Resource collaborative optimization method and device integrating sensing and communication

Abstract

The invention provides a resource collaborative optimization method and device for integrating sensing and communication in an MU-MIMO system, and relates to the technical field of wireless communication, and the method comprises the steps: building a communication-sensing performance coupling model, and carrying out the joint representation of a subcarrier-level beam direction and a resource multiplexing relation, communication rate maximization and radar beam matching error minimization are converted into a multi-target optimization problem; a sensing precision constraint system based on partial subcarrier multiplexing is constructed, multiplexing subcarriers are screened through beam direction similarity measurement, and communication dedicated resources and communication sensing multiplexing resources are separated; and designing a layered iterative optimization mechanism under the hybrid beam forming architecture, taking each iteration for solving the balance problem of resultant rate and beam adaptation as one layer of the deep neural network, and optimizing the weight and step length by using back propagation. According to the invention, while the multi-user communication rate is guaranteed, target sensing precision optimization is realized, and the hardware implementation complexity of a large-scale antenna system is reduced.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and in particular to a resource collaborative optimization method and device for integrating sensing and communication in a MU-MIMO system. Background Art

[0002] With the rapid development of wireless communication technology, millimeter-wave and terahertz massive multi-user multiple-input, multiple-output (MU-MIMO) technology has become a key technology for future wireless communication systems due to its abundant spectrum resources in high-frequency bands and the spatial freedom provided by large-scale antenna arrays. As 5G-A / 6G mobile communication networks expand into millimeter-wave and terahertz high-frequency bands, the need for spectrum sharing for communication and radar sensing is becoming increasingly urgent. Integrated communication and sensing technology, achieving dual functionality through a unified hardware architecture and spectrum resources, has become a core research direction for 6G networks.

[0003] In massive multi-user MIMO-OFDM systems, subcarrier allocation strategies are a core challenge in improving ISAC system spectrum efficiency. Traditional approaches typically employ either strict subcarrier partitioning or full subcarrier sharing, but both have significant limitations. The former strictly partitions spectrum resources, allocating independent subcarriers for communication and sensing functions. While this approach can prevent interference between communication and sensing, it struggles to adapt to real-time changes in channel conditions and sensing requirements in dynamic multi-user scenarios, resulting in inefficient spectrum resource utilization. For example, when communication users are sparse, a large number of subcarriers dedicated to communication may remain idle, resulting in wasted resources. Conversely, when communication load surges, sensing subcarriers cannot be flexibly switched to communication use. The latter achieves theoretically maximum spectrum efficiency through full subcarrier sharing, but the randomness of communication data can lead to radar beam pattern mismatch. Especially in multi-user scheduling scenarios, multi-access interference between users further degrades sensing accuracy.

[0004] Therefore, the current synaesthesia integration technology has the problems of limited spectrum efficiency and performance imbalance caused by the rigid subcarrier resource reuse. Summary of the Invention

[0005] To address the existing issues of limited spectrum efficiency and performance imbalance caused by rigid subcarrier resource reuse in the integrated sensing technology, the present invention provides a method and apparatus for collaborative resource optimization of integrated sensing and communication in a MU-MIMO system. The technical solution is as follows:

[0006] In one aspect, a method for collaboratively optimizing resources for integrated sensing and communication in a MU-MIMO system is provided, comprising:

[0007] Acquire subcarrier resources for sensing and communication;

[0008] A communication-perception performance coupling model is established. By jointly characterizing the subcarrier-level beam direction and resource reuse relationship, the communication rate maximization and radar beam matching error minimization are transformed into a multi-objective optimization problem.

[0009] Based on a multi-objective optimization problem, a perception accuracy constraint system based on partial subcarrier reuse is constructed. The subcarrier reuse strategy is defined with binary variables. The reused subcarriers are screened by beam direction similarity measurement, and the communication-specific resources and communication perception reuse resources in the subcarrier resources are separated.

[0010] Based on the separated communication-specific resources and communication-sensing multiplexing resources, a hierarchical optimization mechanism is determined for the hybrid beamforming architecture. Through a deep unfolding model, each iteration of solving the communication rate-sensing accuracy trade-off problem is treated as a layer of a deep neural network. Backpropagation is used to optimize the weights and step size to generate a hybrid beamforming matrix.

[0011] Subcarrier resources are allocated according to the hybrid beamforming matrix.

[0012] Optionally, the communication-perception performance coupling model includes:

[0013] Use orthogonal frequency division multiplexing waveform to transmit synaesthesia integrated signal and define subcarrier set Each subcarrier The reuse state is determined by the binary variable r k Characterization:

[0014]

[0015] Among them, the collection represents the subcarrier set multiplexed for communication and sensing functions, r k =1 indicates that the subcarrier is a communication sensing multiplexing resource, r k =0 means the subcarrier is a dedicated communication resource;

[0016] Build a hybrid beamforming transmit signal model:

[0017] x m [k]=F RF F BBm [k]s m [k],

[0018] in is the frequency-flat analog beamforming matrix that constitutes the hybrid beamforming, and is the frequency-dependent digital beamforming matrix, is the signal vector transmitted to the mth user, where m is a positive integer m∈[0,M] and M is the total number of users. tis the number of transmitting antennas, N s N is the number of independent data streams transmitted in parallel for a single user on a subcarrier. RF is the number of RF chains;

[0019] Determine the system communication spectrum efficiency:

[0020]

[0021] Where K is the number of system subcarriers, M is the total number of users, and β m is the communication priority of user m, which is a weight coefficient belonging to (0,1). by{u m [k],F RF ,F BBm [k]}, which represents the spectrum efficiency of user m on subcarrier k and can be expressed as:

[0022]

[0023] Among them, N s It is the number of independent data streams transmitted in parallel for a single user on one subcarrier; is the interference plus noise covariance matrix, where H m [k] is the channel matrix, represents the receiving beamforming vector of user m for the kth subcarrier, N r is the number of antennas equipped by the user equipment, and I is the identity matrix;

[0024] Perception beam covariance mismatch The actual subcarrier k is transmitted

[0025] The covariance matrix of the signal is F is the norm, R[k] is solved

[0026] stC1:

[0027] C2: R[k]≥0

[0028] C3: R[k]=R[k] H

[0029] The obtained ideal beamforming covariance matrix;

[0030] in It is the ideal beam pattern corresponding to the target on subcarrier k, which has a beam width B in the target direction. width The binary vector G(θt ,f k ) is the desired transmit beam pattern, P BS is the total base station transmit power;

[0031] Construct a multi-objective optimization problem to improve communication-perception performance:

[0032]

[0033] stC1:

[0034]

[0035] Where ω can be regarded as a regularization factor, τ is the covariance mismatch of the perceptual beam, and the constraint C1 is a partial connection structure constraint, i.e., the analog precoder F RF Belongs to a group of block matrices where each block is a unit module N t / N RF dimensional vector.

