A method for optimizing a common-sense modal adjustable ISAC system

Abstract

This invention discloses an optimization method for an integrated sensing and communication (ISAC) system with adjustable sensing modes, relating to the field of integrated sensing and communication system optimization technology. The method includes: constructing an integrated sensing and communication (ISAC) system model comprising a base station, users, sensing targets, and a reconfigurable smart surface (RIS); calculating the importance score of each RIS unit based on the state information of the base station-RIS channel and the RIS-user channel, and determining the configuration strategy of active and passive reflection units based on the importance scores to form a hybrid RIS structure; using the weighted sum of the communication rate and sensing beam gain of the ISAC system as the objective optimization function of the hybrid RIS unit, and jointly optimizing the complex reflection coefficient matrix of the RIS and the transmission precoding matrix of the base station under the constraint of the base station transmit power, thereby optimizing the ISAC system to operate in the optimal sensing mode. This optimization method enables free adjustment of various modes of the ISAC system, comprehensively improving both the communication rate and the sensing beam gain.

Description

Technical Field

[0001] This invention generally relates to the field of integrated sensing and communication system optimization technology, and specifically to an optimization method for an ISAC system with adjustable sensing modes. Background Technology

[0002] Integrated Sensing and Communications (ISAC) technology, as a core pillar of sixth-generation wireless communication systems, integrates communication and radar sensing functions by sharing hardware and spectrum resources. It aims to improve spectrum efficiency, reduce device power consumption and cost, and provide native support for emerging applications such as autonomous driving and smart cities.

[0003] In this technological system, the Reconfigurable Intelligent Surface (RIS), as a novel intelligent metamaterial surface, enables real-time intelligent control of incident electromagnetic waves through software programmability, providing a new dimension of low power consumption and low cost for optimizing wireless transmission environments. However, currently, RIS-assisted integrated sensing and communication systems are suitable for power-constrained conditions. Active RIS suffers from limited reflection gain, while passive RIS suffers from channel noise coupling, leading to mutual constraints on sensing functions. Random-phase RIS, on the other hand, suffers from non-convexity and relaxation issues, resulting in slow convergence rates. Therefore, there is an urgent need in this field for a novel RIS-assisted ISAC system optimization method to achieve intelligent tunability of sensing modes and synergistic optimization of system performance. Summary of the Invention

[0004] In view of the above-mentioned defects or deficiencies in the prior art, it is desirable to provide an optimization method for an ISAC system with adjustable synesthesia modality.

[0005] This invention provides an optimization method for a tunable sensing modality ISAC system, comprising: An ISAC system model is constructed, comprising a base station, a user, a sensing target, and a RIS-based system; wherein the RIS comprises reflective units with different reflective characteristics; the reflective units with different reflective characteristics include: active reflective units and passive reflective units; Based on the status information of the base station to RIS channel and the status information of the RIS to user channel, the importance score of each RIS unit is calculated, and the configuration strategy of active reflection unit and passive reflection unit is determined based on the importance score to form a hybrid RIS structure. The weighted sum of the communication rate and sensing beam gain of the ISAC system is used as the objective optimization function of the hybrid RIS unit, and the reflection beam parameters of the RIS and the transmission beam parameters of the base station are jointly optimized under the base station transmit power constraint. The optimized reflected beam parameters and transmitted beam parameters are loaded into the ISAC system corresponding to the ISAC system model to drive the ISAC system to operate in the optimal synesthetic mode.

[0006] According to the technical solution provided by the present invention, determining the configuration strategy of active and passive reflective units based on the importance score includes: The importance scores of each RIS unit are sorted in descending order to obtain a priority index; Based on the priority index and the preset number of active reflection units, the RIS units corresponding to the index of the first preset number of active reflection units in the priority index vector are configured as active reflection units; the remaining RIS units in the priority index vector P are configured as passive reflection units.

[0007] According to the technical solution provided by the present invention, the weighted sum of the communication rate and sensing beam gain of the ISAC system is used as the objective optimization function of the hybrid RIS unit, including: Obtain the total communication rate of the ISAC system and the sensing beam gain in the ISAC system in the spatial direction associated with the sensing target. By introducing a sensing adjustment coefficient and combining it with a weighted calculation of the total communication rate and the sensing beam gain, a target optimization function is constructed.

