A robust resource allocation method for backscatter communication system facing eavesdropping users

Abstract

The present application belongs to the field of Internet of Things, and particularly relates to a robust resource allocation method for a backscatter communication system for eavesdropping users, comprising: establishing a backscatter communication system for eavesdropping users; deploying an information receiver, a backscatter device and an eavesdropping user in the backscatter communication system, wherein each backscatter device has a backscatter information function, and the backscatter device is equipped with an energy collection rectifier circuit; constructing a robust resource allocation model based on channel uncertainty and maximum rate; converting the resource optimization model into a convex optimization problem based on a block coordinate descent method; solving the optimization problem by using a convex optimization toolbox to obtain a base station beamforming vector, an artificial noise vector and a backscatter device time for backscattering, i.e. obtaining a resource allocation scheme; the method can effectively curb the influence of eavesdropping users on the system while ensuring the maximization of system and throughput, and meets the service quality requirements of users.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of Internet of Things, and particularly relates to a robust resource allocation method for a backscatter communication system for eavesdropping users. BACKGROUND

[0002] To meet the development needs of low-power Internet of Things technology industry, backscatter communication technology as a new technology has attracted extensive attention and research in the academic and industrial circles. Backscatter communication is well-known for its low power consumption, low cost deployment and other characteristics, and the combination of backscatter communication and Internet of Things technology can significantly improve the system performance. In the Internet of Things system based on backscatter communication, the reflection node can use the environmental electromagnetic wave signal for energy collection and at the same time realize the reflection transmission of the received signal.

[0003] Although many works have greatly promoted the development of backscatter communication resource allocation technology, the existing works only consider single-antenna signal transmission, which cannot further improve the system capacity like multi-antenna technology; in addition, due to the passive characteristics of the reflection signal, it is easy to be attacked by the network or eavesdropped by the user, so that the security of the above system information is greatly discounted; finally, due to the influence of channel feedback delay, nonlinear circuit of reflection node and other factors, the resource allocation based on perfect channel state information is too ideal, which may cause a large interruption event. SUMMARY

[0004] In order to guarantee the security performance of the system and maximize the sum throughput of the system under the condition of channel uncertainty, the application provides a robust resource allocation method for a backscatter communication system for eavesdropping users, which comprises the following steps:

[0005] S1: establishing a backscatter communication system for eavesdropping users;

[0006] S2: deploying an information receiver, a backscatter device and an eavesdropping user in the backscatter communication system, wherein each backscatter device has the function of backscattering information, and the backscatter device is equipped with an energy collection rectifier circuit;

[0007] S3: constructing a sum rate maximum robust resource allocation model based on channel uncertainty;

[0008] S4: based on the block coordinate descent method, converting the resource optimization model into a convex optimization problem;

[0009] S5: using a convex optimization toolbox to solve the optimization problem to obtain the beamforming vector of the base station, the artificial noise vector and the time used for backscattering of the backscatter device, that is, to obtain the resource allocation scheme.

[0010] Further, the construction of the sum rate maximum robust resource allocation model based on channel uncertainty comprises the following steps:

[0011] S31: Establish a resource allocation model with maximum system throughput based on time constraints, energy collection constraints and base station transmission power constraints;

[0012] S32: On the basis of the resource allocation model with maximum system throughput, consider the impact of eavesdropping users on the system, and optimize the resource allocation model into a resource allocation model with secrecy rate constraints;

[0013] S33: On the basis of the resource allocation model with secrecy rate constraints, consider the impact of channel uncertainty on the system, and establish a robust resource allocation model based on bounded spherical channel uncertainty.

