MULTI-CARRIER RESOURCE ALLOCATION METHOD BASED ON WIRELESS-POWERED BACKSCATTER COMMUNICATION NETWORK

Abstract

The present invention relates to a multi-carrier resource allocation method based on a wireless-powered backscatter communication network. The method comprises following steps: S1. constructing a wireless-powered backscatter communication system; S2: according to circuit power and transmit power constraints, establishing a resource allocation optimization problem taking a maximum total transmission rate of the system as an objective function; S3: according to the objective function and constraint conditions, decomposing an optimization sub-problem taking the transmit power of the backscatter transmitter as a variable; S4: after substituting optimal transmit power of the backscatter transmitter into an original problem, decomposing a sub-problems taking an energy allocation coefficient as a variable from an original optimization problem; S5: converting non-convex problems containing coupling variables into convex problems, creating a Lagrangian function, obtaining an optimal solution form according to a KKT condition, and iteratively updating corresponding variables using a gradient descent method until convergence.

Claims

1. A multi-carrier resource allocation method based on a wireless-powered backscatter communication network, comprising following steps: S1. constructing a wireless-powered backscatter communication system, including a backscatter transmitter, an information receiver and a power station, wherein the backscatter transmitter is provided with an energy harvesting module and a backscatter circuit; adopting a multi-carrier downlink transmission mode, and dividing a whole transmission process into two stages: in first stage, the power station transmits radio-frequency signals to the backscatter transmitter, and the backscatter transmitter reflects some signals and converts remaining radio-frequency signals into energy for storage; and in second stage, the backscatter transmitter uses harvested energy for data transmission and circuit power consumption; S2: according to circuit power consumption constraints and transmit power constraints, establishing a resource allocation optimization problem taking a maximum total transmission rate of the system as an objective function; S3: according to the objective function and constraint conditions, decomposing an optimization sub-problem taking transmit power of the backscatter transmitter as a variable; S4: after substituting optimal transmit power of the backscatter transmitter into an original problem, decomposing a sub-problem taking an energy allocation coefficient as a variable from an original optimization problem; and S5: converting non-convex problems containing coupling variables into convex problems using a variable substitution method, creating a Lagrangian function using a Lagrange dual method, obtaining an optimal solution form according to KKT condition, and iteratively updating corresponding variables using a gradient descent method until convergence.

2. The multi-carrier resource allocation method based on a wireless-powered backscatter communication network according to claim 1, characterized in that step S1 comprises: using T.sub.1 and T.sub.2 to represent time spent in the two stages of the transmission process, T=T.sub.1+T.sub.2 representing time required for the whole transmission process; dividing the total system bandwidth B into K orthogonal subcarriers, so the bandwidth of each subcarrier is B.sub.e=B/K; defining a subcarrier set as ∀k∈K={1, 2, . . . , K}, assuming that each subcarrier experiences flat fading; in the first stage T.sub.1, the signals received by the backscatter transmitter are represented as: $y\left(t\right)=\underset{k=1}{\overset{K}{.Math.}}\sqrt{p_k{h}_k}{s}_k\left(t\right)+n\left(t\right)$ where p.sub.k represents transmit power allocated by the power station to the backscatter transmitter through subcarrier k, h.sub.k represents channel power gain from the power station to the backscatter transmitter, s.sub.k (t) represents transmit signals, n(t) represents a background noise and n(T)□N(0,σ.sup.−2); dividing the signals received by the backscatter transmitter into two parts by a backscatter coefficient α, where √{square root over (α)}y(t) is used for data backscatter, and √{square root over (1−α)}y(t) is used for energy harvesting; and in the stage T.sub.1, the time used for energy harvesting is represented as τ, τ≤T.sub.1, and the harvested energy is represented as: $E=\tau \eta \left(1-\alpha \right)\stackrel{K}{\underset{k=1}{.Math.}}{p}_k{h}_k$ where η∈[0,1] represents the energy harvesting efficiency, and the harvested energy is used to support circuit power consumption and information transmission in the stage T.sub.2; assuming that the energy used for circuit power consumption is (1−x)E, the remaining energy xE is used for information transmission, where x represents an energy allocation coefficient, thereby obtaining:
(1−x)E≥p.sub.eT.sub.1+p.sub.dT.sub.2 where p.sub.e and p.sub.d respectively represent the circuit power consumption in the T.sub.1 and T.sub.2 time periods respectively, that is, the harvested energy used for circuit power consumption must be greater than or equal to the total power consumption of the circuit in the T.sub.1 and T.sub.2 periods; assuming that the backscatter transmitter can decode backscatter signals, defining the background noise power on each subcarrier as σ.sub.k=σ.sup.−2/K, so the backscatter data rate from the backscatter transmitter to the information receiver is $R_b=T_1\underset{k=1}{\overset{K}{.Math.}}{B}_e{\log }_2\left(1+\frac{\alpha p_k{h}_k{g}_k}{{\sigma }_k}\right)$ where g.sub.k represents the channel power gain from the backscatter transmitter to the information receiver, B.sub.e represents the bandwidth of each subcarrier, defining the transmit power of the backscatter transmitter on each subcarrier as P.sub.k in the stage T.sub.2 so the data rate of each subcarrier is $R_k^h=B_e\kappa T_2{\log }_2\left(1+\frac{P_k{g}_k}{{\sigma }_k}\right)$ where κ∈[0,1] represents the transmission efficiency.