[0008] According to the technical solution provided by the present invention, obtaining the total communication rate of the ISAC system includes: Based on the direct channel from base station to user, the base station to RIS channel, and the RIS to user channel, an equivalent communication channel model including direct links and reflection links is constructed. Based on the equivalent communication channel model and the base station transmitted signal, the user terminal received signal model is determined; Based on the received signal model, the signal-to-interference-plus-noise ratio (SNR) of each user is calculated, and the total communication rate of the system is calculated based on the SNR of each user.

[0009] According to the technical solution provided by the present invention, the reflection beam parameter is the complex reflection coefficient matrix of the RIS; the transmission beam parameter is the transmission precoding matrix of the base station; Under base station transmit power constraints, the joint optimization of the reflection beam parameters of the RIS and the transmit beam parameters of the base station includes: When the transmission precoding matrix is ​​fixed Initialize the complex reflection coefficient matrix and set its elements to satisfy the unit modulus constraint; Calculate the Euclidean gradient of the transformed objective function with respect to the complex reflection coefficient matrix; The Euclidean gradient is projected onto the tangent space of the unit circular manifold to obtain the Riemann gradient; The update amount of the complex reflection coefficient matrix is ​​calculated based on the Riemann gradient and the Riemann-Hesse operator-determined Riemann-Newton direction. The optimization step size is adaptively adjusted based on the norm of the Riemann gradient and the norm of the Riemann-Hesse operator. The complex reflection coefficient matrix is ​​iteratively updated according to the update amount and the step size until the convergence condition is met, so as to output the jointly optimized complex reflection coefficient matrix.

[0010] According to the technical solution provided by the present invention, under the constraint of base station transmit power, jointly optimizing the reflection beam parameters of the RIS and the transmit beam parameters of the base station includes: When the complex reflection coefficient matrix is ​​fixed Lagrange multipliers are introduced, and an optimization objective equation for the transmit precoding matrix is ​​constructed based on the mean square error of the total communication rate and the objective optimization function. Taking the partial derivative of the Lagrange function with respect to the transmit precoding matrix and setting it to zero yields the optimal solution expression for the transmit precoding matrix with respect to the Lagrange multipliers. Substitute the optimal solution expression into the optimization objective equation to construct the power equation for the Lagrange multiplier; The power equation is solved by an iterative algorithm, the values ​​of the Lagrange multipliers are updated, and the updated Lagrange multipliers are substituted into the optimal solution expression to obtain the jointly optimized transmit precoding matrix.

[0011] According to the technical solution provided by the present invention, the complex reflection coefficient matrix integrates the phase shift component shared by the active reflection unit and the passive reflection unit, as well as the amplitude gain component possessed by the active reflection unit.

[0012] According to the technical solution provided by the present invention, the inductive adjustment coefficient is used to adjust the relative weights of the total communication rate and the sensing beam gain; When the value of the sensing adjustment coefficient is within the first numerical range, the ISAC system is configured as a sensing-first mode with sensing performance as the primary consideration. When the value of the induction adjustment coefficient is within the second numerical range, the ISAC system is configured as a communication priority mode that prioritizes communication performance. When the value of the synesthesia adjustment coefficient is in the third numerical range, the ISAC system is configured as a balanced mode with balanced synesthesia performance.

[0013] In summary, this technical solution specifically discloses an optimization method for an inductively modal adjustable ISAC system, including: constructing an ISAC system model comprising a base station, a user, a sensing target, and a RIS-based system; wherein the reconfigurable smart surface comprises reflective units with different reflective characteristics; calculating the importance score of each RIS unit based on the state information of the base station to RIS channel and the state information of the RIS to user channel, and determining the configuration strategy of active and passive reflective units based on the importance score to form a hybrid RIS structure; using the weighted sum of the communication rate and sensing beam gain of the ISAC system as the objective optimization function of the hybrid RIS unit, and jointly optimizing the reflection beam parameters of the RIS and the transmission beam parameters of the base station under the constraint of the base station transmit power; loading the optimized reflection beam parameters and transmission beam parameters into the ISAC system corresponding to the ISAC system model to drive the ISAC system to operate in the optimal inductive mode.

