Base station system, method, apparatus and device for power consumption optimization of a base station system

Abstract

The application relates to a base station system, a power consumption optimization method, device and equipment for the base station system, wherein the base station system comprises a hybrid architecture transmitter and a perception receiver; the hybrid architecture transmitter is used for simultaneously transmitting a communication signal to a plurality of information decoding receivers, transmitting a perception signal to a plurality of perception targets, and transmitting an energy signal to a plurality of energy collection receivers; the perception receiver is used for receiving echo signals of the plurality of perception targets; the hybrid architecture transmitter comprises a digital beamformer, a plurality of transmission links, an analog beamformer and a plurality of transmission antennas, the analog beamformer comprises a plurality of phase shifters, and the perception receiver comprises a plurality of receiving antennas. The problem that there is no research on the base station system for multiple communication users, multiple perception targets and multiple energy collection receivers in the prior art is solved.

Description

Technical Field

[0001] This application relates to the field of signal base stations, and in particular to a base station system, a power consumption optimization method, apparatus and device for a base station system. Background Technology

[0002] Integrating wireless sensing, power transfer, computing, and even control as new functions into future sixth-generation (6G) wireless networks has attracted increasing research interest to support diverse environmental sensing applications with large-scale distributed sensors, particularly in future Internet of Things (IoT) and industrial scenarios. The integration of multiple functions not only improves resource utilization efficiency and reduces system overhead but also promotes synergistic cooperation between different functions. Much current work focuses on integrating one or two of these functions, such as Integrated Communication and Sensing (ISAC) and Wireless Powered Communication (SWIPT). However, considering complex wireless communication systems combining multiple different functions remains a challenging task.

[0003] Several studies have already considered multifunctional wireless systems integrating sensing, communication, and power transfer capabilities. However, some of the earliest research advocating for the integration of sensing, communication, and wireless power supply only considered one specific technology, such as radio frequency identification (RFID) and sensing (RFID&S); others only studied the Pareto boundary of MIMO systems with a single communication user, a single sensing target, and a single energy harvesting receiver. Currently, no research has attempted to understand the energy efficiency or power consumption issues of multifunctional wireless systems integrating sensing, communication, and power transfer capabilities, nor has any research considered energy optimization design in general scenarios with multiple communication users, multiple sensing targets, and multiple energy harvesting receivers.

[0004] The integration of multiple functions inevitably leads to higher energy consumption. In 6G systems, challenges such as high energy consumption, high operating costs, and significant carbon emissions have emerged, attracting widespread research interest. To alleviate these issues, extensive research efforts are dedicated to achieving green and carbon-neutral communications. These efforts take multiple perspectives, covering technologies such as energy-efficient transmit beamforming and antenna selection, user association and base station activation, and the utilization of renewable energy in base stations. Among these, the joint optimization of transmit beamforming and antenna selection has been a focus of attention. This technology aims to improve the energy efficiency of multi-antenna base stations by designing transmit beamforming and selectively shutting down certain transmit antennas. Typically, finding the optimal antenna subarray relies on exhaustive search or branch-and-bound (BAB) search, which are limited by extremely high computational complexity. Furthermore, the use of machine learning (ML) to solve the antenna selection problem, which is essentially a combinatorial optimization problem, has attracted increasing attention. Although these studies have made some progress, they are all based on the assumption of fixed power amplifier efficiency, which does not hold in reality and may lead to reduced energy efficiency in system design.

[0005] Furthermore, the addition of hybrid beamforming brings more difficulties and challenges. Hybrid beamforming has been widely recognized as a reliable technique for controlling hardware overhead and further reducing power consumption through spatial multiplexing. In hybrid beamforming, a limited number of RF links are connected to a large-scale antenna array via an analog component network (composed of switches and / or phase shifters). Current work has explored the design of hybrid architectures from an energy efficiency perspective, providing valuable insights into the energy efficiency performance of hybrid architectures. Meanwhile, combining antenna selection with hybrid beamforming is a relatively new research topic. However, current discussions based on hybrid architectures are often limited to antenna selection, and exploration of RF link and phase shifter selection remains lacking.

[0006] In practical systems, the power amplifier power consumption of a base station accounts for approximately 50%-80% of its total power consumption, and the efficiency of each power amplifier is typically highly nonlinear. Specifically, maximum power amplifier efficiency is achieved only when the input signal power reaches the saturation point of the maximum output signal power. When the input signal power deviates significantly from this saturation point, the power amplifier efficiency may decrease significantly, leading to a rapid reduction in system energy efficiency. Some studies have considered practical nonlinear power amplifier efficiency models and revealed their advantages over traditional ideal linear models. On the other hand, energy harvesting models in practical systems are also nonlinear. In traditional linear energy harvesting models, the RF-to-DC power conversion efficiency is assumed to be independent of the input power level of the energy harvesting circuit, failing to accurately characterize the energy harvesting efficiency, which is actually related to the input power level. This discrepancy presents challenges for resource allocation. Currently, no research has focused on the optimization of multifunctional wireless network transmission based on practical nonlinear energy harvesting models.

[0007] There is currently no effective solution to the problem of the lack of research on base station systems for multiple communication users, multiple sensing targets, and multiple energy harvesting receivers in related technologies. Summary of the Invention

[0008] This invention provides a base station system, a power consumption optimization method, apparatus, and device for a base station system, in order to address the lack of research in related technologies on base station systems with multiple communication users, multiple sensing targets, and multiple energy harvesting receivers.

[0009] In a first aspect, the present invention provides a base station system that integrates communication, sensing and power transmission, the base station system comprising: a hybrid architecture transmitter and a sensing receiver;

[0010] The hybrid architecture transmitter is used to simultaneously send communication signals to multiple information decoding receivers, send sensing signals to multiple sensing targets, and send energy signals to multiple energy harvesting receivers. The sensing receivers are used to receive echo signals from multiple sensing targets.

[0011] The hybrid architecture transmitter includes a digital beamformer, multiple transmit links, an analog beamformer, and multiple transmit antennas. The analog beamformer and multiple transmit antennas include multiple phase shifters. The sensing receiver includes multiple receive antennas.

[0012] The number of transmission links is greater than the number of information decoding receivers and less than or equal to the number of transmission antennas, and the number of receiving antennas is greater than the number of transmission antennas.

[0013] Secondly, this invention provides a power consumption optimization method for a base station system, wherein the base station system is the integrated communication-sensing-power transmission base station system described in the first aspect, and the power consumption optimization method includes:

[0014] Establish a total power consumption model and system constraints for the base station system regarding the digital beamformer, the analog beamformer, and the covariance matrix used to generate the sensing signal. The total power consumption model includes the total power consumption of the transmit antenna, the total power consumption of the radio frequency link, and the total power consumption of the phase shifter.

[0015] Based on the total power consumption model and the system constraints, the covariance matrix of the digital beamformer, the analog beamformer, and the sensed signal is optimized with the goal of minimizing the total power consumption of the base station system.

[0016] In some of these embodiments, the system constraints include:

[0017] The constraints are as follows: minimum signal-to-interference-plus-noise ratio constraint of the information decoding receiver, maximum estimation Cramer-Rao bound constraint of the sensing target, minimum harvesting power constraint of the energy harvesting receiver, transmit power constraint of the transmitting antenna, and constant mode constraint of the analog beamformer.

[0018] In some of these embodiments, the total power consumption model is:

[0019]

[0020] in, The total power consumption of the base station system is represented by F, which represents the analog beamformer. k This represents the digital beamformer corresponding to the k-th signal decoding receiver, and S represents the covariance matrix of the sensed signal;

[0021] P PA (F,{w k},S) represents the total power consumption of the transmitting antenna:

[0022]

[0023] N T η represents the number of transmitting antennas. n η represents the power amplifier efficiency of the nth transmitting antenna. max This represents the maximum power amplifier efficiency of the transmitting antenna. This represents the actual transmit power of the nth transmitting antenna. K represents the maximum transmit power of the nth transmitting antenna, β represents the efficiency factor, and K ID Indicates the number of information decoding receivers. This represents a zero matrix where all elements except the nth diagonal element are 0. Indicates a length of N T A vector whose elements are all zero except for the nth element which is 1;

[0024] P RF ({w k},S) represents the total power consumption of the radio frequency link:

[0025]

[0026] N RF Indicates the number of radio frequency links. This indicates the power consumption of a single radio frequency link when it is enabled. Indicates a length of N RF A vector whose elements are all zeros except for the nth element which is 1. This represents a zero matrix where all elements except the nth diagonal element are 0, and I represents the indicator function:

[0027]

[0028] P PS (F) represents the total power consumption of the phase shifter:

[0029]

[0030] This indicates the power consumption of a single phase shifter when it is turned on.

[0031] P static This represents the static power consumption of components in the base station system other than the transmitting antenna, the radio frequency link, and the phase shifter.