3. The multi-carrier resource allocation method based on a wireless-powered backscatter communication network according to claim 2, characterized in that in step S2, in order to improve the transmission efficiency of the entire system, a maximum total transmission rate of the information receiver is obtained by jointly optimizing the transmit power, the time allocation and the backscatter coefficient, and the resource allocation optimization problem is established as: $\underset{T_1,T_2,p_k,P_k,x,\tau ,\alpha }{\max }R$ $s.t.C1:\left(1-x\right)\tau \eta \left(1-\alpha \right)P_s\ge p_e{T}_1+p_d{T}_2,$ $C2:T_2\underset{k=1}{\overset{K}{.Math.}}{P}_k\le x\tau \eta \left(1-\alpha \right)P_s,$ $C3:T_1+T_2=T,$ $C4:\underset{k=1}{\overset{K}{.Math.}}{p}_k\le P^{\max },$ $C5:\tau \le T_1,$ $C6:0\le \alpha \le 1,$ $C7:0\le x\le 1,$ $C8:p_k\ge 0,P_k\ge 0,T_1\ge 0,T_2\ge 0,\tau \ge 0.$ where P.sub.max represents the maximum transmit power of the power station, $P_s=\underset{k=1}{\overset{K}{.Math.}}{p}_k{h}_k$ represents the total power of the signals received by the backscatter transmitter, $R=R_b+\underset{k=1}{\overset{K}{.Math.}}{R}_k^h$ represents the total data rate, C1 represents the minimum circuit power consumption constraint, C2 represents the maximum transmit power constraint of the backscatter transmitter in the stage T.sub.2, and C4 represents the maximum transmit power constraint of the power station.

4. The multi-carrier resource allocation method based on a wireless-powered backscatter communication network according to claim 3, characterized in that in step S3, the transmit power P.sub.k of the backscatter transmitter is only restricted by C2, and an optimization sub-problem is decomposed and solved by using a water-filling algorithm, so the optimal P.sub.k* is found first and is substituted into the optimization problem; consider the unit bandwidth over each subcarrier, an optimization sub-problem of the transmit power of the backscatter transmitter is written as: $\underset{P_k}{\max }{T}_2\kappa \underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(1+\frac{P_k{g}_k}{{\sigma }_k}\right)$ $s.t.C2:T_2\underset{k=1}{\overset{K}{.Math.}}{P}_k\le x\tau \eta \left(1-\alpha \right)P_s.$ the above problem is a convex optimization problem since the constraint condition is linear constraint and the objective function is a concave function; g.sub.k=σ.sub.k/g.sub.k is defined, and the optimal transmit power obtained by using the water-filling algorithm is $P_k={\left[\frac{1}{\lambda \ln 2T_2}-{\stackrel{\_}{g}}_k\right]}^{+}$ where [x].sup.+=max (0,x), λ represents a Lagrange multiplier, and λ satisfies the following formula: $T_2\underset{k=1}{\overset{K}{.Math.}}\left(\frac{1}{\lambda \ln 2T_2}-{\stackrel{\_}{g}}_k\right)=x\tau \eta \left(1-\alpha \right)P_s$ $\lambda =\frac{K}{\ln 2\left(T_2\underset{k=1}{\overset{K}{.Math.}}{\stackrel{\_}{g}}_k+x\tau \eta \left(1-\alpha \right)P_s\right)}$ is obtained, and the transmit power P.sub.k is substituted, thus obtaining an optimal power allocation policy in the stage T.sub.2: $P_k^{*}=\frac{\underset{k=1}{\overset{K}{.Math.}}{\overline{g}}_k}{K}+\frac{x\tau \eta \left(1-\alpha \right)P_s}{K{T}_2}-{\overline{g}}_k$ if more harvested energy is used for data transmission, the transmit power P.sub.k* may be greater, and the total data rate may be greater; if the transmit power p.sub.k of the power station becomes larger, P.sub.k* also becomes larger; and since the backscatter transmitter has no energy storage device, the transmission rate is limited by the power station.

5. The multi-carrier resource allocation method based on a wireless-powered backscatter communication network according to claim 4, characterized in that in step S4, P.sub.k* is substituted into the original optimization problem, and an optimization sub-problem with an energy coefficient is decomposed, $\underset{x}{\max }{T}_1\underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(1+\frac{\alpha p_k{h}_k}{{\stackrel{\_}{g}}_k}\right)+T_2\kappa \underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(\stackrel{\_}{K}+\frac{x\tau \eta \left(1-\alpha \right)P_s{g}_k}{{\stackrel{\_}{g}}_k{\mathrm{KT}}_2}\right)$ $s.t.C1:\left(1-x\right)\tau \eta \left(1-\alpha \right)P_s\ge p_e{T}_1+p_d{T}_2,$ $C7:0\le x\le 1.$ where $\stackrel{\_}{K}=\left(\underset{k=1}{\overset{K}{.Math.}}{g}_j/g_k\right)$ represents a constant, because the objective function of the above optimization sub-problem is a monotonically increasing function about the energy allocation coefficient x, according to the constraints C1 and C7, the value of x is $x=\max \left(1,1-\frac{p_e{T}_1+p_d{T}_2}{\tau \eta \left(1-\alpha \right)P_s}\right);$ because the right side of the inequality of the constraint C1 is greater than zero, if x=1, the constraint condition is not satisfied, so the optimal energy allocation coefficient is $x^{*}=1-\frac{p_e{T}_1+p_d{T}_2}{\tau \eta \left(1-\alpha \right)P_s};$ and because x≥0, the backscatter coefficient α must satisfy $\alpha \le 1-\frac{p_e{T}_1+p_d{T}_2}{\tau \eta P_s};$ similarly, the optimal energy harvesting time τ*=T.sub.1=T−T.sub.2 is obtained by using the same method.