[0014] Beneficial effects: The ISAC system model constructed in this application includes both active and passive units, fundamentally solving the inherent problems of channel noise coupling and limited reflection gain faced by single-type RIS in ISAC systems. At the same time, by using the weighted sum of communication rate and sensing beam gain as the optimization objective, it achieves the integration of communication and sensing, enabling the ISAC system to intelligently find the optimal balance between communication and sensing performance according to top-level requirements. This allows for the free adjustment of various modes of the ISAC system, enabling the integrated sensing and communication system to effectively enhance the signal strength pointing towards the user and sensing target, while significantly suppressing interference and noise in other directions, thus synchronously and significantly improving the overall performance of the integrated sensing and communication system. Attached Figure Description

[0015] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram of the architecture of an ISAC system with adjustable synesthetic modes.

[0016] Figure 2 This is a flowchart illustrating an optimization method for a synesthetic modality-tunable ISAC system.

[0017] Figure 3 This is a flowchart illustrating step S300 in an optimization method for a synesthetic modality-tunable ISAC system.

[0018] Figure 4 This is a flowchart illustrating step S301 in an optimization method for a synesthetic modality-tunable ISAC system.

[0019] Figure 5This is a comparison chart of communication-sensing curves under different methods.

[0020] Figure 6 A comparison of communication-sensing curves for different numbers of active RIS reflector units.

[0021] Figure 7 This is a comparison chart of communication-sensing curves at different signal-to-noise ratios. Detailed Implementation

[0022] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0023] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0024] Example 1 To make the technical solutions of the embodiments of the present invention clearer and easier to understand, the application background of the embodiments of the present invention will be introduced below.

[0025] Integrated sensing and communication technologies, as the core pillar of sixth-generation wireless communication systems, achieve the integration of communication and radar sensing functions by sharing hardware and spectrum resources. This aims to improve spectrum efficiency, reduce equipment power consumption and cost, and provide native support for emerging applications such as autonomous driving and smart cities.

[0026] like Figure 1 As shown, integrated sensing and communication technologies, through the design of a unified system architecture and signal waveforms, enable base stations to simultaneously perform two core functions: first, as communication access points, transmitting data with user terminals; and second, as sensing centers, detecting targets by analyzing the reflected echoes of their transmitted signals in the environment (including the sensed targets). This means that the signals transmitted by the base station carry user communication data on one hand, and serve as radar waves for detecting the environment on the other. Users are the recipients of communication services, while targets in the environment become the source of sensing information. Through the coordinated scheduling of base stations and intelligent reflectors, both share the same spectrum and hardware resources, ultimately achieving a paradigm shift from separate communication and sensing to a symbiotic integration of sensing and communication, thereby improving system efficiency and intelligence.

[0027] In this technological system, the Reconfigurable Intelligent Surface (RIS), as a novel type of intelligent metamaterial surface, intelligently controls incident electromagnetic waves in real time through software programmability. It is often deployed in wireless propagation environments between base stations, users, and sensing targets. It dynamically controls the electromagnetic response of each unit through integrated control circuits to achieve intelligent reflection of wireless signals. In terms of communication, RIS can generate high-gain reflected beams pointing towards users to compensate for signal blind spots in obstructed areas and enhance the signal for users; or generate null beams to suppress interference to other users, thereby significantly improving communication speed and coverage. In terms of sensing, RIS can work in conjunction with the base station's transmitted beams to form a highly concentrated beam that accurately illuminates the sensing target to enhance the echo signal, thereby significantly improving the accuracy and distance of sensing.

[0028] Furthermore, RIS (Radio Reflectors) are divided into active reflectors and passive reflectors. Active reflectors are units within the RIS that possess signal amplification capabilities. They typically integrate active circuitry such as power amplifiers, and when reflecting incident signals, they not only change the signal's phase but also increase its amplitude (i.e., provide an amplitude gain greater than 1) to overcome path loss and noise, thereby improving communication distance and the strength of the sensed signal. Passive reflectors, on the other hand, are units within the RIS that do not possess signal amplification capabilities. They are composed of passive components and, when reflecting incident signals, can only change the signal's phase, not its amplitude. This allows for flexible beam steering and waveform shaping with extremely low power consumption, intelligently reconstructing the wireless propagation environment.

