Wireless sensor network resource optimization allocation method based on millimeter wave wireless energy transmission

Abstract

This invention discloses a method for optimizing resource allocation in wireless sensor networks based on millimeter-wave wireless power transmission. The method includes the following steps: constructing a single-cell wireless sensor network model based on millimeter-wave energy harvesting; fixing power allocation coefficients and, with the objective of maximizing the total energy received by all wireless sensors, using coalition formation game theory to obtain the optimal wireless sensor grouping and antenna allocation; employing linear programming optimization to obtain the optimal power allocation for each wireless sensor within the energy harvesting time, maximizing the total harvested energy of all sensors; and utilizing the Lagrange dual multiplier method and KKT conditions to maximize system energy efficiency, obtaining the optimal allocation between uplink and downlink transmissions in the single-cell wireless sensor network model. This invention not only enables flexible beamforming to enhance the total received energy of wireless sensors in multi-sensor application scenarios but also improves the overall energy efficiency of the system.

Description

Technical Field

[0001] This invention relates to the field of fifth-generation mobile communication technology (5G), and more particularly to a method for optimizing the allocation of wireless sensor network resources based on millimeter-wave wireless power transmission. Background Technology

[0002] Radio frequency (RF) signal-based energy-carrying communication technology has attracted widespread research from industry and academia due to its advantages such as being unaffected by weather, location, and hardware size. Energy-carrying communication technology is not only suitable for cellular networks and IoT systems, but also for autonomous driving systems, mobile robotics systems, and distributed sensor networks. However, the currently used sub-6GHz communication frequency bands cannot fully support these services. Millimeter waves (30–300GHz) combined with massive MIMO can significantly improve data transmission rates and provide greater power and energy coverage than sub-6GHz bands, and are considered one of the key technologies for 5G / 6G mobile communication networks. Therefore, applying millimeter wave energy-carrying communication technology to wireless sensor networks can effectively improve system energy efficiency and provide an effective solution to address the bandwidth and energy demands brought about by device growth.

[0003] The paper (G. Kwon, H. Park, and MZ Win, Joint beamforming and power splitting for wideband millimeter wave SWIPT systems, IEEE Journal of Selected Topics in Signal Processing, vol. 15, no. 5, pp. 1211-1227, Aug. 2021.) proposes a joint optimization scheme for power allocation, beamforming, and power splitting in wideband millimeter wave SWIPT downlink systems, addressing the limited channel state information and beamforming structure in millimeter wave MIMO wireless information-energy-sharing (SWIPT) systems. The main strategy is to maximize the achievable rate-weighted sum while considering user fairness. The literature (L. Chen, B. Hu, G. Xu, and S. Chen, Energy-efficient power allocation and splitting for mmWave beamspace MIMO-NOMA with SWIPT, IEEE Sensors Journal, vol. 21, no. 114, pp. 16381-16394, Jul. 2021.) studies the application of SWIPT technology to improve energy efficiency in millimeter-wave non-orthogonal multiple access (NOMA) MIMO systems. To achieve the goal of simultaneously ensuring the minimum data rate and energy requirements of each node, the authors maximize energy efficiency through effective allocation of base station power and power splitting of each node. Due to inter-beam interference, the joint optimization problem exhibits a non-convex form; therefore, a two-layer iterative algorithm is proposed using the Dinkelbach method and alternating optimization methods to solve the non-convex optimization problem. The paper (ANUwaechia and NMMahyuddin, Spectrum and energy efficiency optimization for hybrid precoding-based SWIPT-enabled mmWave mMIMO-NOMA systems, IEEE Access, vol. 8, pp. 139994-140007, 2020.) studies hybrid precoding for millimeter-wave NOMA-MIMO systems based on wireless energy transmission, aiming to maximize the system's total achievable rate and total energy efficiency. The authors first propose an optimization problem for user packets, then design a hybrid analog-digital precoder, and finally decompose the non-convex optimization problem into four independent sub-optimization problems, using an alternating optimization algorithm to obtain the optimal power allocation and power division parameters for the receiving nodes.

[0004] Currently, most millimeter-wave-based multi-user communication systems assume that a single analog beam can transmit data to multiple users simultaneously. If the users are located relatively close together, several single analog beams can be used to complete the information transmission for each user. However, in practical applications, the users to be served are usually located in dispersed locations, and the analog beams formed by the antenna array are also limited. Therefore, it is necessary to consider optimizing user grouping and using beam splitting technology to achieve information / energy transmission to multiple users with limited transmission resources.

[0005] Current efforts to improve system energy efficiency mainly focus on user grouping and power allocation, with little attention paid to optimizing antenna allocation and uplink / downlink transmission time. In millimeter-wave power-carrying communication systems based on Time Division Multiplexing (TDD), the optimized allocation of uplink / downlink transmission time has a significant impact on system energy efficiency, thus necessitating optimization analysis. Summary of the Invention

[0006] In a single-cell wireless sensor network based on millimeter-wave wireless power transmission, the base station first transmits energy signals to multiple sensors simultaneously via a downlink transmission link. Subsequently, the sensors use the collected energy to transmit their respective data back to the base station via an uplink using time-division multiplexing (TDD). By analyzing the realizable rate expression of the base station's received information, the shape of the base station's transmitted energy beam depends on the sensor grouping and antenna allocation. The energy collected by the sensors and the information transmission also depend on the base station's power allocation to the sensors and the time allocation for uplink and downlink transmissions. To improve the overall energy efficiency of the sensors, this invention proposes a method for joint optimization of sensor grouping, antenna allocation, power allocation, and time allocation. Specifically, the joint optimization method consists of two stages. In the first stage, with fixed power allocation coefficients, the optimal sensor grouping and antenna allocation are obtained using coalition formation game theory, aiming to maximize the total energy received by all sensors. In the second stage, based on the sensor grouping and antenna allocation information obtained in the first stage, a linear programming optimization method is used to obtain the optimal power allocation for each sensor within the energy acquisition time, thereby maximizing the total acquisition energy of all sensors. Then, using the Lagrange dual multiplier method and the Karush-Kuhn-Tucker (KKT) conditions, the optimal allocation between uplink and downlink transmissions is obtained with the goal of maximizing system energy efficiency. Simulation results show that, compared with other fixed-parameter allocation schemes, the proposed optimized allocation scheme can effectively improve the system's energy efficiency.

[0007] The objective of this invention is achieved by at least one of the following technical solutions.

[0008] A method for optimizing resource allocation in wireless sensor networks based on millimeter-wave wireless power transmission includes the following steps:

[0009] S1. Construct a single-cell wireless sensor network model based on millimeter-wave energy harvesting, including a millimeter-wave base station BS and K wireless sensors;

[0010] S2. With a fixed power allocation coefficient, the optimal wireless sensor grouping and antenna allocation are obtained by using a coalition-forming game theory approach, with the goal of maximizing the total energy received by all wireless sensors.

[0011] S3. Based on the obtained wireless sensor grouping and antenna allocation information, the optimal power allocation of each wireless sensor during the energy acquisition time is obtained by using the linear programming optimization method, so as to maximize the total acquisition energy of all sensors. Then, using the Lagrange dual multiplier method and KKT conditions, with the goal of maximizing the system energy efficiency, the optimal allocation between uplink and downlink transmissions of the single-cell wireless sensor network model is obtained.

[0012] Furthermore, in step S1, K wireless sensors are deployed in a Poisson distribution with density λ within a circle of radius ρ, and the millimeter-wave base station BS is placed at the center of the circle.

[0013] Millimeter-wave base stations (BS) have a continuous power supply (P) BS Since wireless sensors lack a continuous power supply, wireless sensor networks employ a 'collect energy first, then transmit information' transmission protocol. Specifically, within downlink transmission time slot τ0, the millimeter-wave base station BS generates multiple energy beams to provide power to the wireless sensors. Subsequently, the wireless sensors use the acquired energy to transmit information back to the millimeter-wave base station BS in uplink transmission time slots 1-τ0 using time division multiplexing (TDD). The k-th wireless sensor S... k The allocated transmission time slot is τ k ;

[0014] The millimeter-wave base station BS is equipped with a linear antenna array consisting of M antennas, which can form N radio frequency links; the number of antennas allocated to each radio frequency link is set to M. RF Equal, i.e., M RF =M / N, if M RF If the value is not an integer, the remaining antennas are assigned to the last RF link.

[0015] All wireless sensors are equipped with only a single antenna and are connected to only one radio frequency link.