[0014] Further, the resource allocation model with maximum system throughput is represented as:

[0015]

[0016] Constraint condition C1:

[0017] C2:

[0018] C3:

[0019] C5: W k ≥0, Z≥0

[0020] C6: Rank(W k )=1

[0021] Where W k is the kth beamforming covariance matrix; Z is the artificial noise covariance matrix; α k is the backscattering time of the kth backscattering device; K is the total number of backscattering devices; R k is the throughput of the information receiving end in the kth backscattering time; Tr(·) represents the trace of the matrix; P max is the maximum transmission power threshold of the base station; T is the transmission frame length of the system; is the absorption power at the kth backscattering device; represents the received power of the kth backscattering device in the backscattering time; represents the received power of the kth backscattering device in other time except the backscattering time; is the circuit energy consumption at the kth backscattering device; Rank(·) represents the rank of the matrix; ≥ represents the matrix semidefinite symbol.

[0022] Further, the resource allocation model with secrecy rate constraints is represented as:

[0023]

[0024] Constraint: C1:

[0025] C2:

[0026] C3:

[0027] C4:

[0028] C5: W k ≥ 0, Z ≥ 0

[0029] C6: Rank(W k ) = 1

[0030] where, r k denotes the rate of information receiver in the kth backscattering time; denotes the rate of eavesdropper in the kth backscattering time; denotes the system minimum secrecy rate threshold in the kth backscattering time.

[0031] Further, the robust resource allocation model is represented as:

[0032]

[0033] Constraint: C1:

[0034] C2:

[0035] C3:

[0036] C4:

[0037] C5: W k ≥ 0, Z ≥ 0

[0038] C6: Rank(W k ) = 1

[0039] C7:

[0040] where, Δf denotes the estimation error of channel coefficient between base station and eavesdropper; denotes the uncertain channel set.

[0041] Further, the uncertain channel set considers the bounded spherical channel uncertainty model to be constructed, represented as:

[0042]

[0043] wherein, denotes an estimate of the channel coefficient, Δf denotes an estimation error of the channel coefficient; ||·|| denotes a norm of a vector; ε denotes an upper bound of the channel estimation error.

[0044] Further, in solving the robust resource allocation model, the constraint containing uncertainty is optimized based on variable relaxation method, and the optimized robust resource allocation model is expressed as:

[0045]

[0046] Constraint condition: C1:

[0047] C2:

[0048] C3:

[0049]

[0050]

[0051]

[0052]

[0053]

[0054]

[0055]

[0056] C5: W k ≥ 0, Z ≥ 0,

[0057] C6: Rank(W k ) = 1

[0058] wherein, v k , θ k , γ k , ω k , χ, y 2 , λ1, λ2 are relaxation variables; are respectively values of the last iteration; denotes an estimate of the channel coefficient; ε denotes an upper bound of the channel estimation error. denotes a noise power; I is a unit matrix; Λ, Θ denote sets of relaxation variables, and

[0059] Further, the optimized robust resource allocation model is split into a beamforming and artificial noise component optimization problem and a time optimization problem, including:

[0060] By fixing the time factor, the beamforming and artificial noise component optimization problem is expressed as:

[0061]

[0062] Constraint: C1:

[0063] C3:

[0064]

[0065]

[0066]

[0067]

[0068]

[0069]

[0070]

[0071] C5: W k ≥ 0, Z ≥ 0

[0072] C6: Rank(W k ) = 1

[0073] After relaxing C6 based on the semi-definite relaxation method, the beamforming and artificial noise component optimization problem is solved by the CVX toolbox;

[0074] By fixing the beamforming and artificial noise component covariance matrix, the time optimization problem is expressed as:

[0075] Find a k

[0076] Constraint: C2:

[0077] C3:

[0078] The time optimization problem is solved by the CVX toolbox.

[0079] Further, the throughput R of the information receiving end in the kth backscattering timek is represented as:

[0080]

[0081] wherein h k represents the channel gain from the kth backscatter device to the information receiver; β k represents the power reflection coefficient at the kth backscatter device; represents the channel vector from the base station to the kth backscatter device, and M represents the number of base station antennas; represents the channel vector from the base station to the information receiver; represents the kth information beamforming vector of the base station; represents the artificial noise vector of the base station, which conforms to a Gaussian distribution with a mean of 0 and a variance of Z; represents the noise power of the kth backscatter time information receiver.