[0029] Currently, RIS-assisted integrated sensing and communication systems are suitable for transmit power-constrained conditions. However, some systems suffer from limitations in reflection gain due to the presence of only source reflection units, while others suffer from channel noise coupling due to the presence of only passive reflection units, leading to mutual constraints on sensing and communication functions. Random-phase RIS systems, on the other hand, suffer from non-convexity and relaxation issues, resulting in a slow convergence rate.

[0030] To address the inherent problems of the three types of RIS models mentioned above, a sensory modality-tunable ISAC system optimization method based on a hybrid RIS model improved for ISAC spatial architecture is proposed. Specifically, this method includes: constructing an ISAC system model containing a base station, users, sensing targets, and RIS; wherein the RIS contains reflection units with different reflection characteristics; calculating the importance score of each RIS unit based on the state information of the base station to RIS channel and the state information of the RIS to user channel, and determining the configuration strategy of active and passive reflection units based on the importance score to form a hybrid RIS unit; using the weighted sum of the communication rate and sensing beam gain of the ISAC system as the objective optimization function of the hybrid RIS unit, and jointly optimizing the reflection beam parameters of the RIS and the transmission beam parameters of the base station under the constraint of the base station transmit power; loading the optimized reflection beam parameters and transmission beam parameters into the ISAC system corresponding to the system model to drive the ISAC system to operate in the optimal sensory mode.

[0031] As can be seen, the embodiments of the present invention construct a RIS structure that combines active gain and passive phase, fundamentally solving the inherent problems of channel noise coupling and limited reflection gain faced by a single type of RIS in an ISAC system. Simultaneously, using the weighted sum of communication rate and sensing beam gain as the optimization objective is the core of achieving integrated communication and sensing, enabling the ISAC system to intelligently find the optimal balance between communication and sensing performance based on top-level requirements (by adjusting weights), providing a foundation for achieving modal tunability. Joint optimization of the RIS reflection beam and the base station transmit beam constitutes a smart antenna array capable of collaborative operation. In practical applications, this effectively enhances the signal strength in the target direction while suppressing interference, thereby maximizing the overall system performance under power constraints.

[0032] Please refer to the following. Figure 2 The flowchart shown in this embodiment illustrates an optimization method for an ISAC system with adjustable inductive modality. The execution entity of this embodiment can be the ISAC system itself or a host computer deployed at or connected to a base station (specifically, it can be a baseband processing unit inside the base station, an edge server connected to the base station, or a control center on the network side). For clarity in subsequent descriptions, we will define an ISAC system assisted by a hybrid RIS structure, where the base station has… M A uniform linear array of antennas, employing a spatial hybrid structure design, can provide... K A single-antenna user provides communication services and is able to sense N A single-point target; the following is a further explanation of the steps of this invention, which includes the following steps: S100. Construct an ISAC system model that includes base stations, users, sensing targets, and RIS-based components; wherein, RIS includes reflection units with different reflection characteristics; the reflection units with different reflection characteristics include: active reflection units and passive reflection units; See Figure 1 To build an integrated sensing and communication technology system, this embodiment of the invention first defines the basic architecture of the ISAC system, which includes four core entities: base station, user, sensing target, and RIS. At the same time, by introducing RIS composed of units with different reflection characteristics, it lays the groundwork for subsequent intelligent resource allocation and performance.

[0033] The different reflection units with different reflection characteristics mentioned here refer to the subsequent active and passive reflection units in this embodiment of the invention. Passive reflection units are used to inherit their low-power, low-cost wide-area beamforming capabilities, enabling flexible signal redirection over a wide range. Active reflection units are used to provide critical signal amplification capabilities, overcome path loss and noise on the most important links, and improve the signal-to-noise ratio, so that the ISAC system can both focus and enhance critical signals and adjust the propagation direction of a large number of signals at low cost.

[0034] S200. Based on the status information of the base station to RIS channel and the status information of the RIS to user channel, calculate the importance score of each RIS unit, and determine the configuration strategy of active reflection unit and passive reflection unit based on the importance score to form a hybrid RIS structure. The configuration of active and passive reflection units follows clear principles in this invention. By calculating and ranking the importance score of each RIS unit, units can be dynamically and selectively configured as active or passive. This method ensures that limited active resources are precisely allocated to the links most critical to improving system performance, thereby forming an optimal hybrid RIS hardware structure and maximizing hardware resource utilization efficiency.