[0036] Optionally, the subcarrier multiplexing strategy includes:

[0037] For each subcarrier k, determine its communication beam covariance matrix

[0038] Difference in Frobenius norm from R[k]

[0039] Press τ k Arrange all subcarriers in ascending order, select the first J subcarriers to join the subcarrier set for communication and perception functions in Wherein J is a positive integer determined according to the actual scenario requirements.

[0040] Optionally, for the coupling problem of beamforming and multiplexed subcarrier selection, a three-stage decoupling optimization is performed:

[0041] First stage fixation Optimize on all subcarriers {F RF ,F BBm [k]} to maximize

[0042] The second stage is based on the {F RF ,F BBm [k]}determine {τ k}, sort and intercept, generate

[0043] The third stage Maximize Get the final {F RF ,F BBm [k]}, where α represents the importance of communication and β represents the importance of perception.

[0044] Optionally, the hierarchical optimization mechanism includes:

[0045] Determine a deep unfolding architecture: unfold the X outer iterations of the Projected Gradient Ascent (PGA) algorithm into X layers of neural networks, each containing Z inner iterations for updating F RF and F BBm [k];;

[0046] The parameter optimization process corresponding to the deep expansion architecture includes:

[0047] With the channel matrix H m [k], noise variance and power budget P BS As input, the step size parameters μ(i,j) and λ(i) are dynamically adjusted through unsupervised training, and the optimized hybrid beamforming matrix is output.

[0048] Optionally, the loss function corresponding to the deep expansion architecture includes:

[0049]

[0050] Where ω is a regularization factor, for a fixed F RF ,F BBm [k] can be updated in the i+1th iteration by a projected gradient ascent step, i.e.:

[0051]

[0052] in, Indicates R for F RF Find the gradient, and μ (i,j) are the precoder and step size of the jth inner iteration in the i-th outer iteration, respectively, and It is the final precoder obtained after the i-th outer iteration after completing all inner iterations. Given F RF , F BBm [k] is updated in the i+1th iteration as:

[0053]

[0054] use and Indicates the step size of the outer and inner layers, initial input Channel matrix H m [k], power budget Pmax and noise variance As input, and output in the outer layer of i=1,…,X Each outer layer contains a sub-network, including Y layers to output F RF , the operations within each layer include calculating gradients and performing projections, where X and Y are positive integers. .

[0055] Optionally, the hybrid beamforming architecture satisfies: analog beamforming matrix Each sub-block f i Is an L=N t / N RF dimensional constant modulus constraint vector, i.e., |fi(j)|=1, j=1,…,L, B is a constant; the digital beamforming matrix F BBm [k] Satisfy the total power constraint of the base station The receiving end uses full digital beamforming, and the received signal of user m on subcarrier k is:

[0056]

[0057]

[0058] where u m [k] is the minimum mean square error receiver vector, H m [k] is the channel matrix, n m [k] is the channel additive white Gaussian noise.

[0059] On the other hand, an embodiment of the present invention provides a resource collaborative optimization device for integrated sensing and communication in a MU-MIMO system. The resource collaborative optimization device for integrated sensing and communication in a MU-MIMO system is used to implement the resource collaborative optimization method for integrated sensing and communication in a MU-MIMO system provided in an embodiment of the present invention. The device includes:

[0060] An acquisition module, used to acquire subcarrier resources for sensing and communication;

[0061] A model building module is used to establish a communication-perception performance coupling model. The communication-perception performance coupling model transforms maximizing communication rate and minimizing radar beam matching error into a multi-objective optimization problem by jointly characterizing the relationship between subcarrier-level beam direction and resource reuse.

[0062] A constraint system construction module is used to construct a perception accuracy constraint system based on partial subcarrier reuse based on a multi-objective optimization problem. The subcarrier reuse strategy is defined with binary variables, and multiplexed subcarriers are screened by beam direction similarity measurement to separate communication-specific resources from communication perception multiplexing resources in subcarrier resources.

[0063] A determination module is used to determine the hierarchical optimization mechanism under the hybrid beamforming architecture based on the separated communication-specific resources and communication-sensing multiplexing resources. Through the deep unfolding model, each iteration of solving the communication rate and perception accuracy trade-off problem is used as a layer of the deep neural network. The weights and step size are optimized through backpropagation to generate the hybrid beamforming matrix.

[0064] An allocation module is configured to allocate subcarrier resources according to the hybrid beamforming matrix.

[0065] On the other hand, an embodiment of the present invention provides a resource collaborative optimization device for integrated sensing and communication in a MU-MIMO system, wherein the resource collaborative optimization device for integrated sensing and communication in the MU-MIMO system includes:

[0066] processor;

[0067] A memory having computer-readable instructions stored thereon, wherein the computer-readable instructions, when executed by the processor, implement the method provided by the embodiment of the present invention.

[0068] On the other hand, a computer-readable storage medium is provided, in which a program code is stored. The program code can be called by a processor to execute the method provided by an embodiment of the present invention.

[0069] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0070] The embodiment of the present invention establishes a communication-perception performance coupling model, and transforms the maximization of communication rate and the minimization of radar beam matching error into a multi-objective optimization problem through the joint characterization of the subcarrier-level beam direction and resource reuse relationship; constructs a perception accuracy constraint system based on partial subcarrier reuse, defines the subcarrier reuse strategy with binary variables, screens the reused subcarriers through the beam direction similarity metric, and separates communication-specific resources from communication perception reuse resources; designs a hierarchical iterative optimization mechanism under the hybrid beamforming architecture, introduces a deep unfolding model, and uses each iteration of solving the trade-off problem between the combined rate and beam adaptation as a layer of a deep neural network, using backpropagation to optimize the weights and step size. The present invention breaks through the limitations of traditional subcarrier full reuse or strict allocation strategies for both communication and perception functions, achieves target perception accuracy optimization while ensuring the communication rate of multiple users, and reduces the hardware implementation complexity of large-scale antenna systems. It is suitable for high-precision environmental perception and high-speed data transmission scenarios such as intelligent transportation and industrial Internet of Things. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0072] Figure 1 This is a flow chart of a resource collaborative optimization method for integrated sensing and communication in a MU-MIMO system provided by an embodiment of the present invention;

[0073] Figure 2 This is a schematic diagram of the structure of a hybrid beamforming model of a communication and perception integrated system provided by an embodiment of the present invention;

[0074] Figure 3 This is a schematic diagram of a depth expansion structure based on gradient projection ascent provided by an embodiment of the present invention;

[0075] Figure 4 1 is a schematic diagram of the structure of a resource collaborative optimization device integrating sensing and communication in a MU-MIMO system provided by an embodiment of the present invention;

[0076] Figure 5 This is a schematic diagram of the structure of a resource collaborative optimization device integrating sensing and communication in a MU-MIMO system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0077] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0078] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0079] In the embodiments of the present invention, the terms "image" and "picture" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same. The terms "of," "corresponding," and "corresponding" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same.

