IRS-assisted maximum sum rate resource optimization method for cognitive SWIPT system

Abstract

The present invention discloses a method for optimizing the maximum sum rate resources of an IRS-assisted cognitive SWIPT system. With maximizing the end-to-end sum rate of the IRS-assisted cognitive SWIPT system as the optimization goal, the cognitive SWIPT receiver adopts nonlinear energy collection. The multi-resource allocation problem is modeled as a nonlinear non-convex optimization problem, which is mutually coupled with the subcarrier, power, and IRS reflection phase shift vector. Therefore, an alternating optimization method is adopted to fix the IRS phase shift vector, and the Lagrange dual transformation and subgradient method are used to obtain the optimal power and subcarrier set for information decoding and nonlinear energy collection. Then, the IRS reflection phase shift vector is obtained by a continuous convex approximation method, thereby maximizing the end-to-end sum rate of the IRS-assisted cognitive SWIPT system. The present invention effectively allocates multiple resources such as subcarriers, power, and IRS reflection phase shift vectors for information decoding and nonlinear energy collection in the cognitive SWIPT system, and maximizes the end-to-end sum rate under the actual SWIPT nonlinear energy collection model.

Description

Technical Field

[0001] The present invention belongs to the field of information and communication engineering technology, and proposes a multi-resource joint optimization method for maximizing the sum rate in a cognitive wireless information and power transfer (SWIPT) system assisted by an intelligent reflecting surface (IRS). The method effectively allocates multiple resources such as subcarriers, power, and IRS phase shift vectors, and achieves system sum rate maximization under actual nonlinear energy harvesting conditions. Background Art

[0002] The implementation of 5G has led to the rapid development of intelligent mobile multimedia terminals and the global Internet of Things industry. The explosive demand for communication services has increased equipment energy consumption requirements. The increase in communication network node density and the expansion of network coverage have led to increasing attention to the energy consumption of communication networks. "Greening" has become one of the research directions for future wireless networks. Energy-efficient (EE) green communications and ubiquitous intelligence (AI) are cutting-edge technologies for future 6G mobile communication systems and key technologies for energy conservation, emission reduction, and environmental protection in future green mobile communication networks. Smart Reflecting Surface (IRS)-assisted Wireless Power Transmission (SWIPT), as one of the high-efficiency, low-power green communication technologies, has been proposed to address the energy efficiency issues and improve communication quality of future 6G mobile communication systems, and has received widespread attention in the industry.

[0003] Cognitive radio (CR) offers a new solution to the conflict between scarce spectrum resources and low spectrum utilization. SWIPT significantly improves network energy efficiency by harvesting energy (EH) from radio frequency signals. Combining the two, cognitive SWIPT leverages CR's opportunistic use of spectrum resources and SWIPT's simultaneous signal-energy transmission to improve both the spectrum and energy efficiency of wireless networks. The goal of SWIPT's multi-user, multi-resource joint optimization is to optimize system resources through the joint allocation and scheduling of multiple resources among different users. Compared to conventional wireless network multi-user scheduling, the joint allocation and scheduling of multiple resources in SWIPT requires a trade-off between information and energy. Currently, allocable resources in SWIPT include time slots, power, bits, and subcarriers. Technical indicators for SWIPT network resource allocation and optimization primarily include sum rate (spectrum efficiency), energy efficiency, and outage probability.

[0004] An intelligent reflecting surface (IRS) is a planar array composed of multiple reconfigurable passive reflective elements, whose operating modes are coordinated by a software controller. Each element can control the reflection angle and intensity of the incident electromagnetic wave, thereby controlling the phase and amplitude of the reflected signal. This allows the reflected electromagnetic wave to independently produce phase shifts, forming a three-dimensional passive beam that meets differentiated communication needs. Each IRS element is connected to a controller for unified control, allowing the intelligent reflecting surface to reconfigure the wireless propagation environment, thereby increasing the degree of freedom of wireless communication network performance. The controller adjusts the phase shift of each element in real time to reflect the signal. This transmitted signal can enhance the received power at the receiving end while weakening the received power of eavesdroppers, thereby improving system security. For example, in wireless communication networks, IRSs can be deployed to intelligently coordinate reflections to address fading and interference issues in wireless channels. For joint uplink and downlink communications, a passive IRS tends to achieve higher weighted sum rates than an active IRS, optimizing the IRS placement separately, because the optimal active IRS placement requires balancing the rate performance of the uplink and downlink, while deploying a passive IRS near the transmitter or receiver is optimal for both uplink and downlink.