[0082] The method of the present application effectively suppresses the influence of eavesdropping users on the system while ensuring system and throughput maximization, and meets the service quality requirements of users. BRIEF DESCRIPTION OF DRAWINGS

[0083] Figure 1 is a system model of the present application;

[0084] Figure 2 is a solution flowchart of the method of the present application;

[0085] Figure 3 is a comparison chart of system outage probability of the method of the present application and different conventional methods under different channel state information. DETAILED DESCRIPTION

[0086] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0087] The present application proposes a robust resource allocation method for a backscatter communication system facing eavesdropping users, as shown in Figure 2 , comprising:

[0088] S1: establishing a backscatter communication system facing eavesdropping users;

[0089] S2: deploying an information receiver, a backscatter device and an eavesdropping user in the backscatter communication system, wherein each backscatter device has the function of backscattering information, and the backscatter device is equipped with an energy collection rectifier circuit;

[0090] S3: Construct a channel uncertainty based robust resource allocation model with maximum sum rate;

[0091] S4: Based on the block coordinate descent method, the resource optimization model is converted into a convex optimization problem;

[0092] S5: Use the convex optimization toolbox to solve the optimization problem to obtain the base station beamforming vector, artificial noise vector and backscattering device time for backscattering, that is, to obtain the resource allocation scheme.

[0093] In this embodiment, the process of constructing a resource allocation model based on channel uncertainty includes:

[0094] S31: Based on the time constraint, the energy collection constraint and the base station transmission power constraint, a resource allocation model with maximum system throughput is established;

[0095] S32: On the basis of the resource allocation model with maximum system throughput, considering the influence of the eavesdropping user on the system, the resource allocation model is optimized to a resource allocation model with a secrecy rate constraint;

[0096] S33: On the basis of the resource allocation model with a secrecy rate constraint, considering the influence of channel uncertainty on the system, a robust resource allocation model is established based on bounded spherical channel uncertainty.

[0097] The system model of this embodiment is as shown in Figure 1 , in which a base station with M antennas serves a single antenna receiver, and K backscattering devices assist transmission, There is a single antenna eavesdropper with uncertain location in the network, which obtains information by eavesdropping on the base station transmission signal. It is assumed that K backscattering devices are deployed near the information receiving end, and there is an obstacle between the eavesdropper and the backscattering devices, and the backscattering signal received by the eavesdropper is weak, so the backscattering signal at the eavesdropper can be ignored; K backscattering devices are powered by the signals transmitted by the base station, and at the same time, the wireless information is reflected to the information receiving end through time division multiple access; define the transmission frame length of the system as T, as shown in Figure 1 , the backscattering time of the kth backscattering device is α k , the energy collection time is (1-α k ), and the total transmission time satisfies

[0098] The base station transmits beam information in the form of time division multiple access, and the signal of each time slot contains N symbol periods. Therefore, the transmission signal of the base station in the system time frame can be expressed as:

[0099]

[0100] wherein, and respectively represent the k-th information beamforming vector and artificial noise vector of the base station, and the artificial noise vector satisfies a complex Gaussian distribution with mean 0 and variance Z, b(n) represents a symbol stream carrying confidential information, Therefore, the signal sent by the base station to the k-th backscatter device is w k b(n) + z.