[0035] Specifically, the process of "calculating the importance score of each RIS unit based on the status information of the base station to the RIS channel and the status information of the RIS to the user channel" in the aforementioned step S200 is as follows: Assume that the ISAC system assisted by a hybrid RIS architecture is composed of L ( L = I × J It consists of ) units, of which, I and J Let be the row and column numbers of the RIS matrix, then for a RIS unit ( i , j The importance score can be calculated using the following formula (1): Formula (1); in, For the RIS to user channel in the RIS unit ( i , j The coefficient of ) For the base station to RIS channel in the RIS unit ( i , j The coefficient of ) For RIS unit ( i , j Importance score.

[0036] It should be noted that here This refers to the status information from the RIS to the user channel. This refers to the state information of the channel from the base station to the RIS. These two states can be directly obtained and determined by the ISAC system through its built-in channel estimation function. For example, by sending known pilot signals from the base station and the user, a series of preset guidance vectors can be reconstructed for the smart surface configuration. Through joint processing and feedback of the signals at the receiving end, the channel coefficients can be estimated. Alternatively, the ISAC system can utilize its own sensing capabilities to process reflected or scattered signals from the environment to assist in inferring and updating the channel state information.

[0037] Next, the aforementioned process for determining the configuration strategy of active and passive reflective units based on importance scores includes: Step A1: Sort the importance scores of each RIS unit in descending order to obtain the priority index; Step A21: Based on the priority index and the preset number of active reflection units, configure the RIS unit corresponding to the index of the first preset number of active reflection units in the priority index vector as an active reflection unit. Step A3: Configure the remaining RIS units in the priority index vector P as passive reflection units.

[0038] Specifically, after calculating the importance score of each RIS unit using formula (1), the importance score matrix can be obtained. The importance score matrix can be vectorized in column-major order to obtain the vector. The priority index P obtained by sorting it in descending order satisfies the following formula (2): Formula (2); in, p i For the first priority index in the priority index vector P i Units, p1 corresponds to the initial position of the most important RIS unit in the column priority vector; where the importance score is proportional to the channel gain potential that the position of the RIS unit may provide on the base station-user link.

[0039] Next, after obtaining the priority index P, the RIS unit matrix is ​​vectorized in column-major order, and the priority index is applied to generate vectors sorted by importance. , Assume that the preset number of active reflective units is N. B The preset number of active reflective units is used as a preset threshold, and the importance ranking of each RIS unit is represented by "0" and "1". For example, if the importance score of a RIS unit is within the preset number of active reflective units in the importance ranking, then the vector... Conversely, if the importance score of a RIS unit is not within the preset number of active reflector units in the importance ranking, then the vector... .

[0040] in, Indicates an active reflective unit. The passive reflection unit is essentially a RIS unit that, based on the priority index P, is associated with the number of RIS units in the entire RIS unit hierarchy that have active reflection units configured before it, sorted by importance, and its corresponding unit identifier is assigned. The output is 1 if the signal is not 1, and 0 otherwise. This ensures that the limited signal amplification capability is dynamically and accurately applied to the channel path that has the greatest potential to improve system performance.

[0041] Of course, the configuration strategy for active and passive reflective units can also be determined by setting an importance threshold, for example, setting the importance threshold to N. A Importance score greater than or equal to N A The RIS unit was determined to be an active reflector unit, and its importance score was less than N. A The RIS unit was determined to be a passive reflection unit.

[0042] Alternatively, we can define a vector D based on the aforementioned vector D, where the proportion of source and reflection units is specified. p The first configuration quantity can be obtained based on the preset active reflection unit ratio. The first configuration quantity is the percentage of active reflection units in the total number of RIS units. For example, if the preset active reflection unit ratio is 70% and there are 100 RIS units, then the first configuration quantity of active reflection units should be 70. Taking the first configuration quantity of 70 as an example, 70 RIS units can be selected from the priority index as active reflection units, and the remaining RIS units can be selected as passive reflection units.