[0080] In the embodiments of the present invention, sometimes a subscript such as W1 may be mistakenly written as a non-subscript form such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0081] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0082] To meet the collaborative needs of high-precision environmental perception and high-speed data transmission in scenarios such as intelligent transportation and the Industrial Internet of Things, and to address the problems of limited spectrum efficiency and performance imbalance caused by the rigid subcarrier resource reuse in tele-sensing integration, the present invention proposes a resource collaborative optimization method and device for integrated sensing and communication in a MU-MIMO system, achieving communication-perception collaborative optimization through partial subcarrier multiplexing and layered beamforming design.

[0083] like Figure 1 As shown, a resource collaborative optimization method for integrating sensing and communication in a MU-MIMO system includes:

[0084] S1. Acquire subcarrier resources for sensing and communication;

[0085] S2. Establish a communication-perception performance coupling model. This model transforms maximizing communication rate and minimizing radar beam matching error into a multi-objective optimization problem by jointly characterizing the relationship between subcarrier-level beam direction and resource reuse.

[0086] S3. Based on a multi-objective optimization problem, a perception accuracy constraint system based on partial subcarrier reuse is constructed. The subcarrier reuse strategy is defined with binary variables. The reused subcarriers are screened by beam direction similarity measurement, and the communication-specific resources and communication perception reuse resources in the subcarrier resources are separated.

[0087] S4. Based on the separated communication-specific resources and communication-sensing multiplexing resources, a hierarchical optimization mechanism is determined for the hybrid beamforming architecture. Each iteration of solving the communication rate-sensing accuracy trade-off problem is treated as a layer of a deep neural network through a deep unfolding model. The weights and step size are optimized through backpropagation to generate a hybrid beamforming matrix.

[0088] S5. Allocate subcarrier resources according to the hybrid beamforming matrix.

[0089] Optionally, the present invention first provides a modeling method for a massive multi-user MIMO communication perception integrated system. Consider a broadband MIMO ISAC system with an ISAC base station equipped with a hybrid beamforming architecture and N r The base station uses OFDM waveform to transmit radar and communication integrated signals, and consists of a base station supporting ISAC and M equipped with Nr The user equipment with 1 antenna receives the signal, and T sensing targets reflect the signal so that the base station can obtain their location. The base station is equipped with a large number of MIMO antennas on the transmitting end and adopts hybrid precoding to reduce the demand on the RF chain. On the receiving end, full digital beamforming is considered to facilitate the implementation of spatial diversity and spatial multiplexing technology on the receiving end, and improve signal processing flexibility. Indicates the subcarriers of the system. In order to ensure communication performance, all subcarriers will be used for communication, and only some subcarriers Can be reused as radar function. Define a binary variable r k Indicates whether the kth subcarrier is reused. If r k =1, it means that the kth subcarrier can be reused; if r k =0, it means it is not reused. Using the subcarrier level beamforming strategy, the multiplexed subcarriers are adjusted by the offset of the beam direction to achieve the trade-off between radar and communication performance. Data is sent on each subcarrier separately, and the baseband data stream transmitted on the kth subcarrier is used Indicates that Corresponding to the signal vector transmitted to the mth user, each information symbol constituting it is a random variable and satisfies The data streams on different subcarriers are independent and orthogonal to each other. s N is the number of independent data streams transmitted in parallel for a single user on a subcarrier. s ≤min(N t ,N r ). The signal-to-interference-plus-noise ratio (SINR) is used as a communication performance indicator, so the distribution of information symbols is not restricted. The transmitted signal is expressed as:

[0090] x m [k]=F RF F BBm [k]s m [k],

[0091] in is the frequency-flat analog beamforming matrix that constitutes the hybrid beamforming, and It is a frequency-dependent digital beamforming matrix. The frequency-flat analog beamforming matrix is the analog precoding component used to process wideband signals in hybrid beamforming. Its characteristic is that it does not change with frequency, that is, the same beamforming matrix is used on all subcarriers, simplifying hardware implementation and reducing complexity. The frequency-dependent digital beamforming matrix precodes each subcarrier in the digital domain, which can be optimized according to the channel characteristics of each subcarrier, improving the system's spectrum efficiency and signal quality. RFis the number of Radio Frequency (RF) chains, M≤N RF ≤N t In view of the characteristics of massive multiple-input, multiple-output (MIMO) with a large number of antennas and RF chains, a partial connection structure is adopted. RF Represented as a block diagonal structure:

[0092]

[0093] in Is an L=N t / N RF dimensional constant modulus constraint vector, i.e. |f i (j)|=1,j=1,…,L; there are no other hardware-related restrictions on the baseband precoder. Due to the high free space path loss, the characteristics of the high-frequency signal propagation environment are characterized by a cluster channel model. The channel modeling adopts the Saleh-Valenzuela (SV) model, and the channel matrix Expressed as:

[0094]

[0095] Where P represents the number of scattering paths, α p is the complex gain of the pth scattering path, a r (θ r,p ,f k ) and a t (θ t,p ,f k ) represent the receiving and transmitting antenna steering vectors, respectively, where θ r,p ,θ t,p and τ p Represent the angle of arrival (AOA), angle of departure (DOA) and delay of the p-th scattering path, respectively. represents the frequency of the kth subcarrier, where B represents the system bandwidth, and f c Represents the system center frequency. The steering vectors of the receiving antenna and the transmitting antenna can be expressed as:

[0096]

[0097] The signal received by user m in the frequency band of subcarrier k can be expressed as:

[0098]

[0099] in, is the channel matrix of user m at subcarrier k, represents the receiving beamforming vector of user m for the kth subcarrier, is the corresponding channel additive white Gaussian noise, In this application, it is assumed that both the transmitter and receiver have perfect channel state information. In practice, CSI can be accurately and efficiently obtained through channel estimation at the receiver and further shared at the transmitter through effective feedback technology. Therefore, based on the received signal model, the spectral efficiency for user m on subcarrier k is expressed as:

[0100]

[0101] The hybrid precoder satisfies P BS is the total transmit power of the base station, is defined as the interference plus noise covariance matrix:

[0102]

[0103] The spectral efficiency of the entire system can be expressed as:

[0104]

[0105] For perception, MIMO radar with OFDM waveform provides a larger virtual array aperture and achieves higher degrees of freedom (DoFs) than traditional phased array radar. Assume that there are T radars of interest to detect targets, with θ t ∈[-π / 2,π / 2], t=1,…,T represents the azimuth angle from the base station to the t-th target. The present invention designs a beamforming strategy to maximize the detection signal power at the target of interest and match the desired transmit beam pattern:

[0106]

[0107] About the covariance matrix design. It is the ideal beam pattern corresponding to the target on subcarrier k, which has a beam width B in the target direction. width A binary vector with 1 in each direction and 0 in the other directions. By finding a suitable R[k], G(θ t ,f k ) can be as equipped as possible with Similar spatial characteristics are expressed as:

[0108]

[0109] stC1:

[0110] C2:R[k]≥0

[0111] C3:R[k]=R[k] H

[0112] Where C1 represents the beamforming power constraint expressed by the covariance matrix, C2 represents the semi-positive definite property of the covariance matrix, and C3 represents the conjugate symmetric property of the covariance matrix. Solving this problem will yield the covariance matrix R[k] that best achieves radar beam alignment. From another perspective, the covariance matrix of the actual transmitted signal on subcarrier k is

[0113]

[0114] Expanded to:

[0115]

[0116] When designing the beamforming strategy, the actual covariance matrix C[k] is made as close as possible to the ideal R[k]. The perceived beam covariance mismatch is expressed as:

[0117]

[0118] The perception function is only This part of the subcarrier is considered, while the other subcarriers are fully used for communication. τ is defined as the perceived beam covariance mismatch, which is used to indirectly measure the degree of deviation between the radar beam direction and the ideal direction.

[0119] Furthermore, the present invention provides a communication-perception coupled multi-objective optimization model and a corresponding transformation method, which simultaneously designs the receive beamformer u m [k], digital precoder F at the transmitter BBm [k], the unit mode part is connected to the analog beamformer F RF and a subcarrier set multiplexed for radar sensing functions To obtain the maximum communication spectrum efficiency and the minimum perception beam mismatch error, it is expressed as:

[0120]

[0121] stC1:

[0122] C2:

[0123] C3:

[0124] The constraint C1 is a partial connection structure constraint, i.e., the analog precoder F RF Belongs to a group of block matrices where each block is a unit modulo N t / N RF dimensional vector, C2 is the base station transmit power budget constraint, and C3 explains that the multiplexed subcarriers belong to a subset of the total subcarrier set. Multi-objective optimization problems often involve multiple conflicting objectives, such as the spectrum efficiency of the communication system and the beam matching error of the perception system. Directly solving multi-objective optimization problems is usually more complicated because it is necessary to consider the trade-offs of multiple objectives at the same time. By converting the multi-objective optimization problem into a single-objective optimization problem, the solution process can be simplified and the problem can be made easier to handle. By assigning importance weights to multiple objectives, the multi-objective optimization problem is converted into:

[0125]

[0126] stC1:

[0127] C2:

[0128] C3:

[0129] Among them, α represents the importance of communication, and β represents the importance of perception. In the collaborative obstacle avoidance scenario of the Internet of Vehicles, it is necessary to prioritize the radar's perception refresh rate of obstacles. In this case, β>α should be set; in the industrial AR quality inspection scenario, the demand for high-definition video streaming is dominant, and α>β should be set. In addition, hybrid scenarios (such as low-altitude inspections by drones) adopt an α≈β balance mode, and dynamically adjust the weights through an online feedback mechanism to adapt to the real-time demand fluctuations of communication-perception. The problem is further transformed into the optimization of communication spectrum efficiency under the constraint of perception accuracy. The larger the beam matching error, the lower the perception accuracy. Therefore, by limiting the beam matching error to ensure that the perception accuracy meets the specific business requirements, the transformed problem is expressed as:

[0130]

[0131] stC1:

[0132] C2:

[0133] C3:τ≤τ0

[0134] C4:

[0135] The newly added constraint C3 is the radar perception performance constraint, which sets a minimum requirement for the alignment of the radar beam with the target on the multiplexed subcarriers to ensure the sensing capability in these desired sensing directions.

[0136] Furthermore, the present invention proposes a subcarrier reuse screening mechanism based on beam direction similarity. The core idea is to dynamically select the subcarriers with the highest beam direction consistency for synaesthesia resource reuse by quantifying the matching degree of communication and perception beams in spatial spectrum distribution, thereby achieving collaborative optimization of spectrum efficiency and perception accuracy. The mechanism first calculates the Frobenius norm difference {τ1,…,τ K}, to characterize the spatial energy focusing deviation between the two in the target direction. On this basis, all subcarriers are sorted in ascending order according to the beam matching error, and the first J subcarriers with the highest matching degree are preferentially selected to form a multiplexing set to ensure that the energy coverage of the multiplexing resources in the direction of the perceived target in the spatial dimension is maximized. J is determined according to the needs of the actual scenario. For example, a larger J is set for scenarios with high requirements for perception accuracy, such as intelligent driving, to improve perception performance. In order to reduce the complexity of system optimization, the mechanism adopts a decoupled design architecture: in the first stage, the system performs beamforming optimization on all subcarriers based on pure communication performance indicators (for example, system communication spectrum efficiency or multi-user weighted combined rate) to generate an initial communication beam pattern; in the second stage, by calculating the similarity measure between the communication beam of each subcarrier and the preset perception beam, subcarriers with high spatial spectrum overlap are selected as the multiplexing candidate set; in the third stage, communication-perception joint beamforming optimization is performed only on the selected multiplexing subcarriers. Furthermore, to address demand fluctuations in dynamic environments, this mechanism embeds an adaptive weight adjustment strategy in the multiplexed subcarrier optimization process. By providing real-time feedback on the perceived accuracy deviation and communication rate threshold, it dynamically adjusts the penalty coefficient for the beam matching error term, thereby establishing a closed-loop control between the rigid constraints of resource reuse and the flexible requirements of performance balance. This mechanism uses beam direction similarity screening to naturally align multiplexed subcarriers in the spatial dimension with the perceived target direction, effectively suppressing the perceived signal-to-noise ratio degradation caused by beam mismatch. The decoupled optimization architecture transforms the combinatorial optimization problem into a low-complexity sorting and subset optimization problem, making the algorithm computationally feasible in real time within millimeter-wave / terahertz massive MIMO systems.