[0016] Furthermore, the millimeter-wave base station BS and the k-th wireless sensor S k millimeter-wave channel h k It can be modeled as:

[0017]

[0018] in, Indicates the orientation towards the k-th wireless sensor S k The array guide vector, φ k This represents the millimeter-wave base station BS and the k-th wireless sensor S. k The line-of-sight link departure angle between them; This represents the wide-range path fading of millimeter-wave channels, d k This represents the millimeter-wave base station BS and the k-th wireless sensor S. k The straight-line distance between them, where α represents the path loss coefficient; g k This represents the millimeter-wave base station BS and the k-th wireless sensor S. k millimeter-wave channel h k Small-scale path decay.

[0019] Furthermore, in practical wireless sensor networks, the number of radio frequency links is often less than the number of wireless sensors. Therefore, it is necessary to group the K sensors to improve the efficiency of energy harvesting. If there are too many wireless sensors on each radio frequency link, the energy received by each sensor will be greatly reduced.

[0020] Therefore, in order to ensure the energy harvesting efficiency of each sensor, each radio frequency link can serve a maximum of two sensors;

[0021] According to the transmission protocol, the entire transmission process can be divided into two stages:

[0022] In the first phase, the base station BS generates multiple energy beams within the downlink transmission time slot τ0 to provide power to K wireless sensors. Therefore, the wireless sensors S associated with the RF link r k The received energy signal can be represented as:

[0023]

[0024] Among them, matrix Represents matrix h k The Hermitian matrix, w r and w t Let p represent the analog beamforming vectors for the r-th RF link and the t-th RF link, respectively; k,r Represented as sensor S k The transmit power allocated to the r-th radio frequency link; if the wireless sensor S k If the r-th radio frequency link is associated, then the wireless sensor S k Correlation coefficient u k,r =1, otherwise the correlation coefficient u k,r =0; e k Indicates a wireless sensor S k unit energy signal; nk Indicates wireless sensor S k The received noise; therefore, the wireless sensor S k The received radio frequency power is:

[0025]

[0026] Wireless sensor S k The collected energy can be expressed as:

[0027]

[0028] Where η represents the energy conversion efficiency, η∈(0,1);

[0029] In the second stage, within the remaining transmission time slots 1-τ0, K wireless sensors use the collected energy to transmit their respective data back to the millimeter-wave base station BS via time-division multiplexing on the uplink; at this time, wireless sensor S k The transmission power is expressed as Where τ k Indicates wireless sensor S k The transmission time slot; at this time, the millimeter-wave base station BS receives the wireless sensor S k The achievable transmission rate r of the transmitted signal k for:

[0030]

[0031] Among them, h k Indicates wireless sensor S k Millimeter-wave channel between the base station (BS) and the millimeter-wave channel.

[0032] Furthermore, due to the narrow beamwidth of millimeter waves, they are generally unable to provide power transfer services for large-area wireless sensors. Therefore, if a group of wireless sensors is located in a single millimeter wave coverage area, power transfer can be achieved using a single beam; otherwise, beam splitting technology is used to divide all antenna elements into several subarrays, generating multiple simulated beams to provide power transfer for multiple sensors within a group, with each beam directed to one wireless sensor.

[0033] The k-th wireless sensor S k and the i-th wireless sensor S i Form a group and associate it with the r-th radio frequency link, the k-th wireless sensor S k and the i-th wireless sensor S i The angles between the base station and the millimeter-wave base station are φ k and φ i Assume φ k and φ iThe angular difference between them is less than or equal to one millimeter wavewidth φ B That is, |φ k -φ i |≤φ B The k-th wireless sensor S can be implemented using a single analog beam. k and the i-th wireless sensor S i Simultaneously, it provides power transmission. At this point, the r-th RF link uses a single analog beam. To transfer energy, It can be represented as:

[0034]

[0035] in, This indicates the aperture direction of the millimeter-wave base station BS on the r-th radio frequency link;

[0036] Furthermore, if the k′-th wireless sensor S is associated with the t-th radio frequency link k′ and the i′-th wireless sensor S i′ The absolute angular difference is greater than one millimeter wavewidth φ B That is, |φ k′ -φ i′ |>φ B Then, beam splitting technology is used to separately use and The antenna elements form two sub-beams, which are respectively directed to the k′-th wireless sensor S. k′ and the i′-th wireless sensor S i′ Provide energy services, of which and satisfy At this moment, the k′-th wireless sensor S is pointed to. k′ and the i′-th wireless sensor S i′ beam vector and They are represented as follows:

[0037]

[0038]

[0039] At this time, M is associated with the t-th radio frequency link. RF The analog beamforming formed by the antenna can be expressed as:

[0040]

[0041] In the constructed single-cell wireless sensor network model based on millimeter-wave energy harvesting, the millimeter-wave base station BS first sends energy signals to K wireless sensors through the downlink, and then the wireless sensors use the harvested energy to send their respective signals to the millimeter-wave base station BS through the uplink in the TDD manner.

[0042] Therefore, the total transmission rate of wireless sensors will become an important indicator for judging system performance. By analyzing the achievable rate expression (5) of the received signal of the millimeter-wave base station (BS), it can be seen that the simulated energy beamforming depends on the grouping of K wireless sensors and the antenna allocation of each link. At the same time, effective energy harvesting and information transmission also depend on the power allocation from the millimeter-wave base station (BS) to the wireless sensors and the time slot allocation for uplink and downlink transmission. Therefore, if we want to improve the energy efficiency of the single-cell wireless sensor network model based on millimeter-wave energy harvesting, we need to jointly optimize the grouping of wireless sensors, base station antenna allocation, power allocation and transmission time slot allocation. In order to effectively obtain these parameters, this invention proposes a two-stage optimization design method, as follows:

[0043] In the first stage, by fixing the power allocation coefficient and aiming to maximize the total energy received by K wireless sensors, the optimal wireless sensor grouping and antenna allocation information associated with each radio frequency link are obtained by adopting the alliance formation game theory.

[0044] In the second stage, based on the wireless sensor grouping and antenna allocation information obtained in the first stage, the optimal power allocation coefficient is obtained by using the linear programming problem in convex optimization with the goal of maximizing the total energy acquired by the K wireless sensors. Then, with the goal of maximizing the system energy efficiency, the optimal time allocation for uplink and downlink transmission is obtained by using the Lagrange dual multiplier method.

[0045] Furthermore, in step S2, the optimization problem is defined as follows:

[0046] Assuming equal power allocation is applied to the K wireless sensors, and beam tracking technology is used, the millimeter-wave base station BS can ensure that it acquires the channel state and location information of all wireless sensors, including the k-th wireless sensor S. k Distance d from millimeter-wave base station BS k and angle φ k Based on grouping and antenna allocation, the optimization problem of maximizing the energy collected by all wireless sensors can be expressed as (P1):

[0047]

[0048] p k,r =P BS Substituting / K into formula (3), the objective function in the first optimization problem (P1) is... Represents the sum of conditionally received radio frequency power of all wireless sensors; constraint C1 indicates that if the k-th wireless sensor S k If the r-th radio frequency link is associated, then the association coefficient u k,r =1, otherwise the correlation coefficient u k,r =0; Constraint C2 indicates that each RF link can provide power transfer services to at most two wireless sensors; Constraint C3 indicates that each wireless sensor can be associated with at most one RF link; Constraint C4 ensures that the number of antennas allocated on the r-th RF link does not exceed M. RF Constraint C5 guarantees that, under beam splitting conditions, the antenna associated with the r-th RF link is allocated to the wireless sensor S. k The minimum number of antennas is M min In the first optimization problem (P1), the objective function is... The effective channel gain follows a triangular periodic function of the number of antennas. Therefore, the optimization problem under consideration is a non-convex integer programming problem. It is impossible to directly obtain the optimized wireless sensor grouping and antenna allocation information using convex optimization. Therefore, the problem to be solved is transformed into a coalition formation game problem to obtain the optimized grouping and antenna allocation.

[0049] The specific game problem of alliance formation is as follows:

[0050] First, consider all K wireless sensors that need to harvest energy as a set. Several alliances are formed through joint participation in the game to maximize the total radio frequency power collected by all wireless sensors. The payoff of alliance G in the game is represented by V(G), which consists of a set of payoff vectors. Each element in the vector represents the payoff of the corresponding wireless sensor in alliance G, i.e., the collected radio frequency power, which is expressed by formula (3). The payoff V(G) of alliance G in the game is not only related to its own participation in the alliance, but also to the cooperation structure of other participants. At the same time, each sensor in alliance G will also be affected by antenna allocation in the case of beam splitting. Therefore, the game has the characteristic of non-transferable utility. In summary, the game under consideration can be regarded as an alliance formation game problem.