6. The multi-carrier resource allocation method based on a wireless-powered backscatter communication network according to claim 5, characterized in that in step S5, x* and τ* are substituted into the original optimization problem, thus obtaining following time and power allocation optimization problem: $\underset{T_1,T_2,p_k,\alpha ,l_k}{\max }{T}_1\underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(1+\frac{\alpha p_k{h}_k}{{\stackrel{\_}{g}}_k}\right)+T_2\kappa \underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(K+l_k\right)$ $s.t.C4:\underset{k=1}{\overset{K}{.Math.}}{p}_k\le P^{\max },$ $C9:0\le \alpha \le 1-B/A,$ $C10:\frac{A\left(1-\alpha \right)-B}{{\stackrel{\_}{g}}_k{\mathrm{KT}}_2}\ge l_k.$ $\mathrm{where}{l}_k=\frac{x\tau \eta \left(1-\alpha \right)P_s}{{\stackrel{\_}{g}}_k{\mathrm{KT}}_2},$ l.sub.k>0 is an auxiliary variable, which represents a lower bound of the second term of the objective function, A=T.sub.1ηP.sub.s represents an auxiliary variable; and B=p.sub.eT.sub.1+p.sub.d T.sub.2 represents the total circuit power consumption.

7. The multi-carrier resource allocation method based on a wireless-powered backscatter communication network according to claim 6, characterized in that $\alpha \le 1-\frac{B+{\stackrel{\_}{g}}_k{\mathrm{KT}}_2{l}_k}{A}$ is obtained according to the constraint C10, so and C9 and C10 are combined into one constraint; the variable substitution L.sub.k=T.sub.2l.sub.k, p.sub.k=αp.sub.k, {tilde over (p)}.sub.k=T.sub.1p, {circumflex over (p)}.sub.k=T.sub.1p.sub.k is defined, substituted into the time and power allocation optimization problem, and converted into a convex optimization problem: $\underset{L_k,T_1,T_2,{\tilde{p}}_k,{\hat{p}}_k}{\max }{T}_1\underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(1+\frac{{\hat{p}}_k{h}_k}{{\stackrel{\_}{g}}_k{T}_1}\right)+T_2\kappa \underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(\stackrel{\_}{K}+\frac{L_k}{T_2}\right)$ $s.t.C4:\underset{k=1}{\overset{K}{.Math.}}{\hat{p}}_k\le T_1{P}^{\max },$ $C10:\eta \underset{k=1}{\overset{K}{.Math.}}{\hat{p}}_k{h}_k-\eta \underset{k=1}{\overset{K}{.Math.}}{\tilde{p}}_k{h}_k\ge L_k{\stackrel{\_}{g}}_{k}K+p_e{T}_1+p_d{T}_2.$