[0005] Compared to traditional relay base stations, the intelligent reflector is a passive device with the notable advantage of low power consumption. The user's received signal-to-noise ratio is proportional to the square of the number of sources in the IRS reflector array. By reflecting signals, the intelligent reflector avoids increasing data rates by increasing system energy consumption, thereby improving system energy efficiency. Furthermore, the IRS operates in full-duplex mode, eliminating antenna noise amplification and self-interference, enhancing the effectiveness of the communication system. Because the IRS is lightweight and easy to deploy, it does not require a dedicated power supply room and can be easily installed and deployed in the required environment. Summary of the Invention

[0006] The present invention designs an IRS-assisted cognitive SWIPT system maximum sum rate resource optimization method. This method takes maximizing the cognitive network end-to-end sum rate as the optimization goal, and constructs a multivariable coupled non-convex nonlinear optimization problem under the conditions of cognitive user transmitter power control, cognitive SWIPT user nonlinear energy collection, IRS reflection coefficient vector modulus constraint, etc. The alternating optimization (AO) method is used to solve the optimization problem. First, the IRS phase shift vector is fixed, and the fractional objective function is converted into a subtraction form using the Dinkelbach method, and the Lagrangian dual method is used to obtain the optimal subcarrier and power. Then, the continuous convex approximation (SCA) method is used to obtain the IRS phase shift vector. Simulation shows that the proposed method effectively allocates multiple resources such as subcarriers, power and IRS phase shift vectors, and achieves system sum rate maximization under actual nonlinear energy collection conditions.

[0007] The technical solution of the present invention comprises the following steps:

[0008] Step 1: Scenario assumptions and modeling of the maximum sum rate resource optimization method for the IRS-assisted cognitive SWIPT system:

[0009] In order to avoid loss of generality, before describing the design strategy in detail, the following assumptions are made:

[0010] (1) Ignore the power of the signal reflected from the IRS twice or more, and the maximum reflection on the IRS is lossless;

[0011] (2) The channel gain of the cascaded channel of “cognitive user transmitter-IRS-cognitive user receiver” obeys quasi-static block fading, and the cognitive user transmitter (base station) can obtain the channel state information of all receivers;

[0012] (3) In the primary network, the primary user transmitter does not interfere with the cognitive user receiver;

[0013] In the IRS-assisted cognitive SWIPT system, the primary network and the cognitive network use an underlay spectrum sharing model. While the primary user receiver receives signals from the primary user transmitter, the cognitive user transmitter (cognitive base station) transmits signals to the cognitive user receiver (cognitive SWIPT user) via a direct link and an IRS forwarding link. Cognitive SWIPT employs a power splitting (PS) architecture, using the received signal for information decoding (ID) and nonlinear energy harvesting (EH).

[0014] Assume that the IRS-assisted cognitive SWIPT network consists of a cognitive transmitter and a cognitive receiver, and adopts Orthogonal Frequency Division Multiple Access (OFDMA) with N subcarriers; the system subcarrier set S = S I ∪S E ={1,2,…,N}, Among them S I is the information decoding subcarrier set, S E is a set of nonlinear energy harvesting subcarriers. Each subcarrier is used for ID or nonlinear EH. It is assumed that IRS-assisted cognitive SWIPT uses channel estimation to obtain statistical channel state information (CSI) and uses an IRS with M reflector array sources as an intelligent passive relay.