[0032] In some embodiments, the minimum signal-to-interference-plus-noise ratio (SINNR) constraint of the information decoding receiver is:

[0033]

[0034] Γ ID This represents the minimum signal-to-interference-plus-noise ratio (SIR) of the information decoding receiver. γ represents the set of information decoding receivers. k (F,{w k},S) represents the signal-to-interference-plus-noise ratio (SIR) of the receiver decoding the k-th information:

[0035]

[0036] This represents the channel vector from the base station system to the k-th information decoding receiver. Indicates noise power, w i This represents the digital beamformer corresponding to the i-th information decoding receiver;

[0037] The maximum estimation Cramer-Rao bound constraint for the perceived target is:

[0038]

[0039] Γ S This represents the maximum estimate of the perceived target, Cramer-Rao bound. The mean Cramer-Rao bound representing the perceived target:

[0040]

[0041] M represents the Fisher information matrix of the perceived target;

[0042] The minimum harvesting power constraint for the energy harvesting receiver is:

[0043]

[0044] Γ EH This indicates the maximum harvesting power of the energy harvesting receiver. Ψ represents the set of energy harvesting receivers. j (F,{w k},S) represents the harvesting power of the j-th energy harvesting receiver:

[0045]

[0046] M j a is a constant representing the maximum harvested power of the j-th energy harvesting receiver under circuit saturation conditions. j and b j R represents a constant related to the circuit specifications of the j-th energy harvesting receiver. X This represents the covariance matrix of the baseband signal. This represents the channel vector from the base station system to the j-th energy harvesting receiver;

[0047] The transmit power constraint of the transmitting antenna is:

[0048]

[0049] This refers to the set of transmitting antennas;

[0050] The constant mode constraint of the simulated beamformer is:

[0051]

[0052]

[0053] This refers to the set of radio frequency links.

[0054] In some embodiments, optimizing the covariance matrix of the digital beamformer, the analog beamformer, and the sensed signal includes:

[0055] The covariance matrix of the digital beamformer, the analog beamformer, and the sensed signal is alternately optimized.

[0056] In some embodiments, the covariance matrix of the digital beamformer, the analog beamformer, and the sensed signal is alternately optimized, including:

[0057] Fix the analog beamformer and optimize the covariance matrix of the digital beamformer and the sensed signal;

[0058] The analog beamformer is optimized by fixing the covariance matrix of the digital beamformer and the sensed signal.

[0059] In some embodiments, fixing the analog beamformer and optimizing the covariance matrix of the digital beamformer and the sensed signal includes:

[0060] By fixing the simulated beamformer in the total power consumption model, a first intermediate problem model is obtained;

[0061] The first intermediate problem model is equivalently transformed using the semidefinite relaxation method to obtain the second intermediate problem model;

[0062] The second intermediate problem model is approximated by the continuous convex approximation method to obtain the third intermediate problem model;

[0063] The third intermediate problem model is solved using a convex optimization tool to obtain the optimized covariance matrix of the digital beamformer and the sensed signal.

[0064] In some embodiments, optimizing the covariance matrix of the digital beamformer, the analog beamformer, and the sensed signal includes:

[0065] The covariance matrix of the digital beamformer, the analog beamformer, and the sensed signal is initially optimized to obtain the initial optimization result;

[0066] Based on the initial optimization results, the beamforming weights of each of the transmitting antennas are determined. With minimizing the total power consumption of the base station system as the shutdown objective, the transmitting antennas are shut down iteratively in ascending order based on their beamforming weights, and the iterative optimization results are obtained.

[0067] To ensure the feasibility of optimizing the covariance matrix of the digital beamformer, the analog beamformer, and the sensing signal, the minimum number of radio frequency links is determined through binary search.

[0068] Based on the iterative optimization results, the beamforming weights of each phase shifter are determined. With the goal of minimizing the total power consumption of the base station system, the transmit antennas are shut down iteratively in ascending order based on their beamforming weights, and the final optimization result is obtained.

[0069] Thirdly, this invention provides a power consumption optimization device for a base station system, wherein the base station system is the integrated communication, sensing, and power transmission base station system described in the first aspect, and the power consumption optimization device includes:

[0070] The model building module is used to establish a total power consumption model and system constraints of the base station system regarding the digital beamformer, the analog beamformer, and the covariance matrix used to generate the sensing signal. The total power consumption model includes the total power consumption of the transmitting antenna, the total power consumption of the radio frequency link, and the total power consumption of the phase shifter.

[0071] The beamforming optimization module is used to optimize the covariance matrix of the digital beamformer, the analog beamformer, and the sensing signal based on the total power consumption model and the system constraints, with the goal of minimizing the total power consumption of the base station system.

[0072] In a fourth aspect, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method described in the first aspect above.

[0073] Fifthly, the present invention provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect above.

[0074] Compared with related technologies, this invention provides a base station system for multiple communication users, multiple sensing targets, and multiple energy harvesting receivers. This base station system includes multiple transmit links, multiple transmit antennas, and multiple phase shifters. Dynamic switching control of the transmit antennas, RF links, and phase shifters can be achieved using a switching network. Furthermore, by using different switch combinations, different baseband signals can be transmitted for different communication users, sensing targets, and energy harvesting receivers with the goal of minimizing the total power consumption of the base station. Simultaneously, constraints on the number of different devices in the base station ensure the feasibility of hybrid beamforming design. Therefore, this invention solves the problem of the lack of research on base station systems for multiple communication users, multiple sensing targets, and multiple energy harvesting receivers in related technologies.

[0075] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0076] Figure 1 This is a structural diagram of a base station system provided in an embodiment of the present invention;

[0077] Figure 2 This is a flowchart of a power consumption optimization method for a base station system provided in an embodiment of the present invention;

[0078] Figure 3 This indicates how the total power consumption of the base station system changes with respect to the signal-to-interference-plus-noise ratio (SINR) constraint.

[0079] Figure 4 This indicates how the total power consumption of the base station system changes with respect to the CRB constraint;

[0080] Figure 5 This indicates how the total power consumption of the base station system changes with respect to energy harvesting constraints;

[0081] Figure 6 This indicates the power distribution across different transmitting antennas;

[0082] Figure 7 This indicates how the total power consumption of the base station system changes with respect to the number of transmitted antennas and radio frequency links that are turned off;

[0083] Figure 8 This is a structural diagram of a power consumption optimization device for a base station system provided in an embodiment of the present invention. Detailed Implementation

[0084] To better understand the purpose, technical solution, and advantages of this application, the application is described and illustrated below in conjunction with the accompanying drawings and embodiments.

[0085] An embodiment of the present invention provides a base station system that integrates communication, sensing, and power transmission. Figure 1 This is a structural diagram of a base station system provided in an embodiment of the present invention, such as... Figure 1 As shown, the base station system includes:

[0086] The system comprises a hybrid architecture transmitter and sensing receiver. The hybrid architecture transmitter is used to simultaneously send communication signals to multiple information decoding receivers, sensing signals to multiple sensing targets, and energy signals to multiple energy harvesting receivers. The sensing receiver is used to receive echo signals from multiple sensing targets. The hybrid architecture transmitter includes a digital beamformer, multiple transmit links, an analog beamformer, and multiple transmit antennas. The analog beamformer and multiple transmit antennas include multiple phase shifters. The sensing receiver includes multiple receive antennas. The number of transmit links is greater than or equal to the number of information decoding receivers and less than or equal to the number of transmit antennas. The number of receive antennas is greater than the number of transmit antennas.

[0087] In this scheme, a single base station is K. ID Each information decoding receiver provides services, while also providing services to K. EH An energy harvesting receiver provides energy and senses K. S A perceived target. It can be defined. and These represent the sets of information decoding receivers, energy harvesting receivers, and sensing targets, respectively. The base station is equipped with a hybrid analog-to-digital transmitter, which contains N... T Root transmitting antenna and N RF One radio frequency link, and another with NR A fully digital architecture sensing receiver with a root receiving antenna. It can be defined. and These represent the sets of transmitting antennas and radio frequency links, respectively. A single transmitting antenna and radio frequency link subarray is selected to transmit independent, multi-functional data streams. At the base station end, for the information decoding receiver... The transmitted data stream first passes through a digital beamformer. Processed, and then passed through an analog beamformer This analog beamformer consists of multiple phase shifters with infinite resolution, and its mode length is constant: The elements of matrix F should satisfy in:

[0088]

[0089] When the corresponding phase shifter is active: When the corresponding phase shifter is turned off: [F] i,j =0.

[0090] This scheme introduces a novel hybrid architecture in the base station transmitter, enabling dynamic switching control of the transmit antenna, RF links, and phase shifters via a switching network. The information decoding receiver and the energy harvesting receiver are each equipped with a receiving antenna. To ensure the feasibility of the hybrid beamforming design, the number of data streams, RF links, and transmit antennas are constrained, i.e., K... ID ≤N RF ≤N T In addition, N is set in this scheme. T ≤N R This is to avoid the loss of perception due to a lack of spatial freedom.

