, and channels obey flat block fading, and is characterized in that the relay system adopts the resource allocation method as claimed in claims 1-4 to configure the transmitting power of the information source nodes, the transmitting power of the relay nodes and the number of antennas of the relay nodes.