[0015] Assume that the direct link channel between the cognitive transmitter and the cognitive receiver is represented by h tr ∈C N×1 , the IRS phase shift diagonal matrix is ​​expressed as The IRS reflection phase shift vector is expressed as in represents the reflection coefficient of the mth IRS reflection array source, θ m ∈[0,2π] is the reflection phase shift angle of the mth IRS reflection array source. The channel matrix from the cognitive transmitter to the IRS is expressed as HSI ∈C M×N , the channel matrix from IRS to cognitive receiver is expressed as H IR ∈C M×N , the composite channel gain of the direct link and the cascaded channel from the cognitive transmitter to the cognitive receiver is expressed as where h com ∈C N×1 ;

[0016] The received signal-to-noise ratio (SNR) for information decoding by a cognitive SWIPT user is:

[0017]

[0018] Where P is the transmit power of the cognitive transmitter, S is the subcarrier set, is the IRS reflection phase shift vector; p n is the transmit power of the cognitive transmitter on the nth information decoding subcarrier, h n represents the composite gain of the direct link and the cascaded channel from the cognitive transmitter to the cognitive receiver on the nth subcarrier, is the additive white Gaussian noise (AWGN) channel noise power; the end-to-end sum rate R of the IRS-assisted cognitive SWIPT system is expressed as:

[0019]

[0020] Where B is the system bandwidth;

[0021] If linear energy harvesting is used, the energy harvested by the cognitive SWIPT user can be modeled as:

[0022]

[0023] If nonlinear energy harvesting is used, the energy harvested by the cognitive SWIPT user can be modeled as:

[0024]

[0025] in, Ω is a constant to ensure zero input and zero output response; P sat is the maximum harvested power when the EH circuit is saturated; parameters a and b are constants related to circuit specifications and are positive numbers, such as resistance, capacitance, and diode conduction voltage. In fact, parameter a reflects the nonlinear charging rate relative to the input power, while parameter b is related to the EH circuit turn-on threshold. When the energy harvesting circuit is given, parameter P can be determined by curve fitting. sat , a and b; without loss of generality, we use Ψ n Indicates the energy collected;

[0026] Taking the cognitive SWIPT network end-to-end and rate maximization as the optimization goal, while satisfying multiple constraints such as cognitive user transmitter transmit power control, cognitive SWIPT receiver nonlinear energy collection constraints, and IRS reflection phase shift vector constraints, the constructed mathematical optimization model is expressed as:

[0027]

[0028] in, represents the received signal-to-noise ratio of the cognitive SWIPT user on the information decoding subcarrier, R represents the end-to-end sum rate of the cognitive SWIPT network, Represents cognitive SWIPT user nonlinear energy harvesting; represents the reflection coefficient of the mth IRS reflection array source; p n ,n∈S I represents the transmit power of the cognitive transmitter used for information decoding, p n ,n∈S E represents the transmit power of the cognitive transmitter for energy harvesting, h n represents the composite gain of the direct link and the cascaded channel from the cognitive transmitter to the cognitive receiver on the nth subcarrier, represents the additive white Gaussian noise power, P s represents the cognitive transmitter transmit power threshold, Q min Indicates the minimum harvesting energy threshold for nonlinear energy harvesting, Represents the reflection coefficient modulus of the mth IRS reflection array source.

[0029] Step 2: Optimization of the nonlinear energy harvesting model in the maximum sum rate resource optimization method of the IRS-assisted cognitive SWIPT system:

[0030] After the nonlinear energy harvesting model reaches saturation, it will not increase with the increase of the allocated power in the energy harvesting subcarrier. E is the energy harvesting threshold value that reaches saturation, which introduces the constraint condition of the optimization problem. Then the optimization problem (5) becomes:

[0031]

[0032] Step 3: Information decoding and optimal power allocation of nonlinear energy harvesting in the maximum sum rate resource optimization method of the IRS-assisted cognitive SWIPT system:

[0033] First, the IRS reflection phase shift vector is fixed. Since the power and subcarrier in the objective function and the constraint conditions are dual, the Lagrange dual method is used to solve the optimization problem of Equation (6). Without loss of generality, the system bandwidth and energy conversion efficiency are normalized. The system achievable sum rate under unit frequency band is the spectrum efficiency, which is expressed as The Lagrangian dual function of the optimization problem (6) is:

[0034]

[0035] Among them, α = (α1, α2, α3) is a non-negative Lagrangian dual variable, then the optimization problem (6) can be transformed into the following dual problem:

[0036]

[0037] in, Solving for the dual variable using the subgradient method Reconstruct the Lagrange dual function as:

[0038]

[0039] in, It is related to the transmit power of the cognitive user transmitter in each subcarrier; given a subcarrier set S = S I ∪S E ={1,2,…,N}, in the ID subcarrier set S I With the nonlinear EH subcarrier set S E Internal pair p n Taking the partial derivatives we get:

[0040]

[0041]

[0042] Let the partial derivatives of equations (10) and (11) be zero, and the optimal power of information decoding and energy collection is obtained as follows:

[0043]

[0044]

[0045] Among them, (x) + =max(x,0), the power allocated to the EH subcarrier is related to the channel gain and the dual variable; max With p min Respectively represent the maximum and minimum power constraint values ​​within each subcarrier.

[0046] Step 4: Information decoding and nonlinear energy harvesting for optimal subcarrier allocation in the maximum sum rate resource optimization method of the IRS-assisted cognitive SWIPT system:

[0047] Then, substitute equations (12) and (13) into L(P) to obtain:

[0048]

[0049] in,

[0050] Nonlinear energy harvesting subcarrier set S E Related to U, the subcarrier set that collects the most energy is selected as the optimal nonlinear energy collection subcarrier set, that is:

[0051]

[0052] The remaining subcarriers are used for information decoding, that is:

[0053]

[0054] Furthermore, in step 4, the optimal power and subcarrier allocation for information decoding and nonlinear energy collection using the subgradient method are as follows:

[0055] 4-1 Initialization: Randomly give a set of initial values ​​of α, step size δ, and maximum number of iterations I max and iteration index i = 1, iteration error is ε;

[0056] 4-2 Calculate the subgradient. If the number of iterations is less than the maximum number of iterations, and the updated function value is higher than the original function value, the following loop is executed:

[0057] (a) Update the subgradient function α i+1 =α i +δ i Δα and Lagrangian dual function L(Ρ,S,α);

[0058] (b) Recalculate the subgradient;

[0059] 4-3 The loop ends and the optimal variable α is output * ,S * ,P * .

[0060] Step 5: Optimization of the IRS reflection phase shift vector in the maximum sum rate resource optimization method of the IRS-assisted cognitive SWIPT system:

[0061] Finally, after obtaining the optimal power for information decoding and nonlinear energy collection (Equations (12) and (13)) and the optimal subcarrier set (Equations (15) and (16)), the continuous convex approximation (SCA) method is used to solve the IRS reflection phase shift vector; the optimization problem is constructed as follows:

[0062]

[0063] Composite channel gain Perform N-point discrete Fourier transform (DFT) to obtain the frequency response vector of the composite channel gain as y∈C N×1 ,Right now:

[0064]

[0065] in, is the element of the frequency response vector, f n ∈C N×1 is the discrete Fourier matrix F N ∈C N×N The nth row vector, V∈C M×N is the cascade channel gain;

[0066] Substituting equations (2) and (6) into the optimization problem equation (17), the optimization problem equation (17) is equivalent to:

[0067]

[0068] make definition It is a n and b n A convex differentiable function of and At the point The first-order approximate function of can be used as the lower bound of the original function, that is:

[0069]

[0070] If and only if and When the equation holds true; f n (a n ,b n ) is a n with b n affine function, so it is at the point (a n ,b n ) has the same function At the point The same gradient; Equation (19) can be written as:

[0071]

[0072] This optimization problem is a convex optimization problem, which is solved by using the continuous convex approximation method (SCA) and the MATLAB CVX convex optimization toolbox. At the point The approximate solution of is obtained;

[0073] Furthermore, the specific steps of using alternating optimization to solve the IRS reflection phase shift vector, the optimal power of information decoding and nonlinear energy collection, and subcarrier allocation described in step 5 are as follows:

[0074] 5-1 Fixed IRS reflection phase shift vector The optimal power and subcarrier allocation for information decoding and nonlinear energy collection are obtained through equations (12), (13), (15), and (16);

[0075] 5-2 Fix the optimal power allocation vector P, subcarrier allocation vector S, and initialize the IRS reflection phase shift vector The IRS reflection phase shift vector is updated through SCA and solved using the MATLAB CVX convex optimization toolbox;