[0101] Since the transmission rate of the backscatter device is much lower than that of the radio frequency source, without loss of generality, it is assumed that the symbol period of the backscatter device is N times that of the radio frequency source. Assuming that c is the transmission signal within N radio frequency source symbol periods, and is a random variable satisfying mean 0 and unit variance. The received signals of the information receiving end and the eavesdropper in the k-th time slot can be respectively represented as:

[0102]

[0103]

[0104] wherein, y k (n) represents the received signal of the information receiving end in the k-th time slot, represents the received signal of the eavesdropper in the k-th time slot; respectively represent the channel vectors between the base station and the k-th backscatter device, between the base station and the information receiving end, and between the base station and the eavesdropper, h k represents the channel gain from the k-th backscatter device to the information receiving end; β k represents the power reflection coefficient at the k-th backscatter device, represents the additive white Gaussian noise with mean 0 and variance at the k-th information receiving end in the k-th time slot, represents the additive white Gaussian noise with mean 0 and variance at the eavesdropper.

[0105] The backscatter link usually experiences more attenuation than the main link, therefore, according to the decoding strategy of SIC, the information receiving end first decodes b(n), and then removes b(n) from the received signal before detecting c. Therefore, the signal detected by the information receiving end in the k-th time slot is:

[0106]

[0107] wherein, v s is the combination coefficient of the information receiving end with respect to b(n). For detection, the transmission signal b(n) can be regarded as passing through an equivalent channel Since c is unknown, the information receiver does not know the information about the equivalent channel. Therefore, non-coherent detection is used to detect b(n). Under the premise of decoding c, the signal-to-noise ratio of the information receiver about b(n) in the kth time slot is:

[0108]

[0109] Without loss of generality, it is assumed that the combination coefficients are normalized, i.e., ||v s || 2 = 1. In the case of block fading channels and sufficiently large transmission paths, the capacity of non-coherent detection is consistent with that of coherent detection. For slowly changing channels and sufficiently large N, the throughput of the information receiver and the eavesdropper in the kth backscattering time can be expressed as:

[0110]

[0111]

[0112] where r k and represent the rates of the information receiver and the eavesdropper in the kth backscattering time, respectively. Therefore, the secrecy rate in the kth time slot is:

[0113]

[0114] The backscattering device simultaneously collects energy using the backscattering time and the time after backscattering. Considering the non-linear energy collection model, the absorbed power at the kth backscattering device is:

[0115]

[0116] where B represents the maximum absorption power of the energy collection circuit; κ and μ represent the characteristics of the energy collection circuit, such as resistance and diode turn-on voltage; represents the received power of the kth backscattering device. Define the received power of the kth backscattering device in the backscattering time and the idle time as P k IN and Therefore, it can be expressed as:

[0117]

[0118]

[0119] where is the covariance matrix of the beamforming vector, and satisfies Rank(W k ) = 1, and Z is an artificial noise covariance matrix. Let be the circuit power consumption at the k-th backscattering device. Then, the energy collected by the k-th backscattering device within a time frame should be greater than the circuit power consumption, i.e.:

[0120]

[0121] Due to the inherent randomness and channel feedback delay of wireless communication systems, perfect channel state information for some links is difficult to obtain. Since the locations of the base station, backscattering device, and information receiver are fixed, based on channel estimation theory, it is assumed that perfect channel state information for the links from the base station to the backscattering device, from the base station to the information receiver, and from the backscattering device to the information receiver can be obtained. For eavesdropping links, since the location of the eavesdropping user is not fixed and their location state occurs randomly, this invention considers the channel uncertainty of the channel link between the base station and the eavesdropper to overcome the influence of channel uncertainty. This invention considers a bounded spherical channel uncertainty model, which yields:

[0122]

[0123] in, Δf represents the estimated value of the channel coefficients, and Δf represents the estimation error of the channel coefficients. This represents a set of uncertain channels.