[0051] Alliance G associated with the r-th radio frequency link r In this context, K wireless sensors form several consortiums, denoted as B = {G1, ..., G...} N},r∈{1,…,N},V(G r The value of alliance group B is related to that of alliance G. r Internal antenna allocation strategy M r Antenna allocation strategies of other alliances Ξ\{M r} related, where Ξ={M1,…,M l};

[0052] Therefore, the alliance value V(G) associated with the r-th RF link r B) can be represented as:

[0053] V(G r M r ,B,Ξ\{M r})=v k (G r M r ,B,Ξ\{M r}); (11)

[0054] in, Indicates wireless sensor S k In Alliance G r The proposed optimization algorithm, based on the characteristics of alliance-based games, aims to continuously iterate among wireless sensors to form optimized alliances based on the preference relationships among the alliances. That is, each wireless sensor compares its contribution in different alliances and decides whether to join other alliances or stay in the current alliance. Through several iterations, the total received radio frequency power of all wireless sensors is improved.

[0055] Furthermore, the alliance is set up to form a preference relationship, as follows:

[0056] Alliance Formation Preference Relationship 1: In the iterth iteration, assume the current set of wireless sensors... The alliance groups and the antenna allocation strategies corresponding to each alliance within the alliance groups are respectively B iter ={G1,…,G l} and Ξ iter ={M1,…,M l The wireless sensor S is valid if and only if the following condition is met. k Will leave the current alliance G r And join another alliance G associated with the t-th RF link. t :

[0057]

[0058] Among them, the newly formed alliance groups in the (iter+1)th iteration Represented as Since a consortium can accommodate a maximum of two wireless sensors, when wireless sensor S k Leaving the original alliance G r After that, the new alliance {G r The antenna allocation in \{k}} is represented as follows:

[0059]

[0060] When wireless sensor S k Join the alliance G t At that time, a one-dimensional search is performed on the antenna allocation within the newly formed coalition group, while the antenna allocation of the remaining coalitions is fixed. Find newly formed alliances {G t The maximum received RF power of ∪{k}}, at which point the antenna allocation within the newly formed alliance group is the optimal antenna allocation for that alliance. At this time, the newly formed alliance group B iter+1 The antenna allocation strategy in can be expressed as: In this preference relationship, only the coalition G t When the number of wireless sensors in the system is less than 2, the wireless sensor S k That's why they tried to join the alliance G. t ;

[0061] Alliance Formation Preference Relationship 2: In the iterth iteration, assume the current wireless sensor set... The alliance groups and the antenna allocation strategies corresponding to each alliance within the alliance groups are respectively B iter ={G1,…,G l} and Ξ iter ={M1,…,M l The wireless sensor S is valid if and only if the following condition is met. k ∈G r and wireless sensor S k′ ∈G t They will exchange positions:

[0062]

[0063] Among them, the newly formed alliance group B in the (iter+1)th iteration iter+1 It can be represented as B iter+1 ={B iter \{G r G t}}∪{G r \{k}∪{k′},G t \{k′}∪{k}}; By performing a one-dimensional search on the antenna allocation within the two newly formed alliances, while fixing the antenna allocation of the remaining alliances {Ξ iter+1 \{M r\{k}∪k′ M t\{k′}∪k}}, find a newly formed alliance G r \{k}∪{k′} and G t The maximum received radio frequency power of \{k′}∪{k}, at which point a new alliance group G is formed. r \{k}∪{k′} and G tThe antenna allocation corresponding to \{k′}∪{k} is the optimal antenna allocation for the alliance. and and It can be done by [M] min M RF -M min The integers within the range are obtained by performing a two-dimensional full search, where the computational complexity of obtaining the maximum value is acceptable.

[0064] Furthermore, the wireless sensor grouping and antenna allocation algorithm specifically includes the following steps:

[0065] S2.1. Initialize the algorithm based on the number of RF links N and the number of wireless sensors, as follows:

[0066] When the number of wireless sensors K=N, each radio frequency link randomly corresponds to one wireless sensor, that is, each wireless sensor forms a separate alliance and exclusively uses all the antennas of that radio frequency link;

[0067] When the number of wireless sensors is 2N>K>N, the wireless sensors are randomly divided into alliances. Some alliances have two wireless sensors corresponding to one radio frequency link and the antennas of the radio frequency link are evenly distributed. Some alliances have only one wireless sensor corresponding to one radio frequency link and exclusively enjoy all the antennas of the link.

[0068] When the number of wireless sensors K≥2N, 2N sensors are randomly selected to form N alliances. Each alliance contains two wireless sensors corresponding to one radio frequency link. The wireless sensors are evenly distributed among the antennas of the radio frequency link.

[0069] At this point, the initialization settings can be expressed as: define the initialization iteration index iter = 0, and initialize the alliance group as B0 = {G1, ..., G}. N The antenna allocation strategy for each alliance is Ξ0 = {M1, ..., M}. N};

[0070] S2.2 In the iterth iteration stage, any given member of the union G... r The wireless sensors will access other alliance Gs one by one. t If wireless sensor S k Join the alliance G t When the number of wireless sensors in the alliance is less than or equal to 2, the alliance forming preference relation 1 is used to determine the wireless sensor S. k Should they leave their original alliance G? r Join the alliance G t If wireless sensor S k Join the alliance G tIf the number of wireless sensors in the alliance is greater than 2, then alliance preference relation 2 will be adopted to belong to alliance G. r wireless sensor S k With Alliance G t Wireless sensor S in k′ They exchange positions, and the wireless sensor S is determined based on the calculated radio frequency power collected by all wireless sensors. k With wireless sensor S k′ Should we perform a position swap to obtain a better alliance group?

[0071] S2.3 After multiple iterations, if neither Alliance Formation Preference Relation 1 nor Alliance Formation Preference Relation 2 can form a new alliance group, then the corresponding alliance group B... iter The optimal wireless sensor grouping, and the antenna allocation strategy Ξ for each alliance within the alliance group. iter ={M1,…,M N} represents the optimal antenna allocation strategy.

[0072] Furthermore, in step S3, based on the optimal wireless sensor grouping and antenna allocation information obtained in step S2, it is necessary to further optimize power allocation and time allocation to ensure that the total collection energy of the wireless sensors and the energy efficiency of the base station's received signal are maximized, as follows:

[0073] Optimized power allocation: After obtaining the optimal wireless sensor grouping and antenna allocation, the optimal power allocation of the millimeter-wave base station (BS) for each wireless sensor during the energy harvesting phase can further increase the total harvesting energy of all wireless sensors. First, assuming that the downlink transmission time slot τ0 and the uplink transmission time slot (1-τ0) are constants, the problem of maximizing the total harvesting energy of the wireless sensors can be expressed as a second optimization problem (P2):

[0074]

[0075] Substituting the optimized wireless sensor grouping and antenna allocation information into equation (3) yields the following result: The constraint C6 in equation (15) guarantees that the wireless sensor S k The minimum threshold value for the received radio frequency power is P RF Constraint C7 states that the total energy allocated by the millimeter-wave base station (BS) to all wireless sensors is no greater than the transmission power P of the millimeter-wave base station (BS). BS The constraint C8 indicates that the energy allocated by the millimeter-wave base station (BS) to each wireless sensor must be greater than or equal to 0; by observing the second optimization problem (P2), it can be seen that the objective function... Since both constraints C6 and C7 are linear functions, the second optimization problem (P2) is a linear optimization problem, which can be solved using the corresponding optimization tools in Matlab.

[0076] Optimizing time slot allocation: In the uplink transmission time slot, each wireless sensor uses the collected energy to transmit its own signal to the millimeter-wave base station (BS) in TDD mode. Therefore, by optimizing the energy collection time slot and the information transmission time slot of each wireless sensor, the energy efficiency of the millimeter-wave base station (BS) in receiving signals can be maximized, which can be defined as:

[0077]

[0078] Among them, R sum and E T These represent the total achievable rate of signal reception at the base station (BS) and the total transmit energy of the wireless sensor, respectively, as follows:

[0079]

[0080] This indicates that under optimal conditions of wireless sensor grouping, antenna allocation, and power allocation, the wireless sensor S... k The radio frequency energy collected during downlink transmission; therefore, the third optimization problem (P3) to maximize the energy efficiency of the millimeter-wave base station (BS) received signal can be expressed as:

[0081]

[0082] By analyzing the third optimization problem (P3), its objective function and constraint C9 are concave and linear functions, respectively. Therefore, the objective function in the third optimization problem (P3) can be transformed into its corresponding parametric form, and the optimal time slot allocation can be obtained by using the Lagrange dual multiplier method. The parameterized form of the objective function in the third optimization problem (P3) can be expressed as:

[0083]

[0084] Where θ represents a non-negative parameter; for a given value of θ, the Lagrangian function corresponding to the third optimization problem (P3) can be expressed as:

[0085]

[0086] Here, ξ represents the Lagrange multiplier.