[0076] 5-3 until P, S and Make the objective function converge and get the optimal P * ,S * and

[0077] The beneficial effects of the present invention are as follows:

[0078] The present invention discloses a maximum sum rate resource optimization method for an IRS-assisted cognitive SWIPT system. With maximizing the end-to-end sum rate of the IRS-assisted cognitive SWIPT system as the optimization goal, the cognitive SWIPT receiver adopts nonlinear energy harvesting. The multi-resource allocation problem is modeled as a nonlinear non-convex optimization problem, which is mutually coupled with the subcarrier, power, and IRS reflection phase shift vector. Therefore, an alternating optimization method is adopted to fix the IRS phase shift vector, and the Lagrange dual transformation and subgradient method are used to obtain the optimal power and subcarrier set for information decoding and nonlinear energy harvesting. Then, the IRS reflection phase shift vector is obtained by a continuous convex approximation method, thereby achieving the maximization of the end-to-end sum rate of the IRS-assisted cognitive SWIPT system. Studies have shown that compared with other multi-resource optimization strategies, the proposed method effectively allocates multiple resources such as subcarriers, power, and IRS reflection phase shift vectors for information decoding and nonlinear energy harvesting in the cognitive SWIPT system, and achieves the maximum end-to-end sum rate under the actual SWIPT nonlinear energy harvesting model. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 3D model scene graph of the IRS-assisted cognitive SWIPT system.

[0080] Figure 2 The relationship between system sum rate (spectral efficiency) and number of iterations under different IRS strategies.

[0081] Figure 3Figure 2 is the relationship between the system sum rate (spectral efficiency) and the number of IRS reflector array sources under different IRS strategies.

[0082] Figure 4 Figure 3 is the relationship between system sum rate (spectral efficiency) and transmission power under different IRS reflector array source numbers and optimal power allocation within information decoding / nonlinear energy collection subcarriers. DETAILED DESCRIPTION

[0083] The present invention will be further described below with reference to the accompanying drawings and examples.

[0084] Figure 1 A 3D model scenario diagram of the IRS-assisted cognitive SWIPT system is given. In the IRS-assisted cognitive SWIPT system, the main network and the cognitive network adopt an underlay spectrum sharing model. While the main user receiver receives the signal from the main user transmitter, the cognitive user transmitter (cognitive base station) sends a signal to the cognitive user receiver (cognitive SWIPT user) through a direct link and an IRS forwarding link. Cognitive SWIPT adopts a power splitting (PS) structure and uses the received signal for information decoding (ID) and nonlinear energy harvesting (EH). Assume that the IRS-assisted cognitive SWIPT network includes cognitive base stations and cognitive SWIPT users, and adopts orthogonal frequency division multiple access (OFDMA) with a subcarrier number of N. The system subcarrier set S = S I ∪S E ={1,2,…,N}, Among them S I is the ID subcarrier set, S E is a set of nonlinear EH subcarriers. Each subcarrier is used for ID or nonlinear EH. It is assumed that IRS-assisted cognitive SWIPT uses channel estimation to obtain statistical channel state information (CSI) and uses an IRS with M reflector array sources as an intelligent passive relay.

[0085] Figure 2 A graph showing the relationship between the system sum rate (spectral efficiency) and the number of iterations for different IRS strategies is presented. Comparison of the proposed method with the random IRS reflection phase shift vector strategy and the optimal power allocation strategy without SWIPT shows that the sum rate increases monotonically with the number of iterations and converges rapidly to a specific value. Compared with the other two strategies, the proposed method converges quickly and achieves a high achievable sum rate (spectral efficiency). Furthermore, the achievable sum rate converges to 4 bps / Hz for the proposed method, 3.5 bps / Hz for the optimal power allocation strategy without SWIPT, and 2.5 bps / Hz for the random IRS reflection phase shift vector strategy.