[0137] Furthermore, the present invention provides a hybrid beamforming design method for a communication-aware integrated system based on deep expansion, and develops a multi-objective learning framework based on the projected gradient ascent (PGA) method. This enables efficient updating of {F RF ,F BBm [k]} to simultaneously maximize the communication weighted spectrum efficiency J and minimize the radar beam error τ. For the multiplexed subcarriers, the beamforming design problem is:

[0138]

[0139] stC1:

[0140] C2:

[0141] The importance weight coefficients of communication and perception are changed into one by proportional means, and ω = β / α is set, so the problem is transformed into

[0142]

[0143] stC1:

[0144] C2:

[0145] Where ω can be considered as a regularization factor. In principle, ω should be determined by the maximum beam error. However, here it is considered as a given hyperparameter and iterated in the subsequent depth expansion. For a fixed F RF ,F BBm [k] can be updated in the i+1th iteration by a projected gradient ascent step, i.e.:

[0146]

[0147] in, Indicates R for F RF Find the gradient. Given F RF , F BBm [k] can be updated in the i+1th iteration as:

[0148]

[0149] Afterwards, update F for multiple iterations RF , then update F BBm [k], and Apply weight η so that F RF In the iterative process, BBm [k] keep in sync. Let X represent the number of outer iterations and Y represent the update F RF The number of inner iterations, F RF The update is expressed as:

[0150]

[0151] in and μ (i,j) are the precoder and step size of the jth inner iteration in the i-th outer iteration, respectively, and is the final precoder obtained after the i-th outer layer iteration after completing all inner layer iterations. On the other hand, F BBm [k] Updated as follows:

[0152]

[0153] Next, we design a deep neural network (DNN) based on the unfolded PGA. The algorithm’s X outer iterations are unfolded into X layers of neural networks, each layer contains Y inner iterations for updating F. RF and F BBm [k]. The task of the model is to output feasible analog and digital beamforming with good communication and perception performance {F RF ,F BBm [k]}, that is, maximize use and Indicates the step size of the outer and inner layers, initial input Channel matrix H m [k], power budget P max and noise variance as input and output in the outer layer at i=1,…,I Each outer layer contains a sub-network, including Y layers to output F RF The operations within each layer include calculating gradients and performing projections. The loss function is expressed as:

[0154]

[0155] The loss function follows and the original expression of τ. The loss function L(μ,λ) enables the model to be trained in an unsupervised manner. In particular, the dataset consists of multiple channel realizations, and the noise power is set to And randomly select P for each data sample max ∈[γ min ,γ max ] enhances the learned hyperparameters so that the corresponding signal-to-noise ratio is within the range of interest. The loss L(μ,λ) is a function of the step size {μ,λ}, because depending on and {μ,λ}. The unfolded PGA model is trained to optimize {μ,λ} to achieve the best trade-off within I iterations.

[0156] Finally, the precoding optimization problem at the receiving end can be expressed as:

[0157]

[0158] If a hybrid beamforming receiver is used, the problem becomes a dual problem with the transmitter problem after removing the transmit power constraint, and can also be solved using a deep expansion method.

[0159] As attached Figure 2 and attached Figure 3As shown, the base station is equipped with a massive MIMO antenna array (for example, 256 antennas) and adopts a partially connected hybrid beamforming architecture, with each RF chain connected to an independent subarray. The receiving user equipment adopts a fully digital beamforming structure to support multi-user parallel communication. The system operates in the millimeter wave frequency band and adopts an OFDM waveform. The total number of subcarriers is N, and the bandwidth of each subcarrier is B / N. Parameter configuration is performed, and the communication weight factor α and the perception weight factor β are initialized according to the scenario requirements (for example, α = 0.7 in the intelligent driving scenario); the number of multiplexed subcarriers J is set to a scenario-related value (for example, J = K / 3 for normalized low-altitude intrusion detection, and J = K / 2 for intelligent driving); the transmission power P BS Set according to the number of antennas, channel status, operating frequency band, etc.

[0160] Then establish the communication-perception performance coupling model. Indicates the subcarriers of the system. In order to ensure communication performance, all subcarriers will be used for communication, and only some subcarriers Reused as radar function. Define a binary variable r k Indicates whether the kth subcarrier is reused. If r k =1, it means that the kth subcarrier is reused; if r k =0, it means it is not reused. Using the subcarrier level beamforming strategy, the multiplexed subcarriers are adjusted by the offset of the beam direction to achieve the trade-off between radar and communication performance. Data is sent on each subcarrier separately, and the baseband data stream transmitted on the kth subcarrier is used Indicates that Corresponding to the signal vector transmitted to the mth user, each information symbol constituting it is a random variable and satisfies N s N is the number of independent data streams transmitted in parallel for a single user on a subcarrier. s ≤min(N t ,N r ). The transmitted signal is expressed as:

[0161] x m [k]=F RF F BBm [k]s m [k].

[0162] in is the frequency-flat analog beamforming matrix that constitutes the hybrid beamforming, and is the frequency-dependent digital beamforming matrix. N RF is the number of radio frequency (RF) chains, M≤N RF ≤N tConsidering the characteristics of massive MIMO with a large number of antennas and RF chains, a partial connection structure is adopted. RF Represented as a block diagonal structure:

[0163]

[0164] in Is an L=N t / N RF dimensional constant modulus constraint vector, i.e. |f i (j)|=1,j=1,…,L. The channel model adopts the Saleh-Valenzuela model, and the channel matrix Expressed as:

[0165]

[0166] Where P represents the number of scattering paths, α p is the complex gain of the pth scattering path, a r (θ r,p ,f k ) and a t (θ t,p ,f k ) represent the receiving and transmitting antenna steering vectors, respectively, where θ r,p ,θ t,p and τ p represent the angle of arrival (AOA), angle of departure (DOA) and delay of the p-th scattering path, respectively. represents the frequency of the kth subcarrier, where B represents the system bandwidth, and f c Represents the system center frequency. The steering vectors of the receiving antenna and the transmitting antenna can be expressed as:

[0167]

[0168] The signal received by user m in the frequency band of subcarrier k can be expressed as:

[0169]

[0170] in, is the channel matrix of user m at subcarrier k, represents the receiving beamforming vector of user m for the kth subcarrier, is the corresponding channel additive white Gaussian noise,

[0171] The spectrum efficiency of user m on subcarrier k is expressed as:

[0172]

[0173] The hybrid precoder satisfies P BS is the total transmit power of the base station, is defined as the interference plus noise covariance matrix:

[0174] The spectral efficiency of the entire system can be expressed as:

[0175]

[0176] For perception, MIMO radar with OFDM waveform provides a larger virtual array aperture and achieves higher degrees of freedom (DoFs) than traditional phased array radar. t ∈[-π / 2,π / 2], t=1,…,T represents the azimuth angle from the base station to the t-th target. The present invention designs a beamforming strategy to maximize the detection signal power at the target of interest, or more generally, to match the desired transmit beam pattern:

[0177]

[0178] Next, we will focus on the covariance matrix Design. d (θ t ,f k ) is the ideal beam pattern corresponding to the target on subcarrier k, which is a beam width B in the target direction. width A binary vector with 1 in each direction and 0 in the other directions. By finding a suitable R[k], G(θ t ,f k ) can be as equipped as possible with Similar spatial characteristics are expressed as:

[0179]

[0180] stC1:

[0181] C2:R[k]≥0

[0182] C3:R[k]=R[k] H

[0183] Where C1 represents the beamforming power constraint expressed by the covariance matrix, C2 represents the semi-positive definite property of the covariance matrix, and C3 represents the conjugate symmetric property of the covariance matrix. Solving this problem will yield the covariance matrix R[k] that best achieves radar beam alignment. From another perspective, the covariance matrix of the actual transmitted signal on subcarrier k is

[0184]

[0185] Expanded to:

[0186]

[0187] When designing the beamforming strategy, the actual covariance matrix C[k] is made as close as possible to the ideal R[k]. The perceived beam covariance mismatch is expressed as:

[0188]

[0189] The perception function is only This part of the subcarrier is considered, while the other subcarriers are fully used for communication. τ is defined as the perceived beam covariance mismatch, which is used to indirectly measure the degree of deviation between the radar beam direction and the ideal direction.