[0091] In summary, this invention provides a base station system for multiple communication users (information decoding receivers), multiple sensing targets, and multiple energy harvesting receivers. This base station system includes multiple transmit links, multiple transmit antennas, and multiple phase shifters. Dynamic switching control of the transmit antennas, RF links, and phase shifters can be achieved using a switching network. Furthermore, by using different switch combinations, different baseband signals can be transmitted for different communication users, sensing targets, and energy harvesting receivers with the goal of minimizing the total power consumption of the base station. Simultaneously, constraints on the number of different devices in the base station ensure the feasibility of hybrid beamforming design. Therefore, this invention solves the problem of the lack of research on base station systems for multiple communication users, multiple sensing targets, and multiple energy harvesting receivers in related technologies.

[0092] The embodiments of the present invention also provide a power consumption optimization method for a base station system, wherein the base station system is the base station system provided in the embodiments of the present invention. Figure 2 This is a flowchart of a power consumption optimization method for a base station system provided in an embodiment of the present invention, referred to... Figure 2 The power consumption optimization method includes:

[0093] Step S210: Establish a total power consumption model and system constraints for the base station system regarding the digital beamformer, analog beamformer, and covariance matrix used to generate the sensing signal. The total power consumption model includes the total power consumption of the transmit antenna, the total power consumption of the radio frequency link, and the total power consumption of the phase shifter. Step S220: Based on the total power consumption model and system constraints, optimize the covariance matrix of the digital beamformer, analog beamformer, and sensing signal with the goal of minimizing the total power consumption of the base station system.

[0094] This solution essentially addresses the beamforming problem of base station systems, specifically how to shape the beam to reduce the overall power consumption of the base station system. Beamforming is related to the covariance matrix of the digital beamformer, analog beamformer, and sensed signal, thus requiring optimization. The total power consumption model is used to calculate the total power consumption of the base station system, which is related to variables such as the covariance matrix of the digital beamformer, analog beamformer, and sensed signal. System constraints are the conditions that these variables must satisfy.

[0095] In one embodiment, the system constraints include: a minimum signal-to-interference-plus-noise ratio constraint for the information decoding receiver, a maximum estimation Cramer-Rao bound constraint for the sensed target, a minimum harvesting power constraint for the energy harvesting receiver, a transmit power constraint for the transmit antenna, and a constant mode constraint for the analog beamformer.

[0096] The above scheme optimizes the total power consumption of base station systems with multiple communication users, multiple sensing targets, and multiple energy harvesting receivers. It addresses the lack of research on base station systems with multiple communication users, multiple sensing targets, and multiple energy harvesting receivers in related technologies, and further solves the problem of insufficient research on the energy consumption of multifunctional base station systems.

[0097] In some of these embodiments, the total power consumption model is as follows:

[0098]

[0099] in, The total power consumption of the base station system is represented by F, which represents the analog beamformer, and w is the total power consumption of the base station system. k Let S represent the digital beamformer corresponding to the k-th information decoding receiver, and let S represent the covariance matrix of the sensed signal.

[0100] P PA (F,{w k},S) represents the total power consumption of the transmitting antenna:

[0101]

[0102] N T η represents the number of transmitting antennas. n η represents the power amplifier efficiency of the nth transmitting antenna. max This represents the maximum power amplifier efficiency of the transmitting antenna. This represents the actual transmit power of the nth transmitting antenna. K represents the maximum transmit power of the nth transmitting antenna, β represents the efficiency factor, and K ID Indicates the number of information decoding receivers. This represents a zero matrix where all elements except the nth diagonal element are 0. Indicates a length of N T A vector whose elements are all zero except for the nth element which is 1;

[0103] P RF ({w k},S) represents the total power consumption of the RF link:

[0104]

[0105] N RF Indicates the number of radio frequency links. This indicates the power consumption of a single radio frequency link when it is enabled. Indicates a length of N RF A vector whose elements are all zeros except for the nth element which is 1. This represents a zero matrix where all elements except the nth diagonal element are 0, and I represents the indicator function:

[0106]

[0107] P PS (F) represents the total power consumption of the phase shifter:

[0108]

[0109] This indicates the power consumption of a single phase shifter when it is turned on.

[0110] P static This represents the static power consumption of components in a base station system other than the transmit antenna, RF link, and phase shifter.

[0111] The minimum signal-to-interference-plus-noise ratio (SINNR) constraint for the information decoding receiver is:

[0112] Γ ID This represents the minimum signal-to-interference-plus-noise ratio (SIR) of the information decoding receiver. γ represents the set of information decoding receivers. k(F,{w k},S) represents the signal-to-interference-plus-noise ratio (SIR) of the receiver decoding the k-th information:

[0113]

[0114] This represents the channel vector from the base station system to the k-th information decoding receiver. Indicates noise power, w i This represents the digital beamformer corresponding to the i-th information decoding receiver;

[0115] The maximum estimation Cramer-Rao bound constraint for the perceived target is:

[0116] Γ S This represents the maximum estimate of the perceived target, Cramer-Rao bound. The mean Cramer-Rao bound representing the perceived target:

[0117]

[0118] M represents the Fisher information matrix of the perceived target;

[0119] The minimum harvesting power constraint for the energy harvesting receiver is:

[0120]

[0121] Γ EH This indicates the maximum harvesting power of the energy harvesting receiver. Ψ represents the set of energy harvesting receivers. j (F,{w k},S) represents the harvesting power of the j-th energy harvesting receiver:

[0122]

[0123] M j a is a constant representing the maximum harvested power of the j-th energy harvesting receiver under circuit saturation conditions. j and b j R represents a constant related to the circuit specifications of the j-th energy harvesting receiver. X Represents the covariance matrix of the baseband signal. This represents the channel vector from the base station system to the j-th energy harvesting receiver;

[0124] The transmit power constraint of the transmitting antenna is:

[0125]

[0126] Represents a collection of transmitting antennas;

[0127] The constant mode constraint of the simulated beamformer is:

[0128]

[0129]

[0130] This represents a set of radio frequency links.

[0131] The following is a detailed explanation of the construction process and principles of the total power consumption model and various system constraints in the above embodiments.

[0132] The base station sends communication signals to the target information decoding receiver k. in It is a CSCG (circularly symmetric complex Gaussian) random variable with zero mean and unit variance, i.e. in Indicates symbol index. Besides communication signals {s k In addition to (l), the base station also sends a dedicated sensing / energy signal, which is transmitted via vector. This is indicated to provide sufficient degrees of freedom (DoF) for target perception, where s0(l) is independent of {s}. k (l)} is a matrix with zero mean and covariance matrix of... The CSCG random vector, i.e. {s k (l)} and s0(l) are independent on different signs, for all S is a matrix with a general rank, i.e., 0 ≤ m = rank(S) ≤ N. RF This corresponds to the case with m linearly and statistically independent sensing signal beams, each of which can be obtained through eigenvalue decomposition (EVD) of S. Therefore, the baseband signal transmitted by the base station... Given by the following formula:

[0133]

[0134] Where F and {w k} represent the analog beamformer and digital beamformer to be optimized, respectively. Therefore, the covariance matrix of x(l) is expressed as:

[0135]

[0136] This embodiment considers a quasi-static narrowband channel model in which the wireless channel remains constant during a transport block of interest consisting of L symbols, where L is assumed to be sufficiently large. Let This represents the set of symbols within this specific coherent processing interval. This represents a narrowband multi-functional transmit signal matrix. Let... and Let represent the channel vectors from the base station to the information decoding receiver k and the energy harvesting receiver j, respectively. This represents the target response matrix from the hybrid architecture transmitter to the sensing target i and back to the sensing receiver. The base station has... and All Channel State Information (CSI).

[0137] First, consider a downlink multi-user communication system. This is achieved by transmitting x(l) to K... ID There is an information decoding receiver. The signal received at point k by the information decoding receiver is represented as follows:

[0138]

[0139] in, The AWGN (Additive White Gaussian Noise) at point k in the information decoding receiver is represented. This represents the corresponding noise power. Therefore, the signal-to-interference-plus-noise ratio (SINR) at receiver k in the information decoding unit is:

[0140]

[0141] SINR was selected as the communication performance metric.

[0142] Secondly, consider radar sensing of multiple targets. It is important to note that both the sensing signal s0(l) and the communication signal s0(l) can be utilized to achieve the sensing function, as the sensing receiver possesses all the information about them. Therefore, by transmitting x(l) to the sensing target, the echo signal received by the sensing receiver of the multi-functional base station... Represented as:

[0143]

[0144] in, This represents AWGN (Additive White Gaussian Noise) at the sensing receiver. This represents the corresponding noise power. Considering a monostation radar setup, the arrival direction angle (DoA) and departure direction angle (DoD) are the same. A point model is used for the sensed target. Target response matrix G i Represented as:

[0145] G i =a(θ) i )β i v T (θ i )

[0146] in, It is the complex amplitude of the sensed target, including the round-trip path loss and the target's radar cross-section (RCS). K is relative to the base station S The azimuth angle of a perceived target. and These are the steering vectors of the transmitting and receiving antennas, respectively. To simplify the notation, the echo signal received by the sensing receiver can be expressed as:

[0147] in:

[0148] diag(b) represents a diagonal matrix, and the vector formed by its diagonal elements is b.