[0086] Figure 3A graph showing the relationship between the system sum rate (spectral efficiency) and the number of IRS reflector array sources under different IRS strategies is presented. As can be seen from the graph, the system sum rate performance of all strategies improves with the increase in the number of IRS reflector array sources. Compared with the random IRS reflection phase shift vector strategy, the optimal power allocation strategy without SWIPT, and the optimal power allocation strategy without IRS, the proposed method achieves better system sum rate performance than the other three strategies. As can be seen from the graph, the proposed method allocates optimal power to the subcarriers for information decoding and nonlinear energy harvesting, thus achieving a sum rate gain of 0.5 bps / Hz compared to the optimal power allocation strategy without SWIPT. However, compared with other strategies, the sum rate performance of the proposed method converges to a specific value due to the constraint of the nonlinear energy harvesting threshold.

[0087] Figure 4 A graph shows the relationship between system sum rate (spectral efficiency) and transmit power for different IRS reflector array number and optimal power allocation within the information decoding / nonlinear energy harvesting subcarriers. The graph shows that the sum rate for each strategy increases with increasing transmit power and the number of IRS reflector array sources. With the same number of IRS reflector array sources, the optimal power of the nonlinear energy harvesting subcarriers is superior to the optimal power of the information decoding subcarriers. However, at a transmit power of 36dB, the two values ​​overlap. This is because the optimal power of the nonlinear energy harvesting subcarriers no longer increases with increasing transmit power after reaching their saturation threshold.

[0088] Those skilled in the art should recognize that the above embodiments are merely intended to illustrate the present invention and are not intended to limit the present invention. As long as they are within the scope of the present invention, any changes or modifications to the above embodiments will fall within the scope of protection of the present invention.

Claims

1. IRS-assisted cognitive SWIPT system maximum sum rate resource optimization method, characterized by The following steps are involved: Step 1: Scenario assumption and modeling of the maximum sum rate resource optimization method of the IRS-assisted cognitive SWIPT system; Step 2: Optimization of the nonlinear energy harvesting model in the maximum sum rate resource optimization method of the IRS-assisted cognitive SWIPT system; Step 3: Information decoding and optimal power allocation of nonlinear energy harvesting in the maximum sum rate resource optimization method of the IRS-assisted cognitive SWIPT system; Step 4: Information decoding and nonlinear energy harvesting optimal subcarrier allocation in the maximum sum rate resource optimization method of the IRS-assisted cognitive SWIPT system; Step 5: Optimization of the IRS reflection phase shift vector in the maximum sum rate resource optimization method of the IRS-assisted cognitive SWIPT system.