[0043] As can be seen, by introducing a hybrid RIS structure, this embodiment of the invention achieves performance close to that of a purely active RIS at the cost of a small number of active units, while maintaining the low power consumption advantage of a passive RIS. This avoids the need to deploy expensive and high-power traditional sensing infrastructure (such as dedicated radar), effectively reducing the hardware and operating costs of realizing the integrated sensing function.

[0044] In a preferred embodiment, in order to uniformly characterize the physical behavior of the hybrid RIS structure and to provide a reference for subsequent optimization and synergistic adjustment, a RIS complex reflection coefficient model is introduced. Its core parameter is the complex reflection coefficient matrix, which integrates the phase shift component shared by the active and passive reflection units, as well as the amplitude gain component of the active reflection unit.

[0045] Specifically, assuming vector The first in p i The complex reflection coefficient of each element is given by the following formula (3): Formula (3); in, For phase shift, The amplitude gain of the active reflector unit; For the first p i The complex reflection coefficient of each unit; The cell identifier (active or passive) of the RIS cell serves as a decision factor to distinguish the proportion of active and passive reflective cells. This is the complex exponential form, representing a unit complex vector; For a passive element, its complex reflection coefficient The formula is as follows (4): Formula (4); For an active element, its complex reflection coefficient The formula is as follows (5): Formula (5); Therefore, in this embodiment of the invention, the complex reflection coefficients of the hybrid RIS unit structure can be expressed by a diagonal matrix as follows: (6) Formula (6); in, Represents the phase shift vector, unit modulus phase. This is used for all RIS units. Therefore, the complex reflection coefficient matrix... It includes passive phase shift components and active gain components. Additional gain is introduced for the active elements, determined by the vector. Further capture the additional gain provided by the active unit .

[0046] Based on the setting of the complex reflection coefficient of the hybrid RIS unit structure, it can control the collaborative work of active and passive units while truly reflecting the physical characteristics of the hardware. At the same time, it effectively links the hybrid RIS unit structure with the subsequent joint optimization algorithm, greatly simplifying the design and computational complexity of the subsequent joint optimization algorithm.

[0047] S300: The weighted sum of the communication rate and sensing beam gain of the ISAC system is used as the objective optimization function of the hybrid RIS unit, and the reflection beam parameters of the RIS and the transmission beam parameters of the base station are jointly optimized under the base station transmit power constraint. After confirming the configuration of the RIS unit, the next step is to set the weighted sum of the communication rate and the sensing beam gain as the optimization objective, and jointly optimize the beam parameters of the base station and the RIS under the base station transmit power constraint, so as to transform the qualitative requirement of modal tunability into a solvable quantitative optimization problem.

[0048] Specifically, see Figure 3 and Figure 4 The process of using the weighted sum of the communication rate and sensing beam gain of the ISAC system as the objective optimization function of the hybrid RIS unit includes the following steps: S301. Obtain the total communication rate of the ISAC system and obtain the sensing beam gain in the ISAC system in the spatial direction associated with the sensing target. In this embodiment of the invention, the total communication rate and the sensing beam gain are defined and obtained as follows: (1) This invention obtains the total communication rate of the ISAC system by building an end-to-end multiple-input multiple-output (MIMO) communication system model. The specific steps are as follows: S3011. Based on the direct channel from the base station to the user, the channel from the base station to the RIS, and the channel from the RIS to the user, construct an equivalent communication channel model that includes direct links and reflected links; Assumption For users k The symbols transmitted by the synesthetic function follow a complex Gaussian distribution with zero mean and unit variance, as shown in the following formula (7): Formula (7); According to the principle of multipath propagation in channels, K The signal matrix Y received by each user simultaneously includes the direct channel from the base station to the user and the RIS-assisted reflection channel, as shown in the following formula (8): Formula (8); in, It is a direct channel from the base station to the user. It is the base station to RIS channel. It is the RIS to user channel. It is an equivalent channel matrix that combines direct channels and RIS-assisted reflection channels, that is, an equivalent communication channel model that includes direct links and reflection links; It is the complex reflection coefficient matrix of the hybrid RIS structure. It is an additive white Gaussian noise matrix on the user.

[0049] S3012. Based on the equivalent communication channel model and the base station transmitted signal, determine the user terminal's received signal model; Furthermore, the transmitted signal matrix Defined