[0087] Compared with the prior art, the advantages of the present invention are:

[0088] Existing millimeter-wave-based power-carrying communication transmission protocols mostly focus on downlink transmission, employing simultaneous information and energy transmission. Information and energy are sent from the base station to multiple users according to time division or power allocation, and the impact of antenna allocation on system performance improvement is often ignored during system performance optimization. This invention considers a power-carrying communication application scenario that takes into account both uplink and downlink transmission, i.e., downlink energy transmission and uplink information transmission. Based on this transmission protocol, to improve the overall energy efficiency of the system, this invention proposes a method with low algorithm complexity and easy implementation for joint optimization of wireless sensor grouping, antenna allocation, power allocation, and time allocation. This method not only enables flexible beam routing to increase the total received energy of wireless sensors in multi-wireless sensor applications but also improves the overall energy efficiency of the system. Attached Figure Description

[0089] Figure 1 This is a schematic diagram of the structure of a single-cell wireless sensor network model based on millimeter-wave energy harvesting in an embodiment of the present invention;

[0090] Figure 2 This is a schematic diagram of the transmission protocol in an embodiment of the present invention;

[0091] Figure 3 This is a schematic diagram of variable energy beamforming in an embodiment of the present invention;

[0092] Figure 4 This is a schematic diagram illustrating the total energy received by the wireless sensor in a millimeter-wave wireless sensor network when the present invention is used, a single-beam scheme based on time-division multiplexing is used, and a multi-beam scheme based on fixed power allocation and time slot allocation are used respectively.

[0093] Figure 5 This is a schematic diagram illustrating the total rate at which a base station receives information transmitted by all wireless sensors in a millimeter-wave wireless sensor network when using the present invention, a single-beam scheme based on time-division multiplexing, and a multi-beam scheme based on fixed power allocation and time slot allocation, respectively.

[0094] Figure 6 This diagram illustrates the ratio of the total base station reception rate to the total base station reception rate in a millimeter-wave wireless sensor network, using the present invention, a single-beam scheme based on time-division multiplexing, and a multi-beam scheme based on fixed power allocation and time slot allocation, respectively. Detailed Implementation

[0095] To make the objectives, technical solutions, and advantages of the present invention clearer, the specific implementation of the present invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0096] Example 1:

[0097] A method for optimizing resource allocation in wireless sensor networks based on millimeter-wave wireless power transmission includes the following steps:

[0098] S1, such as Figure 1 As shown, a single-cell wireless sensor network model based on millimeter-wave energy harvesting is constructed, including a millimeter-wave base station BS and K wireless sensors;

[0099] K wireless sensors are deployed in a Poisson distribution with density λ within a circle of radius ρ, and a millimeter-wave base station BS is placed at the center of the circle.

[0100] Millimeter-wave base stations (BS) have a continuous power supply (P) BS Since wireless sensors lack a continuous power supply, wireless sensor networks employ a 'collect energy first, then transmit information' transmission protocol. Specifically, within downlink transmission time slot τ0, the millimeter-wave base station BS generates multiple energy beams to provide power to the wireless sensors. Subsequently, the wireless sensors use the acquired energy to transmit information back to the millimeter-wave base station BS in uplink transmission time slots 1-τ0 using time division multiplexing (TDD). The k-th wireless sensor S... k The allocated transmission time slot is τ k ;

[0101] The millimeter-wave base station BS is equipped with a linear antenna array consisting of M antennas, which can form N radio frequency links; the number of antennas allocated to each radio frequency link is set to M. RF Equal, i.e., M RF =M / N, if M RF If the value is not an integer, the remaining antennas are assigned to the last RF link.

[0102] All wireless sensors are equipped with only a single antenna and are connected to only one radio frequency link.

[0103] like Figure 2 As shown, the millimeter-wave base station BS and the k-th wireless sensor S k millimeter-wave channel h k It can be modeled as:

[0104]

[0105] in, Indicates the orientation towards the k-th wireless sensor S k The array guide vector, φ k This represents the millimeter-wave base station BS and the k-th wireless sensor S. k The line-of-sight link departure angle between them; This represents the wide-range path fading of millimeter-wave channels, d k This represents the millimeter-wave base station BS and the k-th wireless sensor S.k The straight-line distance between them, where α represents the path loss coefficient; g k This represents the millimeter-wave base station BS and the k-th wireless sensor S. k millimeter-wave channel h k Small-scale path decay.

[0106] In practical wireless sensor networks, the number of radio frequency links is often less than the number of wireless sensors. Therefore, it is necessary to group the K sensors to improve the efficiency of energy harvesting. If there are too many wireless sensors on each radio frequency link, the energy received by each sensor will be greatly reduced.

[0107] Therefore, in order to ensure the energy harvesting efficiency of each sensor, each radio frequency link can serve a maximum of two sensors;

[0108] like Figure 3 As shown, according to the transmission protocol, the entire transmission process can be divided into two stages:

[0109] In the first phase, the base station BS generates multiple energy beams within the downlink transmission time slot τ0 to provide power to K wireless sensors. Therefore, the wireless sensors S associated with the RF link r k The received energy signal can be represented as:

[0110]

[0111] Among them, matrix Represents matrix h k The Hermitian matrix, w r and w t Let p represent the analog beamforming vectors for the r-th RF link and the t-th RF link, respectively; k,r Represented as sensor S k The transmit power allocated to the r-th radio frequency link; if the wireless sensor S k If the r-th radio frequency link is associated, then the wireless sensor S k Correlation coefficient u k,r =1, otherwise the correlation coefficient u k,r =0; e k Indicates a wireless sensor S k unit energy signal; n k Indicates wireless sensor S k The received noise; therefore, the wireless sensor S k The received radio frequency power is:

[0112]

[0113] Wireless sensor S k The collected energy can be expressed as:

[0114]

[0115] Where η represents the energy conversion efficiency, η∈(0,1);

[0116] In the second stage, within the remaining transmission time slots 1-τ0, K wireless sensors use the collected energy to transmit their respective data back to the millimeter-wave base station BS via time-division multiplexing on the uplink; at this time, wireless sensor S k The transmission power is expressed as Where τ k Indicates wireless sensor S k The transmission time slot; at this time, the millimeter-wave base station BS receives the wireless sensor S k The achievable transmission rate r of the transmitted signal k for:

[0117]

[0118] Among them, h k Indicates wireless sensor S k Millimeter-wave channel between the base station (BS) and the millimeter-wave channel.

[0119] Because millimeter waves have very narrow beamwidths, they are generally not suitable for providing power transfer services for large-area wireless sensors. Therefore, if a group of wireless sensors is located in a single millimeter wave coverage area, a single beam can be used for power transfer; otherwise, beam splitting technology is used to divide all antenna elements into several subarrays, generating multiple simulated beams to provide power transfer for multiple sensors within a group, with each beam directed to one wireless sensor.

[0120] The k-th wireless sensor S k and the i-th wireless sensor S i Form a group and associate it with the r-th radio frequency link, the k-th wireless sensor S k and the i-th wireless sensor S i The angles between the base station and the millimeter-wave base station are φ k and φ i Assume φ k and φ i The angular difference between them is less than or equal to one millimeter wavewidth φ B That is, |φ k -φ i |≤φ B The k-th wireless sensor S can be implemented using a single analog beam. k and the i-th wireless sensor S i Simultaneously, it provides power transmission; at this time, the r-th RF link uses a single analog beam. To transfer energy, It can be represented as:

[0121]

[0122] in, This indicates the aperture direction of the millimeter-wave base station BS on the r-th radio frequency link;

[0123] Furthermore, if the k′-th wireless sensor S is associated with the t-th radio frequency link k′ and the i′-th wireless sensor S i′ The absolute angular difference is greater than one millimeter wavewidth φ B That is, |φ k′ -φ i′ |>φ B Then, beam splitting technology is used to separately use and The antenna elements form two sub-beams, which are respectively directed to the k′-th wireless sensor S. k′ and the i′-th wireless sensor S i′ Provide energy services, of which and satisfy At this moment, the k′-th wireless sensor S is pointed to. k′ and the i′-th wireless sensor S i′ beam vector and They are represented as follows:

[0124]

[0125]

[0126] At this time, M is associated with the t-th radio frequency link. RF The analog beamforming formed by the antenna can be expressed as:

[0127]

[0128] In the constructed single-cell wireless sensor network model based on millimeter-wave energy harvesting, the millimeter-wave base station BS first sends energy signals to K wireless sensors through the downlink, and then the wireless sensors use the harvested energy to send their respective signals to the millimeter-wave base station BS through the uplink in the TDD manner.