8. The multi-carrier resource allocation method based on a wireless-powered backscatter communication network according to claim 7, characterized in that according to the described time and power allocation optimization problem, a Lagrangian function is created using a Lagrange dual method: $L\left(\left\{L_k\right\},T_1,T_2,\left\{{\tilde{p}}_k\right\},\left\{{\hat{p}}_k\right\},\lambda ,\left\{{\beta }_k\right\}\right)=T_1\underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(1+\frac{{\tilde{p}}_k{h}_k}{{\stackrel{\_}{g}}_k{T}_1}\right)+T_2\kappa \underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(\stackrel{\_}{K}+\frac{L_k}{T_2}\right)+\lambda \left(T_1{P}^{\max }-\underset{k=1}{\overset{K}{.Math.}}{\hat{p}}_k\right)+\underset{k=1}{\overset{K}{.Math.}}{\beta }_k\left(\eta \underset{k=1}{\overset{K}{.Math.}}{\hat{p}}_k{h}_k-\eta \underset{k=1}{\overset{K}{.Math.}}{\tilde{p}}_k{h}_k-L_k{\stackrel{\_}{g}}_{k}K\right)-\underset{k=1}{\overset{K}{.Math.}}{\beta }_k\left(T_2{p}_d+T_1{p}_e\right)$ where β.sub.k and λ represent non-negative Lagrange multipliers of corresponding constraints; the Lagrange function is rewritten as: $L\left(\left\{L_k\right\},T_1,T_2,\left\{{\tilde{p}}_k\right\},\left\{{\hat{p}}_k\right\},\lambda ,\left\{{\beta }_k\right\}\right)=\underset{k=1}{\overset{K}{.Math.}}{L}_k\left(\left\{L_k\right\},T_1,T_2,\left\{{\tilde{p}}_k\right\},\left\{{\hat{p}}_k\right\},\lambda ,\left\{{\beta }_k\right\}\right)\mathrm{where}$ $L_k\left(\left\{L_k\right\},T_1,T_2,\left\{{\tilde{p}}_k\right\},\left\{{\hat{p}}_k\right\},\lambda ,\left\{{\beta }_k\right\}\right)=T_1{\log }_2\left(1+\frac{{\tilde{p}}_k{h}_k}{{\stackrel{\_}{g}}_k{T}_1}\right)+T_2\kappa {\log }_2\left(\stackrel{\_}{K}+\frac{L_k}{T_2}\right)-\lambda {\hat{p}}_k-{\beta }_k\left(T_2{p}_d+T_1{p}_e\right)+{\beta }_k\left(\eta \underset{k=1}{\overset{K}{.Math.}}{\hat{p}}_k{h}_k-\eta \underset{k=1}{\overset{K}{.Math.}}{\tilde{p}}_k{h}_k-L_k{\stackrel{\_}{g}}_{k}K\right)+\frac{\lambda T_1{P}^{\max }}{K}$ according to Karush-Kuhn-Tucker condition, optimal solutions are ${\tilde{p}}_k^{*}={T_1\left[\frac{1}{\ln 2{\beta }_k\eta h_k}-\frac{{\stackrel{\_}{g}}_k}{h_k}\right]}^{+},L_k^{*}=T_2\left[\frac{\kappa }{\ln 2{\beta }_k{\stackrel{\_}{g}}_{k}K}-\stackrel{\_}{K}\right],$ where [x].sup.+=max (0, x); {tilde over (p)}.sub.k* and L.sub.k* are substituted into the above Lagrange function, obtaining: ${\stackrel{\_}{L}}_k\left(T_2\left\{{\hat{p}}_k\right\},\lambda ,\left\{{\beta }_k\right\}\right)=-\left(T-T_2\right)\log \left(\ln 2\eta {\beta }_k{\stackrel{\_}{g}}_k\right)+T_2\kappa {\log }_2\left(\frac{\kappa }{\ln 2{\beta }_k{\stackrel{\_}{g}}_{k}K}\right)-\lambda {\hat{p}}_k-{\beta }_k\left(T_2{p}_d+\left(T-T_2\right)p_e\right)+\frac{\lambda \left(T-T_2\right)P^{\max }}{K}-\left(T-T_2\right)\left(\frac{K}{\ln 2}-\eta \underset{k=1}{\overset{K}{.Math.}}{\beta }_k{\stackrel{\_}{g}}_k\right)-\frac{T_2\kappa }{\ln 2}+T_2{\beta }_k{\stackrel{\_}{g}}_{k}K\stackrel{\_}{K}+{\beta }_k\eta \underset{k=1}{\overset{K}{.Math.}}{\hat{p}}_k{h}_k$ based on the gradient descent method, the parameters are iteratively updated, including $T_2\left(t+1\right)={\left[T_2\left(t\right)-{\Delta }_T\times \frac{\partial {\stackrel{\_}{L}}_k}{\partial T_2}\right]}^{+},T_1\left(t\right)=T-T_2\left(t\right)\mathrm{and}$ ${\hat{p}}_k\left(t+1\right)={\left[{\hat{p}}_k\left(t\right)-{\Delta }_p\times \frac{\partial {\stackrel{\_}{L}}_k}{\partial {\hat{p}}_k}\right]}^{+},\mathrm{where}$ $\frac{\partial {\stackrel{\_}{L}}_k}{\partial {\hat{p}}_k}={\beta }_k\eta h_k-\lambda ,$ $\frac{\partial {\stackrel{\_}{L}}_k}{\partial T_2}={\log }_2\left(\ln 2\eta {\beta }_k{\stackrel{\_}{g}}_k\right)+\kappa {\log }_2\left(\frac{\kappa }{K\ln 2{\beta }_k{\stackrel{\_}{g}}_k}\right)-{\beta }_k{p}_d+{\beta }_k{p}_e-\frac{\lambda P^{\max }}{K}-\frac{\kappa }{\ln 2}+{\beta }_{k}K\stackrel{\_}{K}{\stackrel{\_}{g}}_k+\frac{K}{\ln 2}-\eta \underset{k=1}{\overset{K}{.Math.}}{\beta }_k{\stackrel{\_}{g}}_{k}t$ represents iteration times, and Δ.sub.p and Δ.sub.T represent corresponding iteration step sizes; similarly, Lagrange multipliers β.sub.k and λ are updated by using the same method until convergence; according to the relationship between variables defined by the variable substitution, the optimal value ${\alpha }^{*}=\frac{{\tilde{p}}_k^{*}}{{\hat{p}}_k^{*}}$ of the backscatter coefficient and the optimal power allocation policy $p_k^{*}=\frac{{\tilde{p}}_k^{*}}{T_1^{*}{\alpha }^{*}}$ from the power station to the backscatter transmitter are solved; and p.sub.k* , α*, T.sub.1*, T.sub.2* and τ* are substituted into the previous derivation of the energy allocation coefficient and the transmit power of the backscatter transmitter to obtain x* and P.sub.k* .

Description

DESCRIPTION OF DRAWINGS

[0045] To enable the purpose, the technical solution, and the advantages of the present invention to be more clear, the present invention will be preferably described in detail below in combination with the drawings, wherein:

[0046] FIG. 1 is a flow diagram of the multi-carrier resource allocation method based on a wireless-powered backscatter communication network according to the present invention;

[0047] FIG. 2 is a schematic diagram of the system model of the present invention;

[0048] FIG. 3 is a curve chart showing the relationship between the energy harvesting time coefficient and the maximum transmit power threshold of the power station under different algorithms; and

[0049] FIG. 4 is a curve chart showing the relationship between the total data rate and the maximum transmit power threshold of the power station under different algorithms.

DETAILED DESCRIPTION

[0050] Embodiments of the present invention are described below through specific embodiments. Those skilled in the art can understand other advantages and effects of the present invention easily through the disclosure of the description. The present invention can also be implemented or applied through additional different specific embodiments. All details in the description can be modified or changed based on different perspectives and applications without departing from the spirit of the present invention. It should be noted that the figures provided in the following embodiments only exemplarily explain the basic conception of the present invention, and if there is no conflict, the following embodiments and the features in the embodiments can be mutually combined.

[0051] Wherein the drawings are only used for exemplary description, are only schematic diagrams rather than physical diagrams, and shall not be understood as a limitation to the present invention. In order to better illustrate the embodiments of the present invention, some components in the drawings may be omitted, scaled up or scaled down, and do not reflect actual product sizes. It should be understandable for those skilled in the art that some well-known structures and description thereof in the drawings may be omitted.