[0124] The objective of this invention is to maximize the throughput of the information receiver under constraints such as the maximum transmit power of the base station, time allocation constraints, energy constraints of the backscattering device, and security constraints. Therefore, under imperfect channel state information, this optimization problem can be represented by jointly optimizing the beamforming matrix, the covariance matrix of the artificial noise components, and the time factor. Thus, it can be formulated as the following uncertain optimization problem:

[0125] P1:

[0126] stC1:

[0127] C2:

[0128] C3:

[0129] C4:

[0130] C5:W k ≥0, Z≥0

[0131] C6:Rank(W k ) = 1

[0132] C7:

[0133] wherein the throughput of the information receiving end in the kth backscattering time is:

[0134]

[0135] wherein k represents any one backscattering device in the model, K is the total number of backscattering devices; R k is the throughput of the information receiving end in the kth backscattering time; W k is the kth beamforming covariance matrix; Z is an artificial noise covariance matrix; P max is the maximum transmit power threshold of the base station; a k is the backscattering time of the kth backscattering device; T is the transmission frame length of the system; P k IN and are the received power of the kth backscattering device in the backscattering time and the rest of the time, respectively; P k EH is the absorption power at the kth backscattering device; is the circuit energy consumption at the kth backscattering device; Rank(·) represents the rank of the matrix; r k and respectively represent the rate of the information receiving end and the eavesdropper in the kth backscattering time; represents the minimum secrecy rate threshold of the system in the kth backscattering time; Δf represents the estimation error of the channel coefficient; represents an uncertain channel set; in the above resource allocation model with the maximum system throughput, the constraint C1 is the maximum transmit power constraint of the base station, C2 is the time constraint, C3 is the energy constraint at the backscattering device, the energy collected by the device is greater than the consumed energy of the circuit, C4 is the minimum secrecy rate constraint, C5 is the matrix semi-positive definite constraint, C6 is the rank one constraint, and C7 contains the uncertain parameter set.

[0136] For the constraint C5 containing uncertainty and the non-convex objective function, based on the variable relaxation method, the following inequalities are further obtained:

[0137]

[0138]

[0139]

[0140]

[0141]

[0142]

[0143] where v k , θ k , γ k , ω k , χ and are the slack variables. Thus, the constraint C5 can be relaxed as The objective function can be equivalent to

[0144] Obviously, and are still non-convex constraints, by continuous convex approximation method and first order Taylor expansion, can be approximated as:

[0145]

[0146]

[0147]

[0148] where and are the slack variables. and are the values of and in the last iteration. Similarly, can be approximated as:

[0149]

[0150]

[0151]

[0152] where y 2 and are the slack variables. and are the values of and in the last iteration.

[0153] For the uncertainty constraints and we next transform them into deterministic matrix linear inequalities using the S- procedure method in Lemma 1.

[0154] Lemma 1 S-Procedure (S-Procedure): Definition where is a Hermitian matrix, and x represents an N x 1 dimensional vector, N represents the dimension of a vector or matrix can be from iff there exists λ ≥ 0 such that the following linear matrix inequality holds:

[0155]

[0156] Based on Lemma 1, the uncertainty constraint and can be transformed into

[0157]

[0158]

[0159] where λ1and λ2are the slack variables, ω k denotes the slack variable; ε denotes the upper bound of the channel estimation error; denotes the noise power; I is a unit matrix.

[0160] Definition and Based on the above transformation, the following deterministic optimization problem can be obtained:

[0161] P2:

[0162] s.t.C1:

[0163] C2:

[0164] C3:

[0165]

[0166]

[0167]

[0168]

[0169]

[0170]

[0171] C5:W k ≥ 0, Z ≥ 0, C6: Rank(W k ) = 1

[0172] P2 is still a non-convex optimization problem, there are still coupled variables in the objective function and C3, and there is a non-convex rank-one constraint in C6, so it is still difficult to solve.