[0129] Therefore, the total transmission rate of wireless sensors will become an important indicator for judging system performance. By analyzing the achievable rate expression (5) of the received signal of the millimeter-wave base station (BS), it can be seen that the simulated energy beamforming depends on the grouping of K wireless sensors and the antenna allocation of each link. At the same time, effective energy harvesting and information transmission also depend on the power allocation from the millimeter-wave base station (BS) to the wireless sensors and the time slot allocation for uplink and downlink transmission. Therefore, if we want to improve the energy efficiency of the single-cell wireless sensor network model based on millimeter-wave energy harvesting, we need to jointly optimize the grouping of wireless sensors, base station antenna allocation, power allocation and transmission time slot allocation. In order to effectively obtain these parameters, this invention proposes a two-stage optimization design method, as follows:

[0130] In the first stage, by fixing the power allocation coefficient and aiming to maximize the total energy received by K wireless sensors, the optimal wireless sensor grouping and antenna allocation information associated with each radio frequency link are obtained by adopting the alliance formation game theory.

[0131] In the second stage, based on the wireless sensor grouping and antenna allocation information obtained in the first stage, the optimal power allocation coefficient is obtained by using the linear programming problem in convex optimization with the goal of maximizing the total energy acquired by the K wireless sensors. Then, with the goal of maximizing the system energy efficiency, the optimal time allocation for uplink and downlink transmission is obtained by using the Lagrange dual multiplier method.

[0132] S2. With a fixed power allocation coefficient, the optimal wireless sensor grouping and antenna allocation are obtained by using a coalition-forming game theory approach, with the goal of maximizing the total energy received by all wireless sensors.

[0133] The optimization problem is defined as follows:

[0134] Assuming equal power allocation is applied to the K wireless sensors, and beam tracking technology is used, the millimeter-wave base station BS can ensure that it acquires the channel state and location information of all wireless sensors, including the k-th wireless sensor S. k Distance d from millimeter-wave base station BS k and angle φ k Based on grouping and antenna allocation, the optimization problem of maximizing the energy collected by all wireless sensors can be expressed as (P1):

[0135]

[0136] p k,r =P BS Substituting / K into formula (3), the objective function in the first optimization problem (P1) is... Represents the sum of conditionally received radio frequency power of all wireless sensors; constraint C1 indicates that if the k-th wireless sensor Sk If the r-th radio frequency link is associated, then the association coefficient u k,r =1, otherwise the correlation coefficient u k,r =0; Constraint C2 indicates that each RF link can provide power transfer services to at most two wireless sensors; Constraint C3 indicates that each wireless sensor can be associated with at most one RF link; Constraint C4 ensures that the number of antennas allocated on the r-th RF link does not exceed M. RF Constraint C5 guarantees that, under beam splitting conditions, the antenna associated with the r-th RF link is allocated to the wireless sensor S. k The minimum number of antennas is M min In the first optimization problem (P1), the objective function is... The effective channel gain follows a triangular periodic function of the number of antennas. Therefore, the optimization problem under consideration is a non-convex integer programming problem. It is impossible to directly obtain the optimized wireless sensor grouping and antenna allocation information using convex optimization. Therefore, the problem to be solved is transformed into a coalition formation game problem to obtain the optimized grouping and antenna allocation.

[0137] The specific game problem of alliance formation is as follows:

[0138] First, consider all K wireless sensors that need to harvest energy as a set. Several alliances are formed through joint participation in the game to maximize the total radio frequency power collected by all wireless sensors. The payoff of alliance G in the game is represented by V(G), which consists of a set of payoff vectors. Each element in the vector represents the payoff of the corresponding wireless sensor in alliance G, i.e., the collected radio frequency power, which is expressed by formula (3). The payoff V(G) of alliance G in the game is not only related to its own participation in the alliance, but also to the cooperation structure of other participants. At the same time, each sensor in alliance G will also be affected by antenna allocation in the case of beam splitting. Therefore, the game has the characteristic of non-transferable utility. In summary, the game under consideration can be regarded as an alliance formation game problem.

[0139] Alliance G associated with the r-th radio frequency link r In this context, K wireless sensors form several consortiums, denoted as B = {G1, ..., G...} N},r∈{1,…,N},V(G r The value of alliance group B is related to that of alliance G. r Internal antenna allocation strategy M r Antenna allocation strategies of other alliances Ξ\{M r} related, where Ξ={M1,…,M l};

[0140] Therefore, the alliance value V(G) associated with the r-th RF linkr B) can be represented as:

[0141] V(G r M r ,B,Ξ\{M r})=v k (G r M r ,B,Ξ\{M r}); (11)

[0142] in, Indicates wireless sensor S k In Alliance G r The proposed optimization algorithm, based on the characteristics of alliance-based games, aims to continuously iterate among wireless sensors to form optimized alliances based on the preference relationships among the alliances. That is, each wireless sensor compares its contribution in different alliances and decides whether to join other alliances or stay in the current alliance. Through several iterations, the total received radio frequency power of all wireless sensors is improved.

[0143] The alliance is set up to form a preference relationship, as follows:

[0144] Alliance Formation Preference Relationship 1: In the iterth iteration, assume the current set of wireless sensors... The alliance groups and the antenna allocation strategies corresponding to each alliance within the alliance groups are respectively B iter ={G1,…,G l} and Ξ iter ={M1,…,M l The wireless sensor S is valid if and only if the following condition is met. k Will leave the current alliance G r And join another alliance G associated with the t-th RF link. t :

[0145]

[0146] Among them, the newly formed alliance groups in the (iter+1)th iteration Represented as Since a consortium can accommodate a maximum of two wireless sensors, when wireless sensor S k Leaving the original alliance G r After that, the new alliance {G r The antenna allocation in \{k}} is represented as follows:

[0147]

[0148] When wireless sensor S k Join the alliance G tAt that time, a one-dimensional search is performed on the antenna allocation within the newly formed coalition group, while the antenna allocation of the remaining coalitions is fixed. Find newly formed alliances {G t The maximum received RF power of ∪{k}}, at which point the antenna allocation within the newly formed alliance group is the optimal antenna allocation for that alliance. At this time, the newly formed alliance group B iter+1 The antenna allocation strategy in can be expressed as: In this preference relationship, only the coalition G t When the number of wireless sensors in the system is less than 2, the wireless sensor S k That's why they tried to join the alliance G. t ;

[0149] Alliance Formation Preference Relationship 2: In the iterth iteration, assume the current wireless sensor set... The alliance groups and the antenna allocation strategies corresponding to each alliance within the alliance groups are respectively B iter ={G1,…,G l} and Ξ iter ={M1,…,M l The wireless sensor S is valid if and only if the following condition is met. k ∈G r and wireless sensor S k′ ∈G t They will exchange positions:

[0150]

[0151] Among them, the newly formed alliance group B in the (iter+1)th iteration iter+1 It can be represented as B iter+1 ={B iter \{G r G t}}∪{G r \{k}∪{k′},G t \{k′}∪{k}}; By performing a one-dimensional search on the antenna allocation within the two newly formed alliances, while fixing the antenna allocation of the remaining alliances {Ξ iter+1 \{M r\{k}∪k′ M t\{k′}∪k}}, find a newly formed alliance G r \{k}∪{k′} and G t The maximum received radio frequency power of \{k′}∪{k}, at which point a new alliance group G is formed. r \{k}∪{k′} and G t The antenna allocation corresponding to \{k′}∪{k} is the optimal antenna allocation for the alliance. and and It can be done by [M] min M RF -M min The integers within the range are obtained by performing a two-dimensional full search, where the computational complexity of obtaining the maximum value is acceptable.

[0152] The wireless sensor grouping and antenna allocation algorithm specifically includes the following steps:

[0153] S2.1. Initialize the algorithm based on the number of RF links N and the number of wireless sensors, as follows:

[0154] When the number of wireless sensors K=N, each radio frequency link randomly corresponds to one wireless sensor, that is, each wireless sensor forms a separate alliance and exclusively uses all the antennas of that radio frequency link;

[0155] When the number of wireless sensors is 2N>K>N, the wireless sensors are randomly divided into alliances. Some alliances have two wireless sensors corresponding to one radio frequency link and the antennas of the radio frequency link are evenly distributed. Some alliances have only one wireless sensor corresponding to one radio frequency link and exclusively enjoy all the antennas of the link.

[0156] When the number of wireless sensors K≥2N, 2N sensors are randomly selected to form N alliances. Each alliance contains two wireless sensors corresponding to one radio frequency link. The wireless sensors are evenly distributed among the antennas of the radio frequency link.