[0052] Same or similar reference signs in the drawings of the embodiments of the present invention refer to same or similar components. It should be understood in the description of the present invention that terms such as “upper”, “lower”, “left”, “right”, “front” and “back” indicate direction or position relationships shown based on the drawings, and are only intended to facilitate the description of the present invention and the simplification of the description rather than to indicate or imply that the indicated device or element must have a specific direction or constructed and operated in a specific direction, and therefore, the terms describing position relationships in the drawings are only used for exemplary description and shall not be understood as a limitation to the present invention; for those ordinary skilled in the art, the meanings of the above terms may be understood according to specific conditions.

[0053] As shown in FIG. 1, the resource allocation method based on a wireless-powered backscatter communication network provided by the present invention comprises following steps:

[0054] Step 1, as shown in FIG. 2, constructing a wireless-powered backscatter communication system, including a backscatter transmitter, an information receiver, and a power station, wherein the backscatter transmitter is equipped with an energy harvesting module and a backscatter circuit. Adopting a multi-carrier downlink transmit mode in order to improve the data transmission rate, and dividing the whole transmission process into two stages: in the first stage, the power station transmits radio-frequency signals to the backscatter transmitter, and the backscatter transmitter reflects some signals and converts remaining radio-frequency signals into energy for storage; and in the second stage, the backscatter transmitter decodes the reflected signals and uses harvested energy for data transmission and circuit power consumption.

[0055] Step 2: constructing an optimization model with the maximum total transmission rate, decomposing the complex optimization problem into multiple sub-problems, and finally converting the non-convex problems containing coupling variables into convex problems through variable substitution.

[0056] Step 3: jointly optimizing the transmit power, transmission time, backscatter coefficient and energy allocation coefficient by using the water-filling algorithm and the gradient descent method to obtain an optimal resource allocation policy.

[0057] The optimization process for solving resource allocation includes:

[0058] 1) Based on the minimum power consumption constraint and maximum transmit power constraint of a circuit, in consideration of the value range of power allocation, time coefficient, backscatter coefficient, and energy allocation coefficient, establishing an optimization problem taking a maximum total transmission rate of the system as an objective function.

[0059] 2) Decomposing an optimization sub-problem taking transmit power of the backscatter transmitter as an optimization variable from the original optimization problem, wherein the transmit power of the backscatter transmitter is limited by the energy obtained in the energy harvesting process, and obtaining the optimal value of the transmit power of the scattering transmitter by using the water-filling algorithm.

[0060] 3) Decomposing an optimization sub-problem taking the energy allocation coefficient of the backscatter transmitter as an optimization variable from the original optimization problem, wherein the backscatter transmitter uses a part of the harvested energy for circuit power consumption and the other part for data transmission; and obtaining the optimal values of the energy allocation coefficient and energy harvesting time.

[0061] To better understand the resource allocation method based on a wireless-powered backscatter communication network provided by embodiments of the present invention, the embodiments of the present invention will be described in detail. The method specifically comprises following steps:

[0062] S1. Constructing a wireless-powered backscatter communication system, i.e. a backscatter communication system including a backscatter transmitter, an information receiver, and a power station. Dividing the whole transmission process into two stages: in the first stage, the power station transmits radio-frequency signals to the backscatter transmitter, and the backscatter transmitter reflects some signals and harvests energy from the remaining signals; and in the second stage, the backscatter transmitter uses harvested energy for data transmission and circuit power consumption.

[0063] Using T.sub.1 and T.sub.2 to represent time spent in the two stages of the transmission process, T=T.sub.1+T.sub.2 presenting the time required for the whole transmission process. Dividing the total system bandwidth B into K orthogonal subcarriers, so the bandwidth of each subcarrier is B.sub.e=B/K. Defining a subcarrier set as ∀k∈K={1, 2, . . . , K}, assuming that each subcarrier experiences flat fading. Since the power station is far away from the information receiver, the interference therebetween is ignored.

[0064] In the first stage T.sub.1, the signals received by the backscatter transmitter are represented as

[00029]$y\left(t\right)=\underset{k=1}{\overset{K}{.Math.}}\sqrt{p_k{h}_k}{s}_k\left(t\right)+n\left(t\right),$

where p.sub.k represents transmit power allocated by the power station to the backscatter transmitter through subcarrier k, h.sub.k represents the channel power gain from the power station to the backscatter transmitter, s.sub.k (t) represents a transmit signals, n(t) represents a background noise and n(t)□CN (0,σ.sup.2).

[0065] Dividing the signals received by the backscatter transmitter into two parts by a backscatter coefficient α, where √{square root over (α)}y(t) is used for data backscatter, and √{square root over (1−α)}y(t) is used for energy harvesting; and in the stage T.sub.1, due to the limitation of the battery capacity of the backscatter transmitter, the time used for energy harvesting is represented as τ, τ≤T.sub.1. Therefore, the harvested energy is represented as:

[00030]$E=\tau \eta \left(1-\alpha \right)\underset{k=1}{\overset{K}{.Math.}}{p}_k{h}_k$

[0066] where η∈[0,1] represents an energy harvesting efficiency. The harvested energy is used to support circuit power consumption and information transmission in the stage T.sub.2. Assuming that the energy used for circuit power consumption is (1−x)E, the remaining energy xE is used for information transmission, where x represents an energy allocation coefficient, thereby obtaining:

(1−x)E≥p.sub.eT.sub.2+p.sub.dT.sub.2

[0067] where P.sub.e and p.sub.d respectively represent the circuit power consumption in the T.sub.1 and T.sub.2 time periods respectively, that is, the harvested energy used for circuit power consumption must be greater than or equal to the total power consumption of the circuit in the T.sub.1 and T.sub.2 periods.