[0173] For deterministic optimization problem P2, it is still not possible to solve directly by convex optimization toolbox. Based on the research of alternating optimization theory, the present application proposes a robust resource allocation iterative algorithm based on block coordinate descent method. P2 can be decomposed into two sub-problems, which are: 1) beamforming and artificial noise component optimization problem; and 2) time optimization problem, which specifically includes:

[0174] By fixing the time factor, the following sub-problem can be obtained:

[0175] P2-A:

[0176] s.t.C1:

[0177] C3:

[0178]

[0179]

[0180]

[0181]

[0182]

[0183]

[0184] C5:W k ≥0,Z≥0,C6:Rank(W k )=1

[0185] It can be known that the above sub-problem is still non-convex, based on the semi-definite relaxation method, C6 can be relaxed, so that the relaxed version of the above problem is obtained, and it can be solved by CVX toolbox. If the optimal beamforming matrix obtained satisfies Rank(W k )=1, the optimal beamforming vector If the rank-one constraint is not satisfied, a feasible solution can be obtained by Gaussian randomization method.

[0186] By fixing the beamforming and artificial noise component covariance matrix, the time optimization problem can be expressed as:

[0187] P2-B:Findα k

[0188] s.t.C2,C3

[0189] P2-B is a convex optimization problem, which can be solved by CVX toolbox.

[0190] The application effect of the present application is described in detail below in combination with simulation.

[0191] 1) Simulation conditions

[0192] This section verifies the convergence and effectiveness of the algorithm through simulation results. It is assumed that there is one base station, two backscatter devices, one information receiver and one eavesdropper in the system. The base station is located at the origin, the information receiver is located at (10, 0), and the two backscatter devices are located on a circle with a radius of 7 m and the information receiver as the center. It is assumed that the eavesdropper is located at (-5, 0). The system channel fading model considers large-scale fading, and it is assumed that each channel fading model of the system is independent and satisfies distribution, ζ = -30 dB represents path loss, the path loss exponent a is 3, d i represents the distance between any two devices. It is assumed that the upper bound of channel uncertainty is [0, 0.09], and the noise power at the information receiver and the eavesdropper is equal In order to verify the superiority of the algorithm of the present application, the algorithm of the present application is compared with the traditional non-robust algorithm, and other parameters are: K = M = 2, P max = 1 W, β k = 0.2, T = 1 s, represents the iteration accuracy. When the iteration difference is less than the iteration accuracy, the iteration is exited.

[0193] 2) Simulation results

[0194] In this embodiment, the proposed robust resource allocation method for a backscatter communication system facing an eavesdropping user is compared with a non-robust method, and the other parameters are: Figure 3 It can be seen that when the channel uncertainty is small, the outage probability of the present application method compared with the traditional non-robust method and the benchmark method is closer to 0, because the present application method has a higher degree of freedom and consumes more degrees of freedom to ensure strong robustness. The traditional non-robust method and the benchmark method do not take into account the influence of the auxiliary role of the backscatter device and the channel estimation error on the system secrecy throughput, so when the channel uncertainty increases, the traditional non-robust method and the benchmark method show a high outage probability. It is indirectly proved that the present application method has strong robustness compared with the traditional non-robust method.