[0157] At this point, the initialization settings can be expressed as: define the initialization iteration index iter = 0, and initialize the alliance group as B0 = {G1, ..., G}. N The antenna allocation strategy for each alliance is Ξ0 = {M1, ..., M}. N};

[0158] S2.2 In the iterth iteration stage, any given member of the union G... r The wireless sensors will access other alliance Gs one by one. t If wireless sensor S k Join the alliance G t When the number of wireless sensors in the alliance is less than or equal to 2, the alliance forming preference relation 1 is used to determine the wireless sensor S. k Should they leave their original alliance G? r Join the alliance G t If wireless sensor S k Join the alliance G t If the number of wireless sensors in the alliance is greater than 2, then alliance preference relation 2 will be adopted to belong to alliance G. r wireless sensor S k With Alliance Gt Wireless sensor S in k′ They exchange positions, and the wireless sensor S is determined based on the calculated radio frequency power collected by all wireless sensors. k With wireless sensor S k′ Should we perform a position swap to obtain a better alliance group?

[0159] S2.3 After multiple iterations, if neither Alliance Formation Preference Relation 1 nor Alliance Formation Preference Relation 2 can form a new alliance group, then the corresponding alliance group B... iter The optimal wireless sensor grouping, and the antenna allocation strategy Ξ for each alliance within the alliance group. iter ={M1,…,M N} represents the optimal antenna allocation strategy.

[0160] S3. Based on the obtained wireless sensor grouping and antenna allocation information, the optimal power allocation of each wireless sensor during the energy acquisition time is obtained by using the linear programming optimization method, so as to maximize the total acquisition energy of all sensors. Then, using the Lagrange dual multiplier method and KKT conditions, with the goal of maximizing the system energy efficiency, the optimal allocation between uplink and downlink transmissions of the single-cell wireless sensor network model is obtained.

[0161] Based on the optimal wireless sensor grouping and antenna allocation information obtained in step S2, further optimization of power allocation and time allocation is needed to ensure maximum total energy collection by the wireless sensors and energy efficiency of the base station (BS) received signals, as follows:

[0162] Optimized power allocation: After obtaining the optimal wireless sensor grouping and antenna allocation, the optimal power allocation of the millimeter-wave base station (BS) for each wireless sensor during the energy harvesting phase can further increase the total harvesting energy of all wireless sensors. First, assuming that the downlink transmission time slot τ0 and the uplink transmission time slot (1-τ0) are constants, the problem of maximizing the total harvesting energy of the wireless sensors can be expressed as a second optimization problem (P2):

[0163]

[0164] Substituting the optimized wireless sensor grouping and antenna allocation information into equation (3) yields the following result: The constraint C6 in equation (15) guarantees that the wireless sensor S k The minimum threshold value for the received radio frequency power is P RF Constraint C7 states that the total energy allocated by the millimeter-wave base station (BS) to all wireless sensors is no greater than the transmission power P of the millimeter-wave base station (BS). BSThe constraint C8 indicates that the energy allocated by the millimeter-wave base station (BS) to each wireless sensor must be greater than or equal to 0; according to the second optimization problem (P2), the objective function is... Since both constraints C6 and C7 are linear functions, the second optimization problem (P2) is a linear optimization problem, which can be solved using the corresponding optimization tools in Matlab.

[0165] Optimizing time slot allocation: In the uplink transmission time slot, each wireless sensor uses the collected energy to transmit its own signal to the millimeter-wave base station (BS) in TDD mode. Therefore, by optimizing the energy collection time slot and the information transmission time slot of each wireless sensor, the energy efficiency of the millimeter-wave base station (BS) in receiving signals can be maximized, which can be defined as:

[0166]

[0167] Among them, R sum and E T These represent the total achievable rate of signal reception at the base station (BS) and the total transmit energy of the wireless sensor, respectively, as follows:

[0168]

[0169] This indicates that under optimal conditions of wireless sensor grouping, antenna allocation, and power allocation, the wireless sensor S... k The radio frequency energy collected during downlink transmission; therefore, the third optimization problem (P3) to maximize the energy efficiency of the millimeter-wave base station (BS) received signal can be expressed as:

[0170]

[0171] By analyzing the third optimization problem (P3), its objective function and constraint C9 are concave and linear functions, respectively. Therefore, the objective function in the third optimization problem (P3) can be transformed into its corresponding parametric form, and the optimal time slot allocation can be obtained by using the Lagrange dual multiplier method. The parameterized form of the objective function in the third optimization problem (P3) can be expressed as:

[0172]

[0173] Where θ represents a non-negative parameter; for a given value of θ, the Lagrangian function corresponding to the third optimization problem (P3) can be expressed as:

[0174]

[0175] Here, ξ represents the Lagrange multiplier.

[0176] In this embodiment, simulation analysis verifies the effectiveness of the proposed resource optimization allocation method in improving the energy efficiency of sensors in millimeter-wave wireless power supply wireless sensor networks, and compares it with a single-beam scheme based on time-division multiplexing and a multi-beam scheme based on fixed power allocation and time slot allocation. In the single-beam scheme based on time-division multiplexing, uplink and downlink transmissions are evenly allocated across the entire transmission time slot: assuming there are K wireless sensors, the first half of the transmission time slot is divided into K sub-time slots, and the base station (BS) will wirelessly power one sensor in each sub-time slot. The second half of the transmission time slot is also divided into K sub-time slots, and each sensor will be allocated one sub-time slot to send information to the base station (BS). In the multi-beam scheme based on fixed power allocation and time slot allocation, the entire transmission time slot will also be divided into uplink and downlink transmission time slots: in the first half of the transmission time slot, the base station BS will simultaneously transmit wireless power to K sensors, wherein user grouping and antenna allocation adopt the method mentioned in this invention, and the base station will evenly allocate transmission power to the K sensors. In the second half of the transmission time slot, the transmission time slot will be divided into K sub-time slots, and each sensor will be allocated a sub-time slot to send information to the base station BS.

[0177] Figure 4 The paper demonstrates the total energy received by all wireless sensors in the considered millimeter-wave wireless sensor network when employing the resource optimization allocation scheme proposed in this invention, and compares it with a single-beam scheme based on time-division multiplexing and a multi-beam scheme based on fixed power allocation and time slot allocation. The simulation results show that the received energy of all wireless sensors in the above three schemes increases with the transmit power P of the base station (BS). BS The energy received by wireless sensors increases with increasing power and time slot allocation. Under a multi-beam scheme with fixed power and time slot allocation, the wireless sensors can receive the most energy. Under a resource-optimized allocation scheme, all wireless sensors collect more energy than when using a single-beam scheme based on time division multiplexing. While the multi-beam scheme with fixed power and time slot allocation can provide more energy to wireless sensors, it consumes too many energy transmission time slots. This leads to a reduction in uplink information transmission time slots, ultimately resulting in a decrease in the information transmission rate and energy efficiency of the wireless sensors.

[0178] Figure 5 This paper demonstrates the total rate at which the base station receives information transmitted by all wireless sensors in a millimeter-wave wireless sensor network using the resource optimization allocation scheme proposed in this invention. The results are compared with a single-beam scheme based on time-division multiplexing and a multi-beam scheme based on fixed power allocation and time slot allocation. Simulation results show that the total rate at which the base station (BS) receives information transmitted by the wireless sensors under the resource optimization allocation scheme proposed in this invention is significantly better than the other two schemes, highlighting the crucial role of resource optimization allocation in improving the total rate of information transmitted by wireless sensors.

[0179] Figure 6 The system energy efficiency (ratio of total base station reception rate to total base station reception rate) of the resource optimization allocation scheme proposed in this invention is demonstrated in the considered millimeter-wave wireless sensor network, and compared with a single-beam scheme based on time-division multiplexing and a multi-beam scheme based on fixed power allocation and time slot allocation. Simulation results show that the system energy efficiency under the resource optimization allocation scheme proposed in this invention is significantly better than the other two schemes, thus verifying that the resource optimization allocation method proposed in this invention can effectively improve the energy transmission efficiency of millimeter-wave wireless sensor networks.

[0180] Example 2:

[0181] In this embodiment, the number of users K = 8 and the number of radio frequency links N = 5.

[0182] Example 3:

[0183] In this embodiment, the number of users K = 9 and the number of radio frequency links N = 6.