[0068] Assuming that the backscatter transmitter can decode backscatter signals, defining the background noise power on each subcarrier as σ.sub.k=σ.sup.−2/K, so the backscatter data rate from the backscatter transmitter to the information receiver is

[00031]$R_b=T_1\stackrel{K}{\underset{k=1}{.Math.}}{B}_k{\log }_2\left(1+\frac{\alpha p_k{h}_k{g}_k}{{\sigma }_k}\right)$

[0069] where g.sub.k represents a channel power gain from the backscatter transmitter to the information receiver, B.sub.e represents the bandwidth of each subcarrier,

[0070] defining the transmit power of the backscatter transmitter on each subcarrier as P.sub.k in the stage T.sub.2 so the data rate of each subcarrier is

[00032]$R_k^h=B_e\kappa T_2{\log }_2\left(1+\frac{P_k{g}_k}{{\sigma }_k}\right)$

[0071] where κ∈[0,1] represents the transmission efficiency.

[0072] S2: According to circuit power consumption constraints and transmit power constraints, establishing a resource allocation optimization problem taking a maximum total transmission rate of the system as an objective function.

[0073] In order to improve the transmission efficiency of the entire system, a maximum total transmission rate of the information receiver is obtained by jointly optimizing the transmit power, the time allocation and the backscatter coefficient, and the resource allocation optimization problem is established as:

[00033]$\underset{T_1,T_2,p_k,P_k,x,\tau ,\alpha }{\max }R$$s.t.C1:\left(1-x\right)\tau \eta \left(1-\alpha \right)P_s\ge p_e{T}_1+p_d{T}_2,$$C2:T_2\underset{k=1}{\overset{K}{.Math.}}{P}_k\le x\tau \eta \left(1-\alpha \right)P_s,$$C3:T_1+T_2=T,$$C4:\underset{k=1}{\overset{K}{.Math.}}{p}_k\le P^{\max },$$C5:\tau \le T_1,$$C6:0\le \alpha \le 1,$$C7:0\le x\le 1,$$C8:p_k\ge 0,P_k\ge 0,T_1\ge 0,T_2\ge 0,\tau \ge 0.$

[0074] where P.sup.max represents the maximum transmit power of the power station,

[00034]$P_s=\underset{k-1}{\overset{K}{.Math.}}{p}_k{h}_k$

represents the total power of the signals received by the backscatter transmitter,

[00035]$R=R_b+\underset{k-1}{\overset{K}{.Math.}}{R}_k^h$

represents the total data rate, C1 represents the minimum circuit power consumption constraint, C2 represents the maximum transmit power constraint of the backscatter transmitter in the stage T.sub.2, and C4 represents the maximum transmit power constraint of the power station.

[0075] S3: According to the objective function and constraint conditions, decomposing an optimization sub-problem taking transmit power of the backscatter transmitter as a variable.

[0076] In order to find a closed form solution, according to the above optimization problem, it is observed that the transmit power P.sub.k of the backscatter transmitter is only restricted by C2, and an optimization sub-problem is decomposed and solved by using a water-filling algorithm, so the optimal P.sub.k* is found first and substituted into the optimization problem.

[0077] Consider the unit bandwidth over each subcarrier, an optimization sub-problem of the transmit power of the backscatter transmitter may be written as:

[00036]$\underset{P_k}{\max }{T}_2\kappa \underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(1+\frac{P_k{g}_k}{{\sigma }_k}\right)$$s.t.C2:T_2\underset{k=1}{\overset{K}{.Math.}}{P}_k\le x\tau \eta \left(1-\alpha \right)P_s.$

[0078] The above problem is a convex optimization problem since the constraint condition is linear constraint and the objective function is a concave function. g.sub.k=σ.sub.k/g.sub.k is defined, and the optimal transmit power obtained by using the water-filling algorithm is

[00037]$P_k={\left[\frac{1}{\lambda \ln 2T_2}-{\stackrel{\_}{g}}_k\right]}^{+}$

[0079] where [x].sup.+=max (0,x), λ is a Lagrange multiplier, and λ satisfies the following formula:

[00038]$T_2\underset{k=1}{\overset{K}{.Math.}}\left(\frac{1}{\lambda \ln 2T_2}-{\stackrel{\_}{g}}_k\right)=x\tau \eta \left(1-\alpha \right)P_s$$\lambda =\frac{K}{\ln 2\left(T_2\underset{k=1}{\overset{K}{.Math.}}{\stackrel{\_}{g}}_k+x\tau \eta \left(1-\alpha \right)P_s\right)}$

is obtained, and the transmit power P.sub.k is substituted, thus obtaining an optimal power allocation policy in the stage T.sub.2:

[00039]$P_k^{*}=\frac{\underset{k=1}{\overset{K}{.Math.}}{\stackrel{\_}{g}}_k}{K}+\frac{x\tau \eta \left(1-\alpha \right)P_s}{{\mathrm{KT}}_2}-{\stackrel{\_}{g}}_k$

[0080] If more harvested energy is used for data transmission, the transmit power P.sub.k* may be greater, and the total data rate may be greater. If the transmit power p.sub.k of the power station becomes larger, P.sub.k* also becomes larger. Since the backscatter transmitter has no energy storage device, the transmission rate is limited by the energy node (power station).

[0081] S4: Substituting optimal transmit power of the backscatter transmitter into an original problem, and similarly, decomposing a sub-problem taking an energy allocation coefficient as a variable from an original optimization problem.