2. The IRS-assisted cognitive SWIPT system maximum sum rate resource optimization method according to claim 1, characterized in that The scenario assumptions and modeling of the maximum sum rate resource optimization method for the IRS-assisted cognitive SWIPT system described in step 1 are as follows: In order to avoid loss of generality, before describing the design strategy in detail, the following assumptions are made: (1) Ignore the power of the signal reflected from the IRS twice or more, and the maximum reflection on the IRS is lossless; (2) The channel gain of the cascaded channel of "cognitive user transmitter-IRS-cognitive user receiver" obeys quasi-static block fading, and the cognitive user transmitter can obtain the channel state information of all receivers; (3) In the primary network, the primary user transmitter does not interfere with the cognitive user receiver; In the IRS-assisted cognitive SWIPT system, the primary network and the cognitive network adopt an underlay spectrum sharing model. While the primary user receiver receives signals from the primary user transmitter, the cognitive user transmitter sends signals to the cognitive user receiver via a direct link and an IRS forwarding link. Cognitive SWIPT uses a power-splitting structure to use the received signal for information decoding and nonlinear energy harvesting. Assume that the IRS-assisted cognitive SWIPT network includes a cognitive transmitter and a cognitive receiver, and adopts orthogonal frequency division multiple access with N subcarriers; the system subcarrier set S = S I ∪S E ={1,2,…,N}, Among them S I is the ID subcarrier set, S E is a set of nonlinear EH subcarriers; each subcarrier is used for ID or nonlinear EH; it is assumed that IRS-assisted cognitive SWIPT uses channel estimation to obtain statistical channel state information, and uses an IRS with M reflector array sources as an intelligent passive relay; Assume that the direct link channel between the cognitive transmitter and the cognitive receiver is represented by h tr ∈C N×1 , the IRS phase shift diagonal matrix is ​​expressed as The IRS reflection phase shift vector is expressed as in represents the reflection coefficient of the mth IRS reflector array source, θ m ∈[0,2π] is the reflection phase shift angle of the mth IRS reflection array source; the channel matrix from the cognitive transmitter to the IRS is expressed as H SI ∈C M×N , the channel matrix from IRS to cognitive SWIPT receiver is expressed as H IR ∈C M ×N , the composite channel gain of the direct link and the cascaded channel from the cognitive transmitter to the cognitive receiver is expressed as where h com ∈C N×1 ; The received signal-to-noise ratio (SNR) for information decoding by a cognitive SWIPT user is: Where P is the transmit power of the cognitive transmitter, S is the subcarrier set, is the IRS reflection phase shift vector; p n is the transmit power of the cognitive transmitter on the nth information decoding subcarrier, h n represents the composite gain of the direct link and the cascaded channel from the cognitive transmitter to the cognitive receiver on the nth subcarrier, is the channel noise power of additive white Gaussian noise; the end-to-end sum rate R of the IRS-assisted cognitive SWIPT system is expressed as: Where B is the system bandwidth; If linear energy harvesting is used, the energy harvested by the cognitive SWIPT user can be modeled as: If nonlinear energy harvesting is used, the energy harvested by the cognitive SWIPT user can be modeled as: in, Ω is a constant to ensure zero input and zero output response; P sat is the maximum harvested power when the EH circuit is saturated; parameters a and b are constants related to circuit specifications and are positive numbers. When the energy harvesting circuit is given, the parameter P is determined by curve fitting. sat , a and b; using Ψ n Indicates the energy collected; Taking the cognitive SWIPT network end-to-end and rate maximization as the optimization goal, while satisfying multiple constraints such as cognitive user transmitter transmit power control, cognitive SWIPT receiver nonlinear energy collection constraints, and IRS reflection phase shift vector constraints, the constructed mathematical optimization model is expressed as: in, represents the received signal-to-noise ratio of the cognitive SWIPT user on the information decoding subcarrier, R represents the end-to-end sum rate of the cognitive SWIPT network, Represents cognitive SWIPT user nonlinear energy harvesting; represents the reflection coefficient of the mth IRS reflection array source; p n ,n∈S I represents the transmit power of the cognitive transmitter used for information decoding, p n ,n∈S E represents the transmit power of the cognitive transmitter for energy harvesting, h n represents the composite gain of the direct link and the cascaded channel from the cognitive transmitter to the cognitive receiver on the nth subcarrier, represents the additive white Gaussian noise power, P s represents the cognitive transmitter transmit power threshold, Q min Indicates the minimum harvesting energy threshold for nonlinear energy harvesting, Represents the reflection coefficient modulus of the mth IRS reflection array source.

3. The IRS-assisted cognitive SWIPT system maximum sum rate resource optimization method according to claim 2, characterized in that The nonlinear energy harvesting model optimization in the maximum sum rate resource optimization method of the IRS-assisted cognitive SWIPT system described in step 2 is as follows: After the nonlinear energy harvesting model reaches saturation, it will not increase with the increase of the allocated power in the energy harvesting subcarrier. E is the energy harvesting threshold value that reaches saturation, which introduces the constraint condition of the optimization problem. Then the optimization problem (5) becomes:

4. The IRS-assisted cognitive SWIPT system maximum sum rate resource optimization method according to claim 3, characterized in that The optimal power allocation for information decoding and nonlinear energy harvesting in the maximum sum rate resource optimization method of the IRS-assisted cognitive SWIPT system described in step 3 is as follows: First, the IRS reflection phase shift vector is fixed. Since the power and subcarrier in the objective function and the constraint conditions are dual, the Lagrange dual method is used to solve the optimization problem of Equation (6). The system bandwidth and energy conversion efficiency are normalized. The system achievable sum rate under unit frequency band is the spectrum efficiency, which is expressed as The Lagrangian dual function of the optimization problem (6) is: Where α = (α1, α2, α3) is a non-negative Lagrangian dual variable, and the optimization problem (6) is transformed into the following dual problem: in, Solving for the dual variable using the subgradient method Reconstruct the Lagrange dual function as: in, It is related to the transmit power of the cognitive user transmitter in each subcarrier; given a subcarrier set S = S I ∪S E ={1,2,…,N}, in the ID subcarrier set S I With the nonlinear EH subcarrier set S E Internal pair p n Taking the partial derivatives we get: Let the partial derivatives of equations (10) and (11) be zero, and the optimal power of information decoding and energy collection is obtained as: Among them, (x) + =max(x,0), the power allocated to the EH subcarrier is related to the channel gain and the dual variable; in addition, p max With p min Respectively represent the maximum and minimum power constraint values ​​within each subcarrier.