[0149] Choosing the center of a uniform linear array (ULA) receiver antenna as the reference point, and assuming that adjacent receiver antennas are spaced half a wavelength apart and the number of receiver antennas is even, we can then obtain:

[0150]

[0151]

[0152] Its derivative is:

[0153]

[0154]

[0155] Among them, v i and a i v(θ) i ) and a(θ i The i-th element of ). It should be noted that when turning off the antenna, since the physical position of the antenna element cannot be moved, the antenna element that has been turned off needs to be deleted from the above-mentioned guiding vector.

[0156] Now, consider the Cramer-Rao bound (CRB) for the unknown objective parameters θ and b, which represents the lower bound of the variance of any unbiased estimate. Let the objective parameters be θ and b. R and b I ,in The received signal y is known S (n), then regarding θ, b R and b I The Fisher Information Matrix (FIM) can be derived as follows:

[0157]

[0158] in:

[0159]

[0160]

[0161]

[0162]

[0163]

[0164] Therefore, the corresponding CRB matrix is ​​derived from CRB(F,{w k},S)=M -1 Provided.

[0165] The FIM of the target parameters is the covariance matrix R of the multi-function beam X. X A linear function. In this embodiment, the weighted trace of the CRB matrix is ​​used, i.e. It serves as a performance metric for perception because it represents the weighted average CRB of all estimated parameters in FIM.

[0166] Furthermore, consider wireless power transfer (WPT) from the base station to the energy harvesting receiver, where the received radio frequency signal is converted into a direct current (DC) signal by a rectifier for energy harvesting. Energy harvesting receiver The received radio frequency power (energy per unit time, measured in watts) can be expressed as:

[0167]

[0168] in:

[0169]

[0170]

[0171] Ω j This is an introduced constant to ensure that when the input RF power is zero, the energy collected at the output is also zero. M j Ψ is a constant representing the maximum harvested power of the energy harvesting receiver when the EH circuit is in saturation. j (F,{w k},S) is about the received radio frequency power The traditional Logistic function. Parameter a j and b j It is a constant related to circuit specifications (e.g., resistance, capacitance, and diode forward voltage). In practice, the EH hardware circuitry for each energy harvesting receiver is fixed, and the parameter 'a' in the proposed model... j b j and M jThis can be determined using a standard curve fitting tool. We note that the proposed nonlinear energy harvesting model can capture the combined effects of nonlinear phenomena caused by hardware constraints, including circuit sensitivity limitations and current leakage. Due to Ω j Without affecting beamforming design, for simplicity, Ψ can be used directly. j (F,{w k Let},S) represent the power collected at the energy harvesting receiver j.

[0172] Further considering the actual power consumption model of the base station, it consists of two parts: a hybrid architecture transmitter and a fully digital architecture sensing receiver. The transmitter's power consumption includes the nonlinear power consumption of power amplifiers (PAs), the switching power consumption of the RF link and phase shifters, and the static power consumption of the switching network and other components, which are described in detail below. In communication and sensing integration or multi-functional scenarios, there is relatively little research considering the power consumption introduced by sensing-related devices. Generally, the power consumption of the sensing receiver includes the power consumption of the low-noise amplifier (LNA) and the static power consumption of other signal processing components. In our design, we adopt this sensing receiver model.

[0173] 1. Power Consumption of a Nonlinear Efficiency Power Amplifier (PA): In practice, the efficiency of the power amplifier (PA) for each receiving antenna is nonlinear, and most classical models simplify this by treating the PA efficiency as a fixed value, which is ideal but impractical. Here, we consider a nonlinear efficiency PA power consumption model, where the PA efficiency is a nonlinear function of the transmit power on the receiving antenna, i.e.:

[0174]

[0175] in, and These are the maximum transmitted signal power and the actual transmitted signal power on the receiving antenna n, respectively, η max β represents the maximum PA efficiency, and β is the efficiency factor, which depends on the type of PA. For example, for a Class A PA, β = 1, and for a Class B PA, β = 0.5. Based on the signal model above, we can conclude that:

[0176]

[0177] in, It is a vector whose nth element is 1 and the rest are 0, and its dimension will be determined from the context. It is defined as a zero matrix, except that the nth diagonal element is 1. Therefore, all N T The power consumption of a PA is given by the following formula:

[0178]

[0179] Further considering the constraint of the base station's transmit power on the transmission power of a single antenna, that is, the maximum transmit power of each antenna of the base station. Limitations:

[0180]

[0181] 2. Power consumption of RF links with switch control: In practice, the power consumption of a base station's RF links depends on the switch state of each RF link. Specifically, if RF link m is in the on state, the power transmitted through it... The corresponding RF link m will consume a fixed amount of power, as given by the following formula:

[0182]

[0183] Among them, P DAC P mix P filt and P syn These represent the power consumption of the digital-to-analog converter (DAC), mixer, filter, and frequency synthesizer, respectively. Otherwise, if The corresponding RF link m is shut down, and power consumption is zero. This is achieved by combining N... RF The total power consumption of each RF link is given by the following formula:

[0184]

[0185] Where I{·} represents an indicator function, defined as follows:

[0186]

[0187] 3. Power consumption of phase shifters with switch control: In practice, an active phase shifter consumes a constant amount of power. An inactive phase shifter consumes no power. Considering the dynamic switching control of the phase shifter network, the total power consumption of the phase shifter can be expressed as:

[0188]

[0189] The on / off state of each element is represented by an indicator function.

[0190] 4. Static power consumption of other components: In addition to power amplifiers, RF links and phase shifters, other components of the base station, including switching networks, all-digital architecture sensing receivers and other supporting equipment, also consume energy. Their power consumption is usually static and can be specified separately as follows.

[0191] Compared to a phase shifter, the power consumption of a single switch is very small and can be expressed as a static term. At this time, the power consumed by the main switching network is:

[0192]

[0193] The above represents the total power consumed by the switching of the entire antenna, RF link, and phase shifter system.

[0194] In sensing receivers, LNA design is one of the most challenging aspects due to the requirement for a low noise figure. The DC power absorbed by the LNA... The calculation is as follows:

[0195]

[0196] Where G is the gain, and FoM is in mW -1 Let be the quality factor in units, and NF be the noise figure of the LNA. Thus, the total power consumption of the LNA can be expressed as:

[0197]

[0198] When the system hardware configuration is fixed, the above power consumption is a static item.

[0199] In addition, other components in the sensing receiver, including the mixer (LO), A / D converter, and digital beamformer, also consume power. The power consumption of these components is a function of the sampling frequency and is therefore considered a static term P when the system hardware configuration is fixed. RS In addition, other components in the base station, such as remote backhaul, cooling system, baseband processing, and power supply, also consume static power P. c .

[0200] By combining the power consumption models proposed above, the total power consumption of the base station on the transmission block is given by the following formula:

[0201]

[0202] Among them, P static =P SW +P LNA +P c +P RS This represents the total static power consumption.

[0203] Based on the signal and power consumption models described above, the hybrid beamforming problem can be modeled. This problem aims to minimize the total power consumption while ensuring that each information decoding receiver... Minimum signal-to-interference-plus-noise ratio requirement Γ ID Multiple sensing targets Maximum estimation CRB constraint Γ S and each energy harvesting receiver Minimum collected power constraint Γ EH In addition, it includes transmit power constraints for each receiving antenna and constant mode constraints for simulated beamforming. The hybrid beamforming problem model is as follows:

[0204]

[0205]

[0206]

[0207]

[0208]

[0209]

[0210] in,

[0211] As shown above, the total power consumption model and the constraints of each system are explained in detail. Together, they constitute the hybrid beamforming problem model.

[0212] In some embodiments, step S220, optimizing the covariance matrix of the digital beamformer, analog beamformer, and sensing signal, includes step S221, alternately optimizing the covariance matrix of the digital beamformer, analog beamformer, and sensing signal.

[0213] Since there are multiple variables to be optimized, an alternating optimization method is used to facilitate the optimization of these variables. Alternating minimization represents a widely applicable and successful method in practice for optimization problems involving multivariate coupling.

[0214] Specifically, the alternating optimization process can be as follows: Step S221a, fix the analog beamformer and optimize the covariance matrix of the digital beamformer and the sensing signal; Step S221b, fix the covariance matrix of the digital beamformer and the sensing signal and optimize the analog beamformer.

[0215] Specifically, step S221a includes:

[0216] The simulated beamformer in the fixed total power consumption model is used to obtain the first intermediate problem model; the first intermediate problem model is equivalently transformed using a positive semidefinite relaxation method to obtain the second intermediate problem model; the second intermediate problem model is approximated using a continuous convex approximation method to obtain the third intermediate problem model; the third intermediate problem model is solved using a convex optimization tool to obtain the optimized digital beamformer and the covariance matrix of the sensing signal. The above scheme provides a specific optimization process: first, the non-optimized quantities in the model are fixed; then, a positive semidefinite relaxation method is used to process the model; next, a continuous convex approximation method is used; finally, a convex optimization tool is used to solve the model. Correspondingly, step S221b can also be optimized using the above process, which includes:

[0217] By fixing the covariance matrix of the digital beamformer and the sensing signal in the total power consumption model, a fourth intermediate problem model is obtained. The fourth intermediate problem model is equivalently transformed by the positive semidefinite relaxation method to obtain a fifth intermediate problem model. The fifth intermediate problem model is approximated by the continuous convex approximation method to obtain a sixth intermediate problem model. The sixth intermediate problem model is solved by the convex optimization tool to obtain the optimized analog beamformer.