[0082] P.sub.k* is substituted into the original optimization problem, and an optimization sub-problem with an energy coefficient is decomposed:

[00040]$\underset{x}{\max }{T}_1\underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(1+\frac{\alpha p_k{h}_k}{{\stackrel{\_}{g}}_k}\right)+T_2\kappa \underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(K+\frac{x\tau \eta \left(1-\alpha \right)P_s{g}_k}{{\stackrel{\_}{g}}_k{\mathrm{KT}}_2}\right)$$s.t.C1:\left(1-x\right)\tau \eta \left(1-\alpha \right)P_s\ge p_e{T}_1+p_d{T}_2,$$C7:0\le x\le 1.$$\mathrm{where}\stackrel{\_}{K}=\left(\underset{j=1}{\overset{K}{.Math.}}{g}_j/g_k\right)$

represents a constant, because the objective function of the above optimization sub-problem is a monotonically increasing function about the energy allocation coefficient x, according to the constraints C1 and C7, the value of x is

[00041]$x=\max \left(1,1-\frac{p_e{T}_1+p_d{T}_2}{\tau \eta \left(1-\alpha \right)P_s}\right).$

Because the right side of the inequality of the constraint C1 is greater than zero, if x=1, the constraint condition is not satisfied, so the optimal energy allocation coefficient is

[00042]$x^{*}=1-\frac{p_e{T}_1+p_d{T}_2}{\tau \eta \left(1-\alpha \right)P_s};$

and because x≥0, the backscatter coefficient α must satisfy

[00043]$\alpha \le 1-\frac{p_e{T}_1+p_d{T}_2}{\tau \eta P_s}.$

Similarly, the optimal energy harvesting time τ=T.sub.1=T−T.sub.2 may be obtained by using the same method.

[0083] S5: Converting non-convex problems containing coupling variables into convex problems using a variable substitution method, creating a Lagrangian function using a Lagrange dual method, obtaining an optimal solution form according to a KKT condition, and iteratively updating corresponding variables using a gradient descent method until convergence.

[0084] x* and τ* are substituted into the original optimization problem, thus obtaining the following time and power allocation optimization problem:

[00044]$\underset{T_1,T_2,p_k,\alpha ,l_k}{\max }{T}_1\underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(1+\frac{\alpha p_k{h}_k}{{\stackrel{\_}{g}}_k}\right)+T_2\kappa \underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(\stackrel{\_}{K}+l_k\right)$$s.t.C4:\underset{k=1}{\overset{K}{.Math.}}{p}_k\le P^{\max },$$C9:0\le \alpha \le 1-B/A,$$C10:\frac{A\left(1-\alpha \right)-B}{{\stackrel{\_}{g}}_k{\mathrm{KT}}_2}\ge l_k.$

[0085] where l.sub.k>0 is an auxiliary variable, which represents a lower bound of the second term of the objective function, A=T.sub.1ηP.sub.s represents an auxiliary variable; and B=p.sub.eT.sub.1+p.sub.aT.sub.2 represents the total circuit power consumption.

[00045]$\alpha \le 1-\frac{B+{\overline{g}}_{k}K{T}_2{l}_k}{A}$

can be obtained according to the constraint C10, so C9 and C10 can be combined into one constraint.

[0086] The variable substitution L.sub.k=T.sub.2l.sub.k, p.sub.k=αp.sub.k, {tilde over (p)}.sub.k=T.sub.1p.sub.k, {circumflex over (p)}.sub.k=T.sub.1p.sub.k is defined, substituted into the time and power allocation optimization problem, and converted into a convex optimization problem:

[00046]$\underset{L_k,T_1,T_2,{\tilde{p}}_k,{\hat{p}}_k}{\max }{T}_1\underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(1+\frac{{\tilde{p}}_k{h}_k}{{\stackrel{\_}{g}}_k{T}_1}\right)+T_2\kappa \underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(\stackrel{\_}{K}+\frac{L_k}{T_2}\right)$$s.t.C4:\underset{k=1}{\overset{K}{.Math.}}{\hat{p}}_k\le T_1{P}^{\max },$$C10:\eta \underset{k=1}{\overset{K}{.Math.}}{\hat{p}}_k{h}_k-\eta \underset{k=1}{\overset{K}{.Math.}}{\tilde{p}}_k{h}_k\ge L_k{\stackrel{\_}{g}}_{k}K+p_e{T}_1+p_d{T}_2.$

[0087] A Lagrangian function is created using a Lagrange dual method:

[00047]$L\left(\left\{L_k\right\},T_1,T_2,\left\{{\tilde{p}}_k\right\},\left\{{\hat{p}}_k\right\},\lambda ,\left\{{\beta }_k\right\}\right)=T_1\underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(1+\frac{{\tilde{p}}_k{h}_k}{{\stackrel{\_}{g}}_k{T}_1}\right)+T_2\kappa \underset{k=1}{\overset{K}{.Math.}}{\log }_2\left(\stackrel{\_}{K}+\frac{L_k}{T_2}\right)+\lambda \left(T_1{P}^{\max }-\underset{k=1}{\overset{K}{.Math.}}{\hat{p}}_k\right)+\underset{k=1}{\overset{K}{.Math.}}{\beta }_k\left(\eta \underset{k=1}{\overset{K}{.Math.}}{\hat{p}}_k{h}_k-\eta \underset{k=1}{\overset{K}{.Math.}}{\tilde{p}}_k{h}_k-L_k{\stackrel{\_}{g}}_{k}K\right)-\underset{k=1}{\overset{K}{.Math.}}{\beta }_k\left(T_2{p}_d+T_1{p}_e\right)$

[0088] where β.sub.k and λ represent non-negative Lagrange multipliers of corresponding constraints.