Claims

1. A method for optimizing resource allocation in wireless sensor networks based on millimeter-wave wireless power transmission, characterized in that, Includes the following steps: S1. Construct a single-cell wireless sensor network model based on millimeter-wave energy harvesting, including a millimeter-wave base station BS and K One wireless sensor; S2. With a fixed power allocation coefficient, the optimal wireless sensor grouping and antenna allocation are obtained by using a coalition-forming game theory approach, with the goal of maximizing the total energy received by all wireless sensors. S3. Based on the obtained wireless sensor grouping and antenna allocation information, the optimal power allocation of each wireless sensor during the energy acquisition time is obtained by using the linear programming optimization method, so as to maximize the total acquisition energy of all sensors. Then, using the Lagrange dual multiplier method and KKT conditions, with the goal of maximizing the system energy efficiency, the optimal allocation between uplink and downlink transmissions of the single-cell wireless sensor network model is obtained.

2. The method for optimizing resource allocation in wireless sensor networks based on millimeter-wave wireless power transmission according to claim 1, characterized in that, In step S1, K A wireless sensor at a density of The Poisson distribution is deployed on a radius of Inside the circle, the millimeter-wave base station (BS) is placed at the center. Millimeter-wave base stations (BS) have a continuous power supply. Since wireless sensors do not have a continuous power supply, wireless sensor networks adopt a 'collect energy first, then transmit information' transmission protocol, that is, in the downlink transmission time slot Inside, the millimeter-wave base station (BS) generates multiple energy beams to power the wireless sensor, which then uses the acquired energy in the uplink transmission time slot. The internal transmission uses Time Division Multiplexing (TDD) to send information back to the millimeter-wave base station (BS), where the first... k One wireless sensor The allocated transmission time slots are ; Millimeter-wave base station (BS) is equipped with a... M A linear antenna array composed of three antennas forms N Each radio frequency link; sets the number of antennas allocated to each radio frequency link. Equal, that is ,if If the value is not an integer, the remaining antennas are assigned to the last RF link. All wireless sensors are equipped with only a single antenna and are connected to only one radio frequency link.

3. The method for optimizing resource allocation in wireless sensor networks based on millimeter-wave wireless power transmission according to claim 1, characterized in that, In step S1, the millimeter-wave base station BS and the first k One wireless sensor millimeter-wave channels between The model is as follows: (1) in, Indicates facing the first k One wireless sensor The array guide vector, Indicates millimeter-wave base station BS and the first k One wireless sensor The line-of-sight link departure angle between them; This indicates the wide-range path fading of millimeter-wave channels. Indicates millimeter-wave base station BS and the first k One wireless sensor The straight-line distance between them Indicates the path loss coefficient; Indicates millimeter-wave base station BS and the first k One wireless sensor millimeter-wave channels between Small-scale path decay.

4. The method for optimizing resource allocation in wireless sensor networks based on millimeter-wave wireless power transmission according to claim 1, characterized in that, In step S1, for K The sensors are grouped to improve the efficiency of energy harvesting; Each radio frequency link can serve a maximum of two sensors; According to the transmission protocol, the entire transmission process is divided into two stages: In the first phase, the base station (BS) is in the downlink transmission time slot. Multiple energy beams are generated internally. K Each wireless sensor provides power, thus connecting to the radio frequency link. r Associated wireless sensors The received energy signal is represented as: (2) Among them, matrix Representation matrix Hermitian matrix, and They represent the orientation towards the first The first radio frequency link and the first The simulated beamforming vector of the radio frequency link; Represented as a sensor The first r The transmit power allocated to each radio frequency link; if the wireless sensor Related to the first r One radio frequency link, then wireless sensor Correlation coefficient Otherwise, the correlation coefficient ; Indicates orientation towards wireless sensors The unit energy signal; Indicates wireless sensor The received noise; therefore, wireless sensors The received radio frequency power is: (3) Wireless sensors The collected energy is represented as follows: (4) in, Indicates energy conversion efficiency. ; In the second phase, during the remaining transmission time slots... Inside, K Each wireless sensor uses the collected energy to transmit its data back to the millimeter-wave base station (BS) via time-division multiplexing on the uplink; at this time, the wireless sensors... The transmission power is expressed as ,in Indicates wireless sensor The transmission time slot; at this time, the millimeter-wave base station (BS) receives the wireless sensor achievable transmission rate of the transmitted signal for: (5) in, Indicates wireless sensor Millimeter-wave channel between the base station (BS) and the millimeter-wave channel.

5. The method for optimizing resource allocation in wireless sensor networks based on millimeter-wave wireless power transmission according to claim 4, characterized in that, If a group of wireless sensors is located in a single millimeter-wave coverage area, a single beam is used to achieve power transfer; otherwise, beam splitting technology is used to divide all antenna elements into several subarrays, generating multiple simulated beams to provide power transfer for multiple sensors within a group, with each beam directed to a wireless sensor. No. k One wireless sensor and the i One wireless sensor Form a group and with the first The first radio frequency link is associated with the second radio frequency link. k One wireless sensor and the i One wireless sensor The angles between the base station and the millimeter-wave base station are respectively and ; Assumption and The angular difference between them is less than or equal to one millimeter wavewidth. ,Right now This can be achieved using a single analog beam. k One wireless sensor and the i One wireless sensor Simultaneously providing energy transfer, at this time, the first r The radio frequency link uses a single analog beam. To transfer energy, Represented as: (6) in, This indicates that the millimeter-wave base station BS is in the... The aperture direction of the radio frequency link; Furthermore, if with the first The first radio frequency link associated with One wireless sensor and the One wireless sensor The absolute angular difference is greater than one millimeter wavewidth. ,Right now Then, beam splitting technology is used to separately utilize... and The antenna elements form two sub-beams, which respectively supply the first... One wireless sensor and the One wireless sensor Provide energy services, of which and satisfy At this point, it points to the first One wireless sensor and the One wireless sensor beam vector and They are represented as follows: ；(7) ； (8) At this time and the first Related to a radio frequency link The analog beamforming formed by the antenna is represented as follows: (9) In the constructed single-cell wireless sensor network model based on millimeter-wave energy harvesting, the millimeter-wave base station (BS) first transmits data to the network via the downlink. K Each wireless sensor sends an energy signal, and then the wireless sensor uses the collected energy to send its own signal to the millimeter-wave base station BS via the uplink in a TDD manner.

6. The method for optimizing resource allocation in wireless sensor networks based on millimeter-wave wireless power transmission according to claim 5, characterized in that, The sum of the transmission rates of wireless sensors will become an important indicator for judging system performance; according to the achievable rate expression (5) of the received signal of the millimeter-wave base station BS, the simulated energy beamforming depends on K The grouping of wireless sensors and the antenna allocation of each link are related. At the same time, energy harvesting and information transmission also depend on the power allocation from the millimeter-wave base station (BS) to the wireless sensors and the time slot allocation for uplink and downlink transmission. Therefore, if we want to improve the energy efficiency of the single-cell wireless sensor network model based on millimeter-wave energy harvesting, we need to jointly optimize the grouping of wireless sensors, base station antenna allocation, power allocation and transmission time slot allocation. A two-stage optimization design method is adopted, as follows: In the first stage, by fixing the power allocation factor, K With the goal of maximizing the total energy received by each wireless sensor, the optimal wireless sensor grouping and antenna allocation information associated with each radio frequency link are obtained by using the alliance formation game theory. In the second phase, based on the wireless sensor grouping and antenna allocation information obtained in the first phase, firstly... K With the goal of maximizing the total energy acquired by the wireless sensors, the optimal power allocation coefficient is obtained by using linear programming in convex optimization. Then, with the goal of maximizing the system energy efficiency, the optimal time allocation for uplink and downlink transmission is obtained by using the Lagrange dual multiplier method.