[0190] Next, for each subcarrier k, calculate its communication beam covariance matrix Difference in Frobenius norm from R[k] Press τ k Arrange all subcarriers in ascending order, select the first J subcarriers to form the reuse set J, where J is determined based on the actual scenario needs. For example, in scenarios such as intelligent driving that require high perception accuracy, a larger J is set to improve perception performance.

[0191] For multiplexed subcarriers, the beamforming design problem is:

[0192]

[0193] stC1:

[0194] C2:

[0195] The importance weight coefficients of communication and perception are converted into one in a proportional manner, and ω = β / α is set. Then the problem is transformed into:

[0196]

[0197] stC1:

[0198] c2:

[0199] Where ω can be considered as a regularization factor. In principle, ω should be determined by the maximum beam error. However, here it is considered as a given hyperparameter and iterated in the subsequent depth expansion. For a fixed F RF ,F BBm [k] can be updated in the i+1th iteration by a projected gradient ascent step, i.e.:

[0200]

[0201] in, Indicates R for F RF Find the gradient. Given F RF , F BBm [k] can be updated in the i+1th iteration as:

[0202]

[0203] Afterwards, update F for multiple iterations RF , then update F BBm [k], and Apply weight η so that F RF In the iterative process, BBm [k] keep in sync. Let X represent the number of outer iterations and Y represent the update F RF The number of inner iterations, F RF The update is expressed as:

[0204]

[0205] in and μ (i,j) are the precoder and step size of the jth inner iteration in the i-th outer iteration, respectively, and is the final precoder obtained after the i-th outer layer iteration after completing all inner layer iterations. On the other hand, F BBn [k] Updated as follows:

[0206]

[0207] The present invention also provides a deep neural network (DNN) based on the expanded PGA, where the E outer iterations of the algorithm are expanded into an E-layer neural network, each layer contains F inner iterations for updating F RF and F BBm [k]. The task of the model is to output feasible analog and digital beamforming with good communication and perception performance {F RF ,F BBm [k]}, that is, maximize use and Indicates the step size of the outer and inner layers, initial input Channel matrix H m [k], power budget P max and noise variance as input and output in the outer layer at i=1,…,I Each outer layer contains a sub-network, including F layers to output F RF The operations within each layer include calculating gradients and performing projections. The loss function is expressed as:

[0208]

[0209] The loss function follows and the original expression of τ. The loss function L(μ,λ) enables the model to be trained in an unsupervised manner. In particular, the dataset consists of multiple channel realizations, and the noise power is set to And randomly select P for each data sample max ∈[γ min ,γ max ] enhances the learned hyperparameters so that the corresponding signal-to-noise ratio is within the range of interest. The loss L(μ,λ) is a function of the step size {μ,λ}, because depending on and {μ,λ}. The unfolded PGA model is trained to optimize {μ,λ} to achieve the best trade-off within I iterations.

[0210] Finally, the precoding optimization problem at the receiving end can be expressed as:

[0211]

[0212] If a hybrid beamforming receiver is used, the problem becomes a dual problem with the transmitter problem after removing the transmit power constraint, and can also be solved using a deep expansion method.

[0213] Deep unfolding is an emerging deep learning method that stems from the need to address the lack of interpretability and difficulty integrating domain knowledge in traditional deep learning models. By combining traditional optimization algorithms with deep learning, it improves model performance and interpretability. In recent years, deep unfolding has been widely applied and developed in various fields, such as image super-resolution, signal processing, and wireless communications. In wireless communications, deep unfolding has been applied to tasks such as signal detection, channel estimation, and beamforming design, providing strong support for the development of 6G technology. Its core principle is to treat each step of an iterative optimization algorithm as a network layer, thereby transforming the optimization algorithm into a deep neural network. Specifically, an iterative optimization algorithm relevant to the problem is first selected. Each iteration in the algorithm is then unfolded into a layer in the network, each containing learnable parameters. These parameters are optimized in an end-to-end manner during training, allowing the entire network to better adapt to the data. In this way, the deep unfolding network inherits the interpretability and domain knowledge integration capabilities of traditional optimization algorithms while also possessing the efficiency and flexibility of deep learning.

[0214] To address the existing issues of limited spectrum efficiency and performance imbalance caused by rigid subcarrier resource reuse in the integrated sensing technology, the present invention provides a method and apparatus for collaborative resource optimization of integrated sensing and communication in a MU-MIMO system. The technical solution is as follows:

[0215] On the other hand, Figure 4 The embodiment of the present invention provides a resource collaborative optimization device for integrated sensing and communication in a MU-MIMO system. The resource collaborative optimization device for integrated sensing and communication in a MU-MIMO system is used to implement the resource collaborative optimization method for integrated sensing and communication in a MU-MIMO system provided in an embodiment of the present invention. The device includes:

[0216] Acquisition module 401, used to acquire subcarrier resources for sensing and communication;

[0217] A model building module 402 is configured to establish a communication-perception performance coupling model. The communication-perception performance coupling model converts maximizing communication rate and minimizing radar beam matching error into a multi-objective optimization problem by jointly characterizing the relationship between subcarrier-level beam direction and resource reuse.

[0218] Constraint system construction module 403 is used to construct a perception accuracy constraint system based on partial subcarrier reuse based on a multi-objective optimization problem, define the subcarrier reuse strategy with binary variables, screen the reused subcarriers based on beam direction similarity measurement, and separate the communication-specific resources and communication-aware reuse resources in the subcarrier resources;

[0219] Determination module 404 is configured to determine a hierarchical optimization mechanism under a hybrid beamforming architecture based on the separated communication-specific resources and communication-sensing multiplexing resources. Each iteration of solving the communication rate-sensing accuracy trade-off problem is treated as a layer of a deep neural network through a deep unfolding model. Weights and step sizes are optimized through backpropagation to generate a hybrid beamforming matrix.

[0220] The allocation module 405 is configured to allocate subcarrier resources according to the hybrid beamforming matrix.