[0218] The optimization process described above will be explained in detail below with reference to the specific hybrid beamforming problem model given in the above embodiments.

[0219] In this embodiment, due to the constant modulus constraint of F, the nonlinearity introduced by the actual model, and the complex coupling between variables, F and {w} are jointly optimized. k Solving the variables {F, S} and {w} is very difficult. By decoupling the optimization of these three variables, alternating minimization becomes an efficient method to obtain a valid solution. Following the principle of alternating minimization, we will alternately solve for the variable pairs {F, S} and {w}. k}, while fixing another variable.

[0220] First, consider optimizing the digital beamforming matrix {w} k Let S be the perceived covariance matrix, and assume that the simulated beamforming matrix F is fixed. Therefore, the hybrid beamforming problem model can be further simplified to the first intermediate problem model form, where the static term P in the total power model is further temporarily ignored. static And constant mode constraints on the simulated beamforming matrix F. The first intermediate problem model is as follows:

[0221]

[0222] Although the problem model described above is non-convex, it can be relaxed into a convex problem using the classic SDR (semi-definite relaxation) method. Let... in And rank(R)k ) = 1. Therefore, we can obtain:

[0223]

[0224] Next, consider transforming the constraint into a constraint about {R}. k The convex form of} and S.

[0225] First, considering the signal-to-interference-plus-noise ratio (SINR) constraint of the communication signal, let The original constraint can then be equivalently rewritten as:

[0226]

[0227] The above constraints are affine, and therefore with respect to {R} k} and S are convex.

[0228] Secondly, consider the trace of the weighted CRB matrix as a measure of perceived performance, i.e.:

[0229]

[0230] Using the above approach, the perception performance constraint can be rewritten as follows:

[0231]

[0232] And introduce additional constraints:

[0233] in, It is the identity matrix The i-th column, {t i} is the introduced slack variable. Then, by applying Schur's complement, it can be restated as the following inequality:

[0234]

[0235] This form is about {R} k} and S-convex.

[0236] Finally, consider the power constraints for wireless energy transfer harvesting with a practically parameterized nonlinear energy harvesting model, namely:

[0237]

[0238] It can be equivalently transformed into the following form:

[0239]

[0240] This is a linear matrix inequality concerning the received RF power on the left, therefore concerning {R} k} and S are convex.

[0241] The power amplifier and RF link power consumption terms in the first intermediate problem model can also be expressed as {R} k The functional forms of} and S:

[0242]

[0243]

[0244] Next, we can address the indicator function related to RF link power consumption in the first intermediate problem model. This function is binary and therefore non-convex. For this, we equivalently rewrite the indicator function as:

[0245]

[0246] Therefore, the indicator function can be approximated under a properly chosen ε as:

[0247] As ε becomes smaller, the approximation becomes more accurate. Therefore, the total power consumption of the RF link can be approximated as:

[0248]

[0249] Through the above derivation, the first intermediate problem model and all constraints are expressed as about {R}. k The form of the function of} and S. Therefore, when the rank-one constraint rank(R) is relaxed... k When ) = 1, the first intermediate problem model can be approximated as the second intermediate problem model:

[0250]

[0251] It should be noted that in the second intermediate problem model, the variable that needs to be optimized is {R}. k} and S, instead of {w} in the first intermediate problem model k} and S.

[0252] The second intermediate problem model can be solved using the SCA (Continuous Convex Approximation) method. In SCA, the complex problem is approximated as a series of easily solvable convex problems, approximated iteratively. In one embodiment, consider a specific iteration number j ≥ 1. Let... and S (j) In the j-th iteration of SCA, {R} k Let S be a local point. The first-order Taylor expansion of any concave function is its global upper bound. Therefore, given a local point... and S (j) In this case, the PA power consumption term P PA ({R k},S) can be bounded by its upper bound. Approximately, F is known and determined, where:

[0253]

[0254] Similarly, RF link power consumption It can also be further approximated as its upper bound. in:

[0255]

[0256] By P PA ({R k},S) and Replace with and The second intermediate problem model can be transformed into the third intermediate problem model in the j-th iteration of SCA:

[0257]

[0258] The third intermediate problem model is a convex problem, solvable by standard convex optimization tools, such as CVX. Let... and S *(j) Let represent the solutions to the third intermediate problem model obtained in the j-th iteration, and let represent the solutions obtained in the j-th iteration. These solutions will be updated for the next iteration. and S (j+1) .because and They are P PA ({R k},S) and The global upper bound is such that it can be proven that... and S (j+1) The total power consumption obtained must not exceed and S (j) In other words, the power consumption value obtained by solving the second intermediate problem model is monotonically non-increasing during the SCA iteration process. Therefore, the optimal objective function value of the second intermediate problem model is bounded. This guarantees the convergence of the problem. Let the convergence results of communication and sensing covariance be respectively... and in It is usually not of the rank one.

[0259] Based on the obtained solution and The rank-one solution to the first intermediate problem model can be reconstructed:

[0260]

[0261]

[0262] The first equation guarantees the minimum signal-to-interference-plus-noise ratio constraint, while the second equation guarantees other quadratic constraints and also ensures that the model value of the first intermediate problem remains unchanged.

[0263] In the alternating optimization process, the next step is to fix {w}. k Given S, calculate the optimal solution F for the following fourth intermediate problem model:

[0264]

[0265] Similar to the solution process described above, the fourth intermediate problem model can also be relaxed to a convex form using the classic SDR method. Let f = vec(F T ) represents the matrix F T Perform column vectorization. Let Rf = ff H ,in And rank(R) f = 1. Considering the constant modulus constraint and the phase shifter selection, the following relaxed form of constraint can be obtained:

[0266]

[0267] because, in Defined as the following matrix:

[0268]

[0269] in, It is a completely one vector. Therefore, it can be defined as:

[0270]

[0271] It's about R f The linear function. Let r = rank(S), and consider the eigenvalue decomposition of the Hermitian positive semidefinite matrix S. Where λ1≥λ2≥...≥λ r ≥0 are eigenvalues, q1,...,q r These are the corresponding eigenvectors. Therefore, we can define:

[0272]

[0273] R S It's about R f R is a linear function. X yes and R S A linear function of R, therefore also R f A linear function, where:

[0274]

[0275] Since both sensing and energy transfer constraints are R X Linear functions of R, therefore they are also linear functions of R. f A linear function. The minimum signal-to-interference-plus-noise ratio constraint can be reexpressed as follows:

[0276]

[0277] It is about and R S It is affine, therefore convex. The transmit power constraint on a single antenna can be written in the following affine form:

[0278]

[0279] The power consumption terms for the power amplifier and phase shifter in the fourth intermediate problem model can be re-expressed as R f Functions:

[0280]

[0281]

[0282] The power consumption of the phase shifter can be approximated by using an indication function as a function of R. f Concave form:

[0283]

[0284] Based on the above derivation, the fourth intermediate problem model and all constraints are re-expressed as R. f The function. Therefore, when the rank-one constraint is relaxed, rank(R) f When ) = 1, the fourth intermediate problem model can be approximated as the fifth intermediate problem model:

[0285]

[0286] The optimization variable in the fifth intermediate problem model is R. f Rather than F in the fourth intermediate problem model.

[0287] At this point, the SCA method can be used to handle the fifth intermediate problem model. Consider a specific iteration number j≥1. Let R... f (j) This is a local point in the j-th iteration of SCA. The first-order Taylor expansion of any concave function is its global upper bound. Therefore, given a local point R... f (j) In this case, the PA power consumption term P PA (R fIt can be bounded by its upper bound. Approximate. Where:

[0288]

[0289] Similarly, the approximate phase shifter power consumption term It can be bounded to its upper boundary Further approximation. Where:

[0290]

[0291] By P PA (R f )and Replace with and The fifth intermediate problem model can be transformed into the sixth intermediate problem model in the j-th iteration of SCA:

[0292]

[0293] The sixth intermediate problem model is a convex problem and can be solved by standard convex optimization tools, such as CVX. Let R... f *(j) This represents the solution to the sixth intermediate problem model obtained in the j-th iteration, which will be updated to R in the next iteration. f (j+1) .because and They are P PA (R f )and The global upper bound of R is such that it can be proven that R... f (j+1) The total power consumption obtained will definitely not exceed R. f (j) In other words, the power consumption value obtained by solving the fifth intermediate problem model is monotonically non-increasing during the SCA iteration process. Therefore, the optimal objective function value of the fifth intermediate problem model is bounded. This guarantees the convergence of the problem. Let the convergence result of the simulated beamforming matrix covariance be denoted as... It is not usually rank-one.