[0195] Although the embodiments of the present application have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, replacements and variations of these embodiments can be made without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A method for robust resource allocation in a backscatter communication system for eavesdropper users, characterized in that, Specifically comprising the following steps: S1: Establish a backscatter communication system for eavesdropping users; S2: Deploy information receivers, backscatter devices, and eavesdropping users in the backscatter communication system, wherein each backscatter device has a backscatter information function, and the backscatter device is equipped with an energy harvesting rectifier circuit; S3: Based on channel uncertainty, construct a sum-rate maximization robust resource allocation model; S4: Based on the block coordinate descent method, convert the sum-rate maximization robust resource allocation model into a convex optimization problem, specifically comprising: During the solving of the robust resource allocation model, based on the variable relaxation method, optimize the constraint containing uncertainty, and the optimized robust resource allocation model is represented as: in, For the first Beamforming covariance matrix; The covariance matrix of the artificial noise; For the first Backscattering time of each backscattering device; This represents the total number of backscattering devices; This indicates finding the trace of a matrix; This is the maximum transmit power threshold for the base station; The system's transmission frame length; For the first Absorbed power at each backscattering device; Indicates the first The received power of a backscattering device during the backscattering time; Indicates the first The received power of a backscattering device at times other than the backscattering time; For the first Circuit power consumption at the backscattering device; Indicates the first The minimum security rate threshold of the system within a backscattering time; , , , , , , , , , , , , , These are slack variables; , , , They are respectively , , , The value of the previous iteration; This represents the estimated value of the channel coefficients; This represents the upper bound of the channel estimation error; Represents noise power; I is an identity matrix; Let represent the set of slack variables, and , ; Indicates the first Channel gain from a backscattering device to the information receiver; Indicates the first Power reflection coefficient at each backscattering device; Hk represents a channel vector between the base station and the kth Hk represents a channel vector between the base station and the information receiving terminal; σk represents a variance of the additive white Gaussian noise at the information receiving terminal in the kth rank (Hk) represents a rank of the matrix; Hk represents a channel vector between the base station and the information receiving terminal;​​ Split the optimized robust resource allocation model into a beamforming and artificial noise component optimization problem and a time optimization problem, wherein the beamforming and artificial noise component optimization problem is represented by fixing the time factor as: Based on the semi-positive definite relaxation method, relax C6, and solve the beamforming and artificial noise component optimization problem through the CVX toolbox; By fixing the beamforming and artificial noise component covariance matrix, the time optimization problem is represented as: Solve the time optimization problem through the CVX toolbox; S5: Obtain the beamforming vector of the base station, the artificial noise vector, and the time for backscattering of the backscatter device, that is, obtain the resource allocation scheme.

2. The method of claim 1, wherein, Constructing the sum-rate maximization robust resource allocation model based on channel uncertainty comprises: S31: Based on the time constraint, the energy harvesting constraint, and the base station's transmit power constraint, establish a resource allocation model that maximizes system throughput; S32: On the basis of the resource allocation model that maximizes system throughput, consider the influence of eavesdropping users on the system, and optimize the resource allocation model into a resource allocation model with a secrecy rate constraint; S33: On the basis of the resource allocation model with a secrecy rate constraint, consider the influence of channel uncertainty on the system, and based on bounded spherical channel uncertainty, establish a robust resource allocation model.

3. The method of claim 2, wherein, The resource allocation model that maximizes system throughput is represented as: wherein, is the throughput of the information receiving end for the nth backscattering time. backscattering time.

4. The method of claim 2, wherein, The resource allocation model with a secrecy rate constraint is represented as: wherein, is the throughput of the information receiving end for the th backscatter time; denotes the rate of the information receiving end for the th backscatter time; denotes the rate of the eavesdropper for the th backscatter time.

5. The method of claim 2, wherein, The robust resource allocation model is represented as: wherein, denotes the estimation error of the channel coefficient between the base station and the eavesdropper; denotes the set of uncertain channels; is the throughput of the information receiving end in the th backscattering time; denotes the rate of the information receiving end in the th backscattering time; denotes the rate of the eavesdropper in the th backscattering time.

6. The method of claim 5, wherein, The uncertain channel set considers the bounded spherical channel uncertainty model to be constructed and is represented as: wherein denotes an estimate of a channel coefficient; denotes a norm of a vector; denotes an upper bound of a channel estimation error.

7. A method of robust resource allocation for a backscatter communication system facing eavesdroppers according to claim 3, 4 or 5, characterized in that, The first backscattering time information receiving end throughput is expressed as: wherein, denotes the channel gain of the first backscattering device to the information receiving end; denotes the power reflection coefficient at the first backscattering device; denotes the channel vector of the base station to the first backscattering device, M denotes the number of base station antennas; denotes the channel vector of the base station to the information receiving end; denotes the first information beamforming vector of the base station; denotes the artificial noise vector of the base station, which conforms to a Gaussian distribution with a mean of 0 and a variance of Z; denotes the noise power of the first backscattering time information receiving end.

Images