[0221] On the other hand, an embodiment of the present invention provides a resource collaborative optimization device for integrated sensing and communication in a MU-MIMO system, wherein the resource collaborative optimization device for integrated sensing and communication in the MU-MIMO system includes:

[0222] processor;

[0223] A memory having computer-readable instructions stored thereon, wherein the computer-readable instructions, when executed by the processor, implement the method provided by the embodiment of the present invention.

[0224] On the other hand, a computer-readable storage medium is provided, in which a program code is stored. The program code can be called by a processor to execute the method provided by an embodiment of the present invention.

[0225] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0226] The embodiment of the present invention establishes a communication-perception performance coupling model, and transforms the maximization of communication rate and the minimization of radar beam matching error into a multi-objective optimization problem through the joint characterization of the subcarrier-level beam direction and resource reuse relationship; constructs a perception accuracy constraint system based on partial subcarrier reuse, defines the subcarrier reuse strategy with binary variables, screens the reused subcarriers through the beam direction similarity metric, and separates communication-specific resources from communication perception reuse resources; designs a hierarchical iterative optimization mechanism under the hybrid beamforming architecture, introduces a deep unfolding model, and uses each iteration of solving the trade-off problem between the combined rate and beam adaptation as a layer of a deep neural network, using backpropagation to optimize the weights and step size. The present invention breaks through the limitations of traditional subcarrier full reuse or strict allocation strategies for both communication and perception functions, achieves target perception accuracy optimization while ensuring the communication rate of multiple users, and reduces the hardware implementation complexity of large-scale antenna systems. It is suitable for high-precision environmental perception and high-speed data transmission scenarios such as intelligent transportation and industrial Internet of Things.

[0227] Figure 5 FIG is a schematic diagram of a resource collaborative optimization device integrating sensing and communication in a MU-MIMO system according to an embodiment of the present invention. Figure 5As shown, optionally, the resource collaborative optimization device 510 integrating sensing and communication in the MU-MIMO system may include a first processor 2001.

[0228] Optionally, the resource collaborative optimization device 510 integrating sensing and communication in the MU-MIMO system may further include a memory 2002 and a transceiver 2003 .

[0229] The first processor 2001, the memory 2002 and the transceiver 2003 may be connected via a communication bus.

[0230] The following combination Figure 5 The components of the resource collaborative optimization device 510 for integrating sensing and communication in the MU-MIMO system are described in detail.

[0231] The first processor 2001 is the control center of the resource collaborative optimization device 510 for integrated sensing and communication in the MU-MIMO system, and can be a processor or a collective term for multiple processing elements. For example, the first processor 2001 is one or more central processing units (CPUs), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement an embodiment of the present invention, such as one or more microprocessors (digital signal processors, DSPs), or one or more field programmable gate arrays (FPGAs).

[0232] Optionally, the first processor 2001 may execute various functions of the resource collaborative optimization device 510 integrating sensing and communication in the MU-MIMO system by running or executing a software program stored in the memory 2002 and calling data stored in the memory 2002 .

[0233] In a specific implementation, as an embodiment, the first processor 2001 may include one or more CPUs, such as Figure 5 CPU0 and CPU1 are shown in FIG.

[0234] In a specific implementation, as an embodiment, the resource collaborative optimization device 510 integrating sensing and communication in the MU-MIMO system may also include multiple processors, such as Figure 51 and 2. The first processor 2001 and the second processor 2004 are shown in FIG. Each of these processors can be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). A processor herein can refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0235] The memory 2002 is used to store the software program for executing the solution of the present invention, and is controlled by the first processor 2001 for execution. The specific implementation method can refer to the above method embodiment and will not be repeated here.

[0236] Optionally, the memory 2002 may be a read-only memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM) or other type of dynamic storage device that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 2002 may be integrated with the first processor 2001 or exist independently and be connected to the first processor 2001 through the interface circuit ( Figure 5 (not shown) is coupled to the first processor 2001, which is not specifically limited in this embodiment of the present invention.

[0237] The transceiver 2003 is used to communicate with a network device or a terminal device.

[0238] Optionally, the transceiver 2003 may include a receiver and a transmitter ( Figure 5 (not shown separately in the figure). The receiver is used to implement a receiving function, and the transmitter is used to implement a sending function.

[0239] Optionally, the transceiver 2003 may be integrated with the first processor 2001 or may exist independently and be connected to the interface circuit of the resource collaborative optimization device 510 integrating sensing and communication in the MU-MIMO system ( Figure 5(not shown) is coupled to the first processor 2001, which is not specifically limited in this embodiment of the present invention.

[0240] It should be noted that Figure 5 The structure of the resource collaborative optimization device 510 for integrating sensing and communication in the MU-MIMO system shown in the figure does not constitute a limitation on the router. The actual knowledge structure identification device may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0241] In addition, the technical effects of the resource collaborative optimization device 510 integrating sensing and communication in the MU-MIMO system can refer to the technical effects of the multimodal emotion recognition method described in the above method embodiment, and will not be repeated here.

[0242] It should be understood that the first processor 2001 in the embodiment of the present invention may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0243] It should also be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0244] The above embodiments can be implemented in whole or in part through software, hardware (such as circuits), firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the processes or functions described in accordance with the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, motor drive, or data center to another website, computer, motor drive, or data center via infrared, microwave, or other means. The computer-readable storage medium can be any available medium that can be accessed by a computer, or a data storage device such as a motor drive or data center that includes a collection of one or more available media. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, or a magnetic tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0245] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0246] In this disclosure, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, "at least one of a, b, or c" can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural.

[0247] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0248] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0249] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0250] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of the device or unit, which can be electrical, mechanical or other forms.

[0251] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0252] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0253] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a motor driver, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0254] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A resource collaborative optimization method for integrating sensing and communication in a MU-MIMO system, characterized in that: include: Acquire subcarrier resources for sensing and communication; A communication-perception performance coupling model is established. By jointly characterizing the subcarrier-level beam direction and resource reuse relationship, the communication rate maximization and radar beam matching error minimization are transformed into a multi-objective optimization problem. Based on a multi-objective optimization problem, a perception accuracy constraint system based on partial subcarrier reuse is constructed. The subcarrier reuse strategy is defined with binary variables. The reused subcarriers are screened by beam direction similarity measurement, and the communication-specific resources and communication perception reuse resources in the subcarrier resources are separated. Based on the separated communication-specific resources and communication-sensing multiplexing resources, a hierarchical optimization mechanism is determined for the hybrid beamforming architecture. Through a deep unfolding model, each iteration of solving the communication rate-sensing accuracy trade-off problem is treated as a layer of a deep neural network. Backpropagation is used to optimize the weights and step size to generate a hybrid beamforming matrix. Subcarrier resources are allocated according to the hybrid beamforming matrix.