[0294] Finally, the rank-one solution can be recovered using Gaussian randomization. Specifically, a set of randomized solutions can be generated. And based on this, a set of candidate feasible solutions is generated:

[0295] We can choose the f that minimizes the value of the fourth intermediate problem model. * As an optimization result, Gaussian randomization should be performed sufficiently many times to ensure that the value of the fourth intermediate problem model decreases in each iteration of alternating optimization. Let f... *Representing it in matrix form, we obtain the optimized result F of the simulated beamformer. * .

[0296] The above describes a specific optimization process. It should be noted that this optimization method can also be combined with intelligent shutdown of the receiving antenna, RF link, and phase shifter.

[0297] In some embodiments, step S220, optimizing the covariance matrix of the digital beamformer, the analog beamformer, and the sensed signal, includes:

[0298] Step S222: Initial optimization is performed on the covariance matrix of the digital beamformer, analog beamformer, and sensing signal to obtain initial optimization results. Step S223: Based on the initial optimization results, the beamforming weights of each transmit antenna are determined. With minimizing the total power consumption of the base station system as the shutdown objective, the transmit antennas are shut down iteratively in ascending order based on the beamforming weights of each transmit antenna, and the iterative optimization results are obtained. Step S224: On the premise of ensuring the feasibility of optimizing the covariance matrix of the digital beamformer, analog beamformer, and sensing signal, the minimum number of RF links is determined through binary search. Step S225: Based on the iterative optimization results, the beamforming weights of each phase shifter are determined. With minimizing the total power consumption of the base station system as the shutdown objective, the transmit antennas are shut down iteratively in ascending order based on the beamforming weights of each transmit antenna, and the final optimization results are obtained.

[0299] Specifically, this embodiment utilizes the F obtained by solving the hybrid beamforming problem. * , and S * Establish a discrimination criterion. In practical systems, power amplifiers typically consume the vast majority of power, accounting for approximately 50%-80% of the total system power consumption. The power consumption of a single phase shifter is much lower than that of the RF link. At 2.4 GHz, the power consumption of a single phase shifter and switch is generally considered to be within the following range: The power consumption of a single RF link is generally considered to be Therefore, this embodiment employs the following approach: components with higher power consumption should have a higher shutdown priority, thus considering the shutdown of the transmit antenna, RF links, and phase shifters in that order. Specifically, this embodiment first shuts down as many transmit antennas as possible based on the transmit antenna shutdown weight, then performs a binary search to determine the minimum number of RF links that ensures the feasibility of the problem, and finally considers the shutdown of the phase shifters to further reduce system power consumption.

[0300] First, the optimization method described in the above embodiments can be used to initially optimize the covariance matrix of the digital beamformer, analog beamformer, and sensing signal, and the beamforming weights of each transmit antenna can be determined based on the initial optimization results. The beamforming weight of transmit antenna n is defined as follows:

[0301]

[0302] This represents the transmit power on that transmitting antenna. Clearly, the beamforming weight v n Smaller transmit antennas have less impact on beamforming design performance and are therefore prioritized for shutdown. Thus, we sort the beamforming weights of different transmit antennas in ascending order and iteratively determine the on / off state of the transmit antennas. Specifically, we iteratively shut down antennas according to the above sorting and compare their power consumption values ​​achieved through the optimization method described in the above embodiments to obtain the optimal antenna switch selection design that minimizes total power consumption. For example, we can first shut down one or more transmit antennas with the lowest weights, then re-optimize the total power consumption model and determine if the total power consumption decreases. If it decreases, we continue to shut down one or more transmit antennas with the lowest weights and re-optimize the total power consumption model. This iterative process of shutting down transmit antennas and optimizing the total power consumption model continues until the total power consumption of the base station no longer decreases, ultimately determining the optimal shutdown scheme for the transmit antennas.

[0303] After the transmit antennas are turned off, the selection of radio frequency (RF) links is considered. Since each RF link is connected to all transmit antennas, shutting down any one RF link has the same impact on the transmitter system. Therefore, by performing a binary search on the minimum number of RF links that guarantees the feasibility of the total power consumption model (the existence of an optimal solution), the optimal RF link selection can be obtained with low complexity. Specifically, a binary search can be performed starting from the smaller of the total number of RF links configured in the base station and the optimal number of antennas selected from the transmit antennas. Based on the current number of RF links, the feasibility of the total power consumption model is determined. If feasible, the search moves towards fewer RF links; if not, the search moves towards more RF links.

[0304] Finally, the beamforming weights of the phase shifter are determined based on the optimization results, and the phase shifter's turn-off is considered based on these weights. The optimal f is obtained through Gaussian randomization in the final alternating optimization iteration. * and corresponding we will Represented in the form of the corresponding analog beamforming matrix And based on this, define each phase shifter [F] * ] i,j Beamforming weights Where F * It is the optimal simulation domain matrix obtained in practice, taking into account the constant modulus constraint, and This is its form when the constant mode constraint is ignored. Clearly, the beamforming weight c... i,jSmaller phase shifters have a smaller impact on beamforming design performance and are therefore prioritized for shutdown. Thus, we sort all phase shifters by their beamforming weights in ascending order and iteratively determine their on / off states. Specifically, we iteratively shut down phase shifters according to the above sorting, turning off 10% of them at a time, and compare their power consumption values ​​achieved by the previously proposed optimal beamforming algorithm to obtain the optimal phase shifter selection design that minimizes total power consumption. Turning off phase shifters is represented by forcing a subset of elements to zero; we consider adding an additional constraint [R] on the nth phase shifter that is turned off in the analog domain beamforming solution. f ] n,n =0.

[0305] As can be seen from the above embodiments, the base station system and its power consumption optimization method provided by the present invention have the following characteristics and advantages:

[0306] 1. The technical solution of this invention considers the simultaneous integration of communication, sensing, and energy transmission capabilities within a single station system. Compared with existing technical solutions, this solution effectively reduces hardware overhead and improves energy efficiency.

[0307] 2. This invention proposes a system energy consumption optimization method based on joint energy optimization hybrid beamforming and dynamic antenna, radio frequency, and phase shifter network shutdown control. This method minimizes single-site energy consumption while ensuring performance requirements, fully utilizes system resources, and improves energy efficiency.

[0308] 3. This invention proposes an algorithm based on Alternating Optimization (AO), Sustained Convex Approximation (SCA), and SDR to solve the energy-optimized hybrid beamforming problem. Alternating Optimization (AO) decouples multiple variables, Sustained Convex Approximation (SCA) approximates the non-convex power consumption objective function into a convex function through a first-order Taylor expansion, and Semi-definite Relaxation (SDR) transforms quadratic constraints into multivariable linear constraints. Compared to existing algorithms, this scheme achieves more accurate and energy-efficient beamforming.

[0309] 4. This invention proposes a beamforming weighted transmit antenna shutdown control algorithm and a binary search-based RF link shutdown lookup algorithm, respectively used to shut down the transmit antenna and the RF link. Furthermore, it proposes a phase shifter shutdown scheme that guarantees constant modulus constraints in the analog precoding domain. Compared to existing algorithms, this scheme solves this complex combinatorial optimization problem with lower complexity.

[0310] 5. Based on the above algorithm, this invention proposes an energy consumption trade-off for multiple performance indicators in a communication-sensing-power transmission integrated system, and analyzes the impact of energy-optimized hybrid beamforming gain and dynamic shutdown control gain on the total energy consumption of the system.

[0311] Five simulation examples are provided below to verify the feasibility of the technical solution in this invention.

[0312] In the simulation, the base station is equipped with an N T =32 and N S =32 antennas in a uniform linear array, with the spacing between adjacent antennas being half a wavelength, and having N RF =32 radio frequency links. The number of communication receivers, sensing targets, and energy harvesting receivers are each set to K. ID =8,K S =6 and K EH =6. Set the noise power of the information decoding receiver and the sensing receiver to... Regarding power consumption parameters, set P c =10W, η max =0.38 and β=0.5. The signal-to-interference-plus-noise ratio (SIR) constraints for different users and the minimum harvesting power requirements for different energy harvesting receivers were all set to be the same. Next, a Rayleigh fading channel model was considered for the information decoding receiver, a line-of-sight channel model for the sensing target, and a Rician fading channel model with a Rician factor of 3dB for the energy harvesting receiver. The large-scale path loss model was modeled as 51.2 + 41.2log 10 r, where the distance between the base station transmitter and the information decoding receiver is r = 50m, and the distance between the base station transmitter and the energy harvesting receiver is r = 20m. The perceived line-of-sight channel model is set to 51.2 + 41.2log 10 2r S The distance r between the base station transmitter and the sensing target S =10m, which takes into account the round-trip path loss (r and r) S The unit is meters. Target azimuth. It is randomly generated. The transport block length is set to L = 30. For general considerations, the weight matrix Λ of the CRB is... α It is set as the identity matrix. In the nonlinear energy harvesting model, M is set as... j =0.02, a j =6400, b j =0.003, these data are obtained by curve fitting of measurement data from specific hardware devices. For the sensing receiver, G=10dB, FoM=8.46 and NF=3.1dB are set, these are commonly used values ​​for the receiver.