7. The method for optimizing resource allocation in wireless sensor networks based on millimeter-wave wireless power transmission according to claim 6, characterized in that, In step S2, the optimization problem is defined as follows: Assuming that K The wireless sensors are allocated equal power, and beam tracking technology is used to ensure that the millimeter-wave base station (BS) acquires the channel state and location information of all wireless sensors, including the first... k One wireless sensor Distance from millimeter-wave base station (BS) and angle Based on grouping and antenna allocation, the optimization problem of maximizing the energy collected by all wireless sensors is expressed as the first optimization problem (P1): (10) Will Substituting into formula (3), the objective function in the first optimization problem (P1) This represents the sum of the conditional receive radio frequency power of all wireless sensors; Constraint C1 means that if the first... k One wireless sensor Related to the first For each radio frequency link, the correlation coefficient is... Otherwise, the correlation coefficient Constraint C2 indicates that each RF link can provide power transfer services for a maximum of two wireless sensors. Constraint C3 states that each wireless sensor can be associated with at most one radio frequency link; Constraint C4 ensures that in the first... The number of antennas allocated on each radio frequency link shall not exceed Constraint C5 guarantees that, under beam splitting conditions, it is consistent with the first... The antenna associated with each radio frequency link is assigned to the wireless sensor. The minimum number of antennas is In the first optimization problem (P1), the objective function is... The effective channel gain follows a triangular periodic function of the number of antennas. Therefore, the optimization problem under consideration is a non-convex integer programming problem. It is impossible to directly obtain the optimized wireless sensor grouping and antenna allocation information using convex optimization. Therefore, the problem to be solved is transformed into a coalition formation game problem to obtain the optimized grouping and antenna allocation. The specific game problem of alliance formation is as follows: First, put all K A wireless sensor that needs to harvest energy is considered as a set. Several alliances are formed through joint participation in the game to maximize the total radio frequency power collected by all wireless sensors; alliances G The payoff in a game is expressed as It consists of a set of payoff vectors, where each element represents the corresponding wireless sensor's contribution to the alliance. G The benefit, i.e., the collected radio frequency power, is expressed by formula (3). G Gains in the game The value of depends not only on the participants within its own alliance, but also on the cooperative structure of other participants; at the same time, the alliance G Each sensor is also affected by antenna allocation in the case of beam splitting, so the game has a non-transferable utility characteristic; in summary, the game under consideration is regarded as a coalition formation game problem. With the Alliances associated with radio frequency links middle, K Several wireless sensors form several alliance groups, denoted as , , In the alliance group The value and alliance Internal antenna allocation strategy Antenna allocation strategies with other alliances Related, among which ; Therefore, with the first Alliance value associated with each radio frequency link Represented as: ；(11) in, Indicates wireless sensor In the league The proposed optimization algorithm, based on the characteristics of alliance-based games, aims to continuously iterate among wireless sensors to form optimized alliances based on the preference relationships among the alliances. That is, each wireless sensor compares its contribution in different alliances and decides whether to join other alliances or stay in the current alliance. Through several iterations, the total received radio frequency power of all wireless sensors is improved.

8. The method for optimizing resource allocation in wireless sensor networks based on millimeter-wave wireless power transmission according to claim 7, characterized in that, The alliance is set up to form a preference relationship, as follows: Alliance Formation Preference Relationship 1: In the iter In this iteration, assume the current set of wireless sensors The alliance groups and the antenna allocation strategies corresponding to each alliance in the alliance group are as follows: and A wireless sensor is valid if and only if the following conditions are met. Will leave the current alliance And join another with the first Alliances associated with radio frequency links : ；(12) Among them, the iter+1 New alliances formed in the next iteration Represented as Since a consortium can accommodate a maximum of two wireless sensors, therefore when wireless sensors Leave the original alliance After that, the new alliance The antenna allocation in the diagram is represented as follows: ；(13) When wireless sensor Join the alliance At that time, a one-dimensional search is performed on the antenna allocation within the newly formed coalition group, while the antenna allocation of the remaining coalitions is fixed. Find new alliances The maximum received radio frequency power, at which point the antenna allocation within the newly formed alliance group is the optimal antenna allocation for that alliance. The newly formed alliance group at this time The antenna allocation strategy in the text is expressed as: In this preference relationship, only alliances... When the number of wireless sensors in the system is less than 2, the wireless sensors That's why they try to join the alliance. ; Alliance Formation Preference Relationship 2: In the iter In this iteration, assume the current set of wireless sensors The alliance groups and the antenna allocation strategies corresponding to each alliance in the alliance groups are respectively and A wireless sensor is valid if and only if the following conditions are met. and wireless sensors They will exchange positions: ；(14) Among them, the iter+1 New alliances formed in the next iteration Represented as By performing a one-dimensional search on the antenna allocation within the two newly formed coalitions, while fixing the antenna allocation of the remaining coalitions. Find new alliances and The maximum received radio frequency power, at which point a new alliance group is formed. and The antenna allocation corresponding to this is the optimal antenna allocation for the alliance. and ; and Through the The integers within the range are obtained by performing a two-dimensional full search.

9. The method for optimizing resource allocation in wireless sensor networks based on millimeter-wave wireless power transmission according to claim 8, characterized in that, The wireless sensor grouping and antenna allocation algorithm specifically includes the following steps: S2.1, Based on the number of RF links N The algorithm is initialized based on the number of wireless sensors, as follows: When the number of wireless sensors K = N At that time, each radio frequency link randomly corresponds to a wireless sensor, that is, each wireless sensor forms a separate alliance and exclusively uses all the antennas of that radio frequency link; When the number of wireless sensors At that time, wireless sensors are randomly divided into alliances. Some alliances have two wireless sensors corresponding to one radio frequency link and the antennas of the radio frequency link are evenly distributed. Some alliances have only one wireless sensor corresponding to one radio frequency link and exclusively enjoy all the antennas of the link. When the number of wireless sensors At that time, randomly select 2 N Each sensor and form N Each alliance consists of two wireless sensors corresponding to one radio frequency link, with the wireless sensors evenly distributing the antennas of that radio frequency link. At this point, the initialization setting is represented as: defining the initialization iteration index. iter =0, initialize the alliance group as The antenna allocation strategy corresponding to each alliance is as follows: ; S2.2, in the iter In the next iteration phase, any given member of the alliance The wireless sensors will access other alliances one by one. If wireless sensors Join the alliance When the number of wireless sensors in the alliance is less than or equal to 2, the alliance forming preference relation 1 is used to determine the wireless sensor. Should they leave their original alliance? Join the alliance If wireless sensors Join the alliance If the number of wireless sensors in the alliance is greater than 2, then alliance preference relation 2 will belong to the alliance. wireless sensors With Alliance Wireless sensors They exchange positions and determine the wireless sensor based on the calculated radio frequency power collected by all wireless sensors. With wireless sensors Should we perform a position swap to obtain a better alliance group? S2.3 After multiple iterations, if neither Alliance Formation Preference Relation 1 nor Alliance Formation Preference Relation 2 can form a new alliance group, then the corresponding alliance group... For optimal wireless sensor grouping, and the antenna allocation strategy for each alliance within the alliance group. The optimal antenna allocation strategy.

10. The method for optimizing resource allocation in wireless sensor networks based on millimeter-wave wireless power transmission according to any one of claims 1 to 9, characterized in that, In step S3, based on the optimal wireless sensor grouping and antenna allocation information obtained in step S2, it is necessary to further optimize power allocation and time allocation to ensure that the total collection energy of the wireless sensors and the energy efficiency of the base station (BS) received signals are maximized, as follows: Optimized power allocation: After obtaining the optimal wireless sensor grouping and antenna allocation, the millimeter-wave base station (BS) further increases the total harvesting energy of all wireless sensors by allocating optimal power to each wireless sensor during the energy harvesting phase; firstly, assuming downlink transmission time slots... and uplink transmission time slot For a given value, the problem of maximizing the total energy collected by the wireless sensor can be expressed as a second optimization problem (P2): (15) Substituting the optimized wireless sensor grouping and antenna allocation information into equation (3) yields the following result. Constraint C6 in equation (15) guarantees the wireless sensor The minimum threshold value for the received radio frequency power is Constraint C7 states that the total energy allocated by the millimeter-wave base station (BS) to all wireless sensors is no greater than the transmission power of the millimeter-wave base station (BS). The constraint C8 indicates that the energy allocated by the millimeter-wave base station (BS) to each wireless sensor must be greater than or equal to 0; according to the second optimization problem (P2), the objective function is... Since both constraints C6 and C7 are linear functions, the second optimization problem (P2) is a linear optimization problem, which can be solved using the corresponding optimization tools in Matlab. Optimize time slot allocation: In the uplink transmission time slot, each wireless sensor uses the collected energy to transmit its own signal to the millimeter-wave base station (BS) in TDD mode; therefore, by optimizing the energy collection time slot and the information transmission time slot of each wireless sensor, the energy efficiency of the millimeter-wave base station (BS) in receiving signals is maximized, which is defined as: ；(16) in, and These represent the total achievable rate of signal reception at the base station (BS) and the total transmit energy of the wireless sensor, respectively, as follows: (17) This indicates that under optimal conditions of wireless sensor grouping, antenna allocation, and power allocation, the wireless sensor... The radio frequency energy collected during downlink transmission; therefore, the third optimization problem (P3) to maximize the energy efficiency of the millimeter-wave base station (BS) received signal is expressed as: ；(18) By analyzing the third optimization problem (P3), its objective function and constraint C9 are concave and linear functions, respectively. Therefore, by transforming the objective function in the third optimization problem (P3) into its corresponding parametric form, and then using the Lagrange dual multiplier method, the optimal time slot allocation can be obtained. The parameterized form of the objective function in the third optimization problem (P3) is as follows: ；(19) in, It is represented as a non-negative parameter; for a given... The Lagrangian function corresponding to the third optimization problem (P3) is expressed as follows: (20) in, It represents the Lagrange multiplier.

Images