2. The method according to claim 1, characterized in that The communication-perception performance coupling model includes: Use orthogonal frequency division multiplexing waveform to transmit synaesthesia integrated signal and define subcarrier set Each subcarrier The reuse state is determined by the binary variable r k Characterization: Among them, the collection represents the subcarrier set multiplexed for communication and sensing functions, r k =1 indicates that the subcarrier is a communication sensing multiplexing resource, r k =0 means the subcarrier is a dedicated communication resource; Build a hybrid beamforming transmit signal model: x m [k]=F RF F BBm [k]s m [k], in is the frequency-flat analog beamforming matrix that constitutes the hybrid beamforming, and is the frequency-dependent digital beamforming matrix for user m, is the signal vector transmitted to the mth user, where m is a positive integer m∈[0,M], and M is the total number of users; t is the number of transmitting antennas, N s N is the number of independent data streams transmitted in parallel for a single user on a subcarrier. RF is the number of RF chains; Determine the system communication spectrum efficiency: Where K is the number of system subcarriers, M is the total number of users, and β m is the communication priority of user m, which is a weight coefficient belonging to (0,1). by{u m [k],F RF ,F BBm [k]}, which represents the spectrum efficiency of user m on subcarrier k and can be expressed as: Among them, N s It is the number of independent data streams transmitted in parallel for a single user on one subcarrier; is the interference plus noise covariance matrix, where H m [k] is the channel matrix, represents the receiving beamforming vector of user m for the kth subcarrier, N r is the number of antennas equipped by the user equipment, I is the unit matrix, F BBj [k] represents the frequency-dependent user shaping matrix of other users except user m; Perception beam covariance mismatch The covariance matrix of the actual transmitted signal on subcarrier k is F is the norm, R[k] is solved s.t.C1: C2:R[k]≥0 C3:R[k]=R[k] H The obtained ideal beamforming covariance matrix; in It is the ideal beam pattern corresponding to the target on subcarrier k, which has a beam width B in the target direction. width The binary vector G(θ t ,f k ) is the desired transmit beam pattern, P BS is the total base station transmit power; Construct a multi-objective optimization problem to improve communication-perception performance: s.t.C1: C2: Where ω can be regarded as a regularization factor, τ is the covariance mismatch of the perceptual beam, and the constraint C1 is a partial connection structure constraint, i.e., the analog precoder F RF Belongs to a group of block matrices where each block is a unit module N t / N RF dimensional vector.

3. The method according to claim 2, characterized in that The subcarrier multiplexing strategy includes: For each subcarrier k, determine its communication beam covariance matrix Difference in Frobenius norm from R[k] Press τ k Arrange all subcarriers in ascending order, select the first J subcarriers to join the subcarrier set for communication and perception functions in Wherein J is a positive integer determined according to the actual scenario requirements.

4. The method according to claim 3, characterized in that To address the coupling issue between beamforming and multiplexed subcarrier selection, a three-stage decoupling optimization is performed: First stage fixation Optimize on all subcarriers {F RF ,F BBm [k]} to maximize The second stage is based on the {F RF ,F BBm [k]} determine {τ k }, sort and intercept, generate The third stage Maximize Get the final {F RF ,f BBm [k]}, where α represents the importance of communication and β represents the importance of perception.

5. The method according to claim 1, wherein The layered optimization mechanism includes: Determine a deep unfolding architecture: unfold the X outer iterations of the Projected Gradient Ascent (PGA) algorithm into X layers of neural networks, each containing Z inner iterations for updating F RF and F BBm [k]; The parameter optimization process corresponding to the deep expansion architecture includes: With the channel matrix H m [k], noise variance and power budget P BS As input, the step size parameters μ(i,j) and λ(i) are dynamically adjusted through unsupervised training, and the optimized hybrid beamforming matrix is output.

6. The method according to claim 5, characterized in that The loss functions corresponding to the deep expansion architecture include: Where ω is a regularization factor, for a fixed F RF ,F BBm [k] can be updated in the i+1th iteration by a projected gradient ascent step, i.e.: in, Indicates R for F RF Find the gradient, and μ (i,j) are the precoder and step size of the jth inner iteration in the i-th outer iteration, respectively, and It is the final precoder obtained after the i-th outer iteration after completing all inner iterations. Given f RF , F BBm [k] is updated in the i+1th iteration as: use and Indicates the step size of the outer and inner layers, the initial input Channel matrix H m [k], power budget P max and noise variance As input, and output in the outer layer of i=1,…,X Each outer layer contains a sub-network, including Y layers to output F RF , the operations within each layer include calculating gradients and performing projections, where X and Y are positive integers.

7. The method according to claim 2, characterized in that The hybrid beamforming architecture satisfies: analog beamforming matrix Each sub-block f i Is an L=N t / N RF dimensional constant modulus constraint vector, i.e. |f i (j)|=1,j=1,…,L, B is a constant; the digital beamforming matrix F BBm [k] Satisfy the total power constraint of the base station The receiving end uses full digital beamforming, and the received signal of user m on subcarrier k is: where u m [k] is the minimum mean square error receiver vector, H m [k] is the channel matrix, n m [k] is the channel additive white Gaussian noise.

8. A resource collaborative optimization device for integrated sensing and communication in a MU-MIMO system, wherein the resource collaborative optimization device for integrated sensing and communication in a MU-MIMO system is configured to implement the resource collaborative optimization method for integrated sensing and communication in a MU-MIMO system according to any one of claims 1 to 7, characterized in that: The device comprises: An acquisition module, used to acquire subcarrier resources for sensing and communication; A model building module is used to establish a communication-perception performance coupling model. The communication-perception performance coupling model transforms maximizing communication rate and minimizing radar beam matching error into a multi-objective optimization problem by jointly characterizing the relationship between subcarrier-level beam direction and resource reuse. A constraint system construction module is used to construct a perception accuracy constraint system based on partial subcarrier reuse based on a multi-objective optimization problem. The subcarrier reuse strategy is defined with binary variables, and multiplexed subcarriers are screened by beam direction similarity measurement to separate communication-specific resources from communication perception multiplexing resources in subcarrier resources. A determination module is used to determine the hierarchical optimization mechanism under the hybrid beamforming architecture based on the separated communication-specific resources and communication-sensing multiplexing resources. Through the deep unfolding model, each iteration of solving the communication rate and perception accuracy trade-off problem is used as a layer of the deep neural network. The weights and step size are optimized through backpropagation to generate the hybrid beamforming matrix. An allocation module is configured to allocate subcarrier resources according to the hybrid beamforming matrix.

9. A resource collaborative optimization device integrating sensing and communication in a MU-MIMO system, characterized in that: The resource collaborative optimization device integrating sensing and communication in the MU-MIMO system includes: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program code, which can be called by a processor to execute the method according to any one of claims 1 to 7.

Images