[0089] According to the Karush-kuhn-Tucker condition, the optimal solutions are

[00048]${\tilde{p}}_k^{*}={T_1\left[\frac{1}{\ln 2{\beta }_k\eta h_k}-\frac{{\stackrel{\_}{g}}_k}{h_k}\right]}^{+},L_k^{*}=T_2\left[\frac{\kappa }{\ln 2{\beta }_k{\stackrel{\_}{g}}_{k}K}-\stackrel{\_}{K}\right],$

where [x].sup.+=max (0, x). {tilde over (p)}.sub.k* and L.sub.k* are substituted into the above Lagrange function, obtaining:

[00049]${\stackrel{\_}{L}}_k\left(T_2,\left\{{\hat{p}}_k\right\},\lambda ,\left\{{\beta }_k\right\}\right)=-\left(T-T_2\right){\log }_2\left(\ln 2\eta {\beta }_k{\stackrel{\_}{g}}_k\right)+T_2\kappa {\log }_2\left(\frac{\kappa }{\ln 2{\beta }_k{\stackrel{\_}{g}}_{k}K}\right)-\lambda {\hat{p}}_k-{\beta }_k\left(T_2{p}_d+\left(T-T_2\right)p_e\right)+\frac{\lambda \left(T-T_2\right)P^{\max }}{K}-\left(T-T_2\right)\left(\frac{K}{\ln 2}-\eta \underset{k=1}{\overset{K}{.Math.}}{\beta }_k{\stackrel{\_}{g}}_k\right)-\frac{T_2\kappa }{\ln 2}+T_2{\beta }_k{\stackrel{\_}{g}}_{k}K\stackrel{\_}{K}+{\beta }_k\eta \underset{k=1}{\overset{K}{.Math.}}{\hat{p}}_k{h}_k$

[0090] Based on the gradient descent method, the parameters are iteratively updated, including

[00050]$T_2\left(t+1\right)={\left[T_2\left(t\right)-{\Delta }_T\times \frac{\partial {\stackrel{\_}{L}}_k}{\partial T_2}\right]}^{+},T_1\left(t\right)=T-T_2\left(t\right)\mathrm{and}$${\hat{p}}_k\left(t+1\right)={\left[{\hat{p}}_k\left(t\right)-{\Delta }_p\times \frac{\partial {\stackrel{\_}{L}}_k}{\partial {\hat{p}}_k}\right]}^{+},\mathrm{where}$$\frac{\partial {\stackrel{\_}{L}}_k}{\partial {\hat{p}}_k}={\beta }_k\eta h_k-\lambda ,$$\frac{\partial {\stackrel{\_}{L}}_k}{\partial T_2}={\log }_2\left(\ln 2\eta {\beta }_k{\stackrel{\_}{g}}_k\right)+\kappa {\log }_2\left(\frac{\kappa }{K\ln 2{\beta }_k{\stackrel{\_}{g}}_k}\right)-{\beta }_k{p}_d+{\beta }_k{p}_e-\frac{\lambda P^{\max }}{K}-\frac{\kappa }{\ln 2}+{\beta }_{k}K\stackrel{\_}{K}{\stackrel{\_}{g}}_k+\frac{K}{\ln 2}-\eta \underset{k=1}{\overset{K}{.Math.}}{\beta }_k{\stackrel{\_}{g}}_{k}t$

represents iteration times, and Δ.sub.p and Δ.sub.T represent corresponding iteration step sizes. Similarly, Lagrange multipliers β.sub.k and λ are updated by using the same method until convergence. According to the relationship between variables defined by variable substitution, the optimal value

[00051]${\alpha }^{*}=\frac{{\tilde{p}}_k^{*}}{{\hat{p}}_k^{*}}$

of the backscatter coefficient and the optimal power allocation policy

[00052]$p_k^{*}=\frac{{\tilde{p}}_k^{*}}{T_1^{*}{\alpha }^{*}}$

from the power station to the backscatter transmitter are solved. p.sub.k*, α* , T.sub.1* , T.sub.2τ* are substituted into the previous derivation of the energy allocation coefficient and the transmit power of the backscatter transmitter to obtain x* and P.sub.k*.

[0091] The performance of the system based on a wireless-powered backscatter communication network proposed in this embodiment is shown in FIGS. 3 and 4. The hybrid method proposed by the present invention is compared with backscatter communication (x=0) and energy harvesting communication (α=0).

[0092] FIG. 3 depicts a curve chart showing the relationship between the time coefficient in the energy harvesting stage and the maximum transmit power of the power station. If the maximum transmit power threshold of the power station increases, the time duration of energy harvesting becomes longer, because the longer the energy harvesting time, that is, the longer the stage T.sub.1, the more backscatter signals the information receiver receive, and the more the energy is used for data transmission in the stage T.sub.2, thereby further increasing the overall data rate.

[0093] FIG. 4 depicts a curve chart showing the relationship between the total data rate and the maximum transmit power of the power station. The total data rate increases as the maximum transmit power threshold of the power station increases. Compared with the other two algorithms, the algorithm of the present invention has the best performance, while the energy harvesting communication mode has the worst performance. Larger transmit power may be allocated to each subcarrier, and more energy may be harvested for data transmission in the stage T.sub.2, thereby increasing the user rate.

[0094] Finally, it should be noted that the above embodiments are only used for describing, rather than limiting the technical solution of the present invention. Although the present invention is described in detail with reference to the preferred embodiments, those ordinary skilled in the art shall understand that the technical solution of the present invention can be amended or equivalently replaced without departing from the purpose and the scope of the technical solution. The amendment or equivalent replacement shall be covered within the scope of the claims of the present invention.