5. The IRS-assisted cognitive SWIPT system maximum sum rate resource optimization method as described in claim 4 is characterized in that The optimal subcarrier allocation for information decoding and nonlinear energy collection in the maximum sum rate resource optimization method of the IRS-assisted cognitive SWIPT system described in step 4 is as follows: Then, substitute equations (12) and (13) into L(P) to obtain: in, Nonlinear energy harvesting subcarrier set S E Related to U, the subcarrier set that collects the most energy is selected as the optimal nonlinear energy collection subcarrier set, that is: The remaining subcarriers are used for information decoding, that is:

6. The IRS-assisted cognitive SWIPT system maximum sum rate resource optimization method according to claim 5, characterized in that The optimization of the IRS reflection phase shift vector in the maximum sum rate resource optimization method of the IRS-assisted cognitive SWIPT system described in step 5 is as follows: Finally, after obtaining the optimal power and optimal subcarrier set for information decoding and nonlinear energy collection, the continuous convex approximation method is used to solve the IRS reflection phase shift vector; the optimization problem is constructed as follows: Composite channel gain Perform discrete Fourier transform at N points and obtain the frequency response vector of the composite channel gain as y∈C N×1 ,Right now: in, is the element of the frequency response vector, f n ∈C N×1 is the discrete Fourier matrix F N ∈C N×N The nth row vector, V∈C M×N is the cascade channel gain; Substituting equations (2) and (6) into the optimization problem equation (17), the optimization problem equation (17) is equivalent to: make definition It is a n and b n A convex differentiable function of and At the point The first-order approximate function of can be used as the lower bound of the original function, that is: If and only if and When the equation holds true; f n (a n ,b n ) is a n with b n affine function, so it is at the point (a n ,b n ) has the same function At the point The same gradient; Equation (19) can be written as: This optimization problem is a convex optimization problem, which is solved by using the continuous convex approximation method and the MATLAB CVX convex optimization toolbox. At the point An approximate solution is obtained.

7. The IRS-assisted cognitive SWIPT system maximum sum rate resource optimization method according to claim 6, characterized in that Step 4 uses the sub-gradient method to perform optimal power and sub-carrier allocation for information decoding and nonlinear energy collection. The specific steps are as follows: 4-1 Initialization: Randomly give a set of initial values ​​of α, step size δ, and maximum number of iterations I max and iteration index i = 1, iteration error is ε; 4-2 Calculate the subgradient. If the number of iterations is less than the maximum number of iterations, and the updated function value is higher than the original function value, the following loop is executed: (a) Update the subgradient function α i+1 =α i +δ i Δα and Lagrangian dual function L(Ρ,S,α); (b) Recalculate the subgradient; 4-3 The loop ends and the optimal variable α is output * 、S * 、P * .

8. The IRS-assisted cognitive SWIPT system maximum sum rate resource optimization method according to claim 7, characterized in that Step 5 uses alternating optimization to solve the IRS reflection phase shift vector, the optimal power for information decoding and nonlinear energy collection, and the subcarrier allocation. The specific steps are as follows: 5-1 Fixed IRS reflection phase shift vector The optimal power and subcarrier allocation for information decoding and nonlinear energy collection are obtained through equations (12), (13), (15), and (16); 5-2 Fix the optimal power allocation vector P, subcarrier allocation vector S, and initialize the IRS reflection phase shift vector The IRS reflection phase shift vector is updated through SCA and solved using the MATLAB CVX convex optimization toolbox; 5-3 until P, S and Make the objective function converge and get the optimal P * ,S * and

Images