[0313] To make performance comparisons, we consider the following benchmark schemes: a conventional power amplifier model with fixed power amplifier efficiency, a conventional energy-optimized hybrid beamforming design based on optimizing total transmit power, and a scheme that ignores the dynamic switching selection of antennas, RF links, and / or phase shifters. These are compared with our proposed joint energy-optimized hybrid beamforming and dynamic antenna, RF, and phase shifter network shutdown control design.

[0314] 1. Hybrid Beamforming Design with Optimized Total Transmit Power: Considering traditional energy optimization design, this approach minimizes total transmit power instead of total power consumption, while ensuring the same constraints are met: communication / sensing / wireless power transfer performance constraints, transmit power constraints for each antenna, and constant mode constraints for simulated beamforming. The optimization problem can be expressed in the following form:

[0315]

[0316]

[0317]

[0318]

[0319]

[0320]

[0321] To solve this problem, we can consider the objective function P. T An equivalent expression Based on this form, the above problem can be solved similarly using the algorithm proposed earlier. SCA is no longer needed in this case.

[0322] 2. Hybrid Beamforming Design Only: This corresponds to the case where dynamic selection is completely ignored, and all antennas, RF links, and phase shifters are set to active state. The hybrid beamformer can be solved using the proposed algorithm.

[0323] 3. Phase Shifter Selection Only Design: This corresponds to the case where the dynamic selection of the antenna and RF link is ignored, and the focus is on the phase shifter's shutdown control. In this case, we observe the energy efficiency gain brought by the phase shifter selection alone and explore the impact of the phase shifter's switching selection on the total system power consumption.

[0324] 4. RF Link Selection Only Design: This corresponds to the case where the dynamic selection of antennas and phase shifters is ignored, and the focus is on the shutdown control of the RF link. In this case, we observe the energy efficiency gain brought by the RF link selection alone and explore the impact of the RF link switching selection on the total system power consumption. Here, each phase shifter is normalized to a unit mode after beamforming, and shutdown control is not considered.

[0325] 5. Joint All-Digital Beamforming and Dynamic Selection Design: This corresponds to implementing the proposed joint design scheme in an all-digital transmitter architecture. In this case, the analog beamformer F and the digital beamformer {w} k} are combined into a fully digital beamformer that can be arbitrarily controlled {d k The alternating minimization step is no longer needed, and the rest of the proposed algorithm is similarly relocated.

[0326] 6. Joint Design Using a Fixed Power Amplifier Efficiency Model: This corresponds to setting β = 0 in the proposed joint design. In this case, the power consumption model of the power amplifier can be simplified to the following traditional form, which has been extensively studied:

[0327]

[0328] The above model can be solved using the optimization method provided by this invention.

[0329] Reference Figure 3 , Figure 3 This shows that when a given CRB threshold Γ is displayed S =0.1 and energy harvesting constraint Γ EH The figure shows the change in total system power consumption with respect to the signal-to-interference-plus-noise ratio (SINR) constraint at -2 dBm. Several benchmark schemes and the proposed design are presented under the same channel conditions. It can be seen that the total system power consumption achieved by the proposed joint design is significantly lower than other benchmark schemes. Furthermore, we note that when the SINR requirement is relatively small, dynamic gain selection has a more significant effect on reducing total system power consumption compared to optimal beamforming gain, while the opposite is true under relatively large SINR constraints. This is because when the SINR constraint is high, more antennas and RF links must be activated to ensure performance, leading to a reduction in dynamic gain selection. To reveal the trade-offs between SINR, CRB, and energy harvesting constraints, we also find that when the SINR requirement Γ... ID Below a certain inflection point, the total power consumption of the system will be limited by the CRB and energy harvesting constraints, which will cause the total power consumption of the system to no longer decrease with the decrease of the signal-to-interference-plus-noise ratio constraint, and will show horizontal behavior in the figure.

[0330] Figure 4 and Figure 5 The figures show the values ​​for each SINR requirement Γ. ID =6dB and energy harvesting constraint Γ EH The total system power consumption changes with the CRB constraint when the SINR requirement is -2dBm, and the total system power consumption changes with the CRB constraint when the SINR requirement is given. ID =6dB and CRB threshold Γ SThe total system power consumption varies with the energy harvesting constraint when the CRB threshold is 0.1. The proposed design significantly outperforms any benchmark scheme in terms of energy efficiency. Furthermore, we note that when the CRB threshold is high or the energy harvesting constraint is low, the total power consumption is limited by the other two constraints, resulting in the horizontal behavior shown in the figure. Additionally, we can observe that the inflection point for optimizing the total transmit power occurs earlier in the design. This indicates that the energy efficiency achieved by this beamforming design is reduced compared to the proposed design.

[0331] Figure 6 We compared the optimal transmit power allocation on different antennas under two power amplifier power consumption models: one considering nonlinear power amplifier efficiency (i.e., β = 0.5) and the other considering fixed power amplifier efficiency (i.e., β = 0). In the traditional model considering fixed power amplifier efficiency, there is a tendency to activate almost all antennas and distribute power relatively evenly across each antenna. In contrast, our proposed design considering nonlinear power amplifier efficiency reveals a new power allocation pattern: some antennas are turned off, while others receive more power to meet performance requirements. This is because the efficiency of nonlinear power amplifiers increases with the output signal power at each antenna. Therefore, base stations tend to turn off more antennas and increase the transmit power on the remaining antennas to utilize higher power amplifier efficiency. Furthermore, comparing Γ... ID =6dB and Γ ID With a power limit of 15dB, we can further observe that when the performance constraint is higher, the power distribution on different transmitting antennas becomes more polarized, which means that the beam pattern will have stronger directionality at this time.

[0332] Figure 7 This shows how the total system power consumption changes with the number of antennas and RF links that are turned off. We use a greedy algorithm to turn off at most N... T -K ID=24 antennas. We can see that the total power consumption initially decreases rapidly as the number of antennas turned off increases, but then begins to rise after reaching a minimum. Therefore, there exists an optimal number of antennas to minimize total power consumption—activating more or fewer antennas leads to reduced energy efficiency. Comparing the antenna selection behavior with a fixed power amplifier efficiency model and an actual nonlinear model, we can observe that the latter tends to turn off more antennas to utilize higher power amplifier efficiency, thus reaching its optimum when turning off more antennas. The slope of the latter is larger than the former, meaning that when considering the actual nonlinear power amplifier model, antenna selection has a more significant impact than under the fixed power amplifier efficiency model; that is, when the erroneous fixed power amplifier efficiency model is discarded, the actual antenna selection gain is much larger than that obtained under the erroneous model. Comparing the antenna selection behavior at different distances from the receiver to the base station, where r1 > r2, we observe that the farther the receiver is from the base station, the higher the requirement for transmitted signal quality, and the fewer antennas can be turned off, thus requiring more antennas. Furthermore, the total power consumption decreases more slowly when the RF link is turned off than when the antenna is turned off. This means that antenna selection has a greater impact on the reduction of total power consumption and should be given priority. This verifies the effectiveness of the proposed algorithm, namely, we first consider the switching of the antenna and then the switching of the RF link.

[0333] Figure 8 This is a structural block diagram of a power consumption optimization device for a base station system provided in an embodiment of the present invention, such as... Figure 8 As shown, the device includes:

[0334] The model building module 810 is used to establish a total power consumption model and system constraints of the base station system regarding the digital beamformer, the analog beamformer, and the covariance matrix used to generate the sensing signal. The total power consumption model includes the total power consumption of the transmitting antenna, the total power consumption of the radio frequency link, and the total power consumption of the phase shifter.

[0335] The beamforming optimization module 820 is used to optimize the covariance matrix of the digital beamformer, the analog beamformer, and the sensing signal based on the total power consumption model and the system constraints, with the goal of minimizing the total power consumption of the base station system.

[0336] The above scheme optimizes the total power consumption of base station systems with multiple communication users, multiple sensing targets, and multiple energy harvesting receivers. It addresses the lack of research on base station systems with multiple communication users, multiple sensing targets, and multiple energy harvesting receivers in related technologies, and further solves the problem of insufficient research on the energy consumption of multifunctional base station systems.

[0337] It should be noted that the above modules can be functional modules or program modules, and can be implemented through software or hardware. For modules implemented through hardware, the above modules can reside in the same processor; or the above modules can be located in different processors in any combination.

[0338] An embodiment of the present invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0339] Furthermore, in conjunction with the power optimization methods for base station systems provided in the above embodiments, this embodiment can also provide a storage medium for implementation. The storage medium stores a computer program; when executed by a processor, the computer program implements any of the power optimization methods for base station systems described in the above embodiments.

Claims

1. A power consumption optimization method for a base station system, characterized in that, The base station system includes: a hybrid architecture transmitter and a sensing receiver; The hybrid architecture transmitter is used to simultaneously send communication signals to multiple information decoding receivers, send sensing signals to multiple sensing targets, and send energy signals to multiple energy harvesting receivers. The sensing receivers are used to receive echo signals from multiple sensing targets. The hybrid architecture transmitter includes a digital beamformer, multiple radio frequency links, an analog beamformer, and multiple transmit antennas. The analog beamformer includes multiple phase shifters, and the sensing receiver includes multiple receive antennas. Wherein, the number of radio frequency links is greater than or equal to the number of information decoding receivers and less than or equal to the number of transmitting antennas, and the number of receiving antennas is greater than the number of transmitting antennas; The power consumption optimization method includes: Establish a total power consumption model and system constraints for the base station system regarding the digital beamformer, the analog beamformer, and the covariance matrix used to generate the sensing signal. The total power consumption model includes the total power consumption of the transmit antenna, the total power consumption of the radio frequency link, and the total power consumption of the phase shifter. Based on the total power consumption model and the system constraints, the covariance matrix of the digital beamformer, the analog beamformer, and the sensed signal is optimized with the goal of minimizing the total power consumption of the base station system. The minimum signal-to-interference-plus-noise ratio constraint of the information decoding receiver, the maximum estimation Cramer-Rao bound constraint of the sensing target, the minimum harvesting power constraint of the energy harvesting receiver, the transmit power constraint of the transmitting antenna, and the constant mode constraint of the analog beamformer; The total power consumption model is as follows: ； in, This represents the total power consumption of the base station system. This refers to the simulated beamformer. This represents the digital beamformer corresponding to the decoding receiver of the k-th signal. This represents the covariance matrix of the sensed signal; The total power consumption of the transmitting antenna is: ； This indicates the number of transmitting antennas. This represents the power amplifier efficiency of the nth transmitting antenna. This represents the maximum power amplifier efficiency of the transmitting antenna. This represents the actual transmit power of the nth transmitting antenna. This represents the maximum transmit power of the nth transmitting antenna. Represents the efficiency factor. Indicates the number of information decoding receivers. This represents a zero matrix where all elements except the nth diagonal element are 0. Indicates length is A vector whose elements are all zero except for the nth element which is 1; This indicates the total power consumption of the radio frequency link: ； Indicates the number of radio frequency links. This indicates the power consumption of a single radio frequency link when it is enabled. Indicates length is A vector whose elements are all zeros except for the nth element which is 1. This represents a zero matrix where all elements except the nth diagonal element are 0. Indicates the indicator function: ； The total power consumption of the phase shifter is: ； This indicates the power consumption of a single phase shifter when it is turned on. This represents the static power consumption of components in the base station system other than the transmitting antenna, the radio frequency link, and the phase shifter.

2. The power consumption optimization method for a base station system according to claim 1, characterized in that, The minimum signal-to-interference-plus-noise ratio constraint for the information decoding receiver is: ； This represents the minimum signal-to-interference-plus-noise ratio (SIR) of the information decoding receiver. This represents the set of information decoding receivers. The signal-to-interference-plus-noise ratio (SIR) of the receiver decoding the k-th information is: ； This represents the channel vector from the base station system to the k-th information decoding receiver. Indicates noise power. This represents the digital beamformer corresponding to the i-th information decoding receiver; The maximum estimation Cramer-Rao bound constraint for the perceived target is: ； This represents the maximum estimate of the perceived target, Cramer-Rao bound. The mean Cramer-Rao bound representing the perceived target: ； The Fischer information matrix representing the perceived target; The minimum harvesting power constraint for the energy harvesting receiver is: ； This indicates the maximum harvesting power of the energy harvesting receiver. This represents a set of energy harvesting receivers. This represents the harvesting power of the j-th energy harvesting receiver: ； The constant representing the maximum harvested power of the j-th energy harvesting receiver under circuit saturation conditions is denoted by . and Represents a constant related to the circuit specifications of the j-th energy harvesting receiver. Represents the covariance matrix of the baseband signal. This represents the channel vector from the base station system to the j-th energy harvesting receiver; The transmit power constraint of the transmitting antenna is: ； This refers to the set of transmitting antennas; The constant mode constraint of the simulated beamformer is: ； ； This refers to the set of radio frequency links.

3. The power consumption optimization method for a base station system according to claim 1, characterized in that, Optimizing the covariance matrix of the digital beamformer, the analog beamformer, and the sensed signal includes: The covariance matrix of the digital beamformer, the analog beamformer, and the sensed signal is alternately optimized.

4. The power consumption optimization method for a base station system according to claim 3, characterized in that, Alternate optimization of the covariance matrix of the digital beamformer, the analog beamformer, and the sensed signal includes: Fix the analog beamformer and optimize the covariance matrix of the digital beamformer and the sensed signal; The analog beamformer is optimized by fixing the covariance matrix of the digital beamformer and the sensed signal.

5. The power consumption optimization method for a base station system according to claim 4, characterized in that, Fixing the analog beamformer and optimizing the covariance matrix of the digital beamformer and the sensed signal includes: By fixing the simulated beamformer in the total power consumption model, a first intermediate problem model is obtained; The first intermediate problem model is equivalently transformed using the semidefinite relaxation method to obtain the second intermediate problem model; The second intermediate problem model is approximated by the continuous convex approximation method to obtain the third intermediate problem model; The third intermediate problem model is solved using a convex optimization tool to obtain the optimized covariance matrix of the digital beamformer and the sensed signal.

6. The power consumption optimization method for a base station system according to claim 1, characterized in that, Optimizing the covariance matrix of the digital beamformer, the analog beamformer, and the sensed signal includes: The covariance matrix of the digital beamformer, the analog beamformer, and the sensed signal is initially optimized to obtain the initial optimization result; Based on the initial optimization results, the beamforming weights of each of the transmitting antennas are determined. With minimizing the total power consumption of the base station system as the shutdown objective, the transmitting antennas are shut down iteratively in ascending order based on their beamforming weights, and the iterative optimization results are obtained. To ensure the feasibility of optimizing the covariance matrix of the digital beamformer, the analog beamformer, and the sensing signal, the minimum number of radio frequency links is determined through binary search. Based on the iterative optimization results, the beamforming weights of each phase shifter are determined. With the goal of minimizing the total power consumption of the base station system, the transmit antennas are shut down iteratively in ascending order based on their beamforming weights, and the final optimization result is obtained.

7. A power consumption optimization device for a base station system, characterized in that, The base station system includes: a hybrid architecture transmitter and a sensing receiver; The hybrid architecture transmitter is used to simultaneously send communication signals to multiple information decoding receivers, send sensing signals to multiple sensing targets, and send energy signals to multiple energy harvesting receivers. The sensing receivers are used to receive echo signals from multiple sensing targets. The hybrid architecture transmitter includes a digital beamformer, multiple radio frequency links, an analog beamformer, and multiple transmit antennas. The analog beamformer includes multiple phase shifters, and the sensing receiver includes multiple receive antennas. Wherein, the number of radio frequency links is greater than or equal to the number of information decoding receivers and less than or equal to the number of transmitting antennas, and the number of receiving antennas is greater than the number of transmitting antennas; The power consumption optimization device includes: The model building module is used to establish a total power consumption model and system constraints of the base station system regarding the digital beamformer, the analog beamformer, and the covariance matrix used to generate the sensing signal. The total power consumption model includes the total power consumption of the transmitting antenna, the total power consumption of the radio frequency link, and the total power consumption of the phase shifter. The beamforming optimization module is used to optimize the covariance matrix of the digital beamformer, the analog beamformer, and the sensing signal based on the total power consumption model and the system constraints, with the goal of minimizing the total power consumption of the base station system. The minimum signal-to-interference-plus-noise ratio constraint of the information decoding receiver, the maximum estimation Cramer-Rao bound constraint of the sensing target, the minimum harvesting power constraint of the energy harvesting receiver, the transmit power constraint of the transmitting antenna, and the constant mode constraint of the analog beamformer; The total power consumption model is as follows: ； in, This represents the total power consumption of the base station system. This refers to the simulated beamformer. This represents the digital beamformer corresponding to the decoding receiver of the k-th signal. This represents the covariance matrix of the sensed signal; The total power consumption of the transmitting antenna is: ； This indicates the number of transmitting antennas. This represents the power amplifier efficiency of the nth transmitting antenna. This represents the maximum power amplifier efficiency of the transmitting antenna. This represents the actual transmit power of the nth transmitting antenna. This represents the maximum transmit power of the nth transmitting antenna. Represents the efficiency factor. Indicates the number of information decoding receivers. This represents a zero matrix where all elements except the nth diagonal element are 0. Indicates length is A vector whose elements are all zero except for the nth element which is 1; This indicates the total power consumption of the radio frequency link: ； Indicates the number of radio frequency links. This indicates the power consumption of a single radio frequency link when it is enabled. Indicates length is A vector whose elements are all zeros except for the nth element which is 1. This represents a zero matrix where all elements except the nth diagonal element are 0. Indicates the indicator function: ； The total power consumption of the phase shifter is: ； This indicates the power consumption of a single phase shifter when it is turned on. This represents the static power consumption of components in the base station system other than the transmitting antenna, the radio frequency link, and the phase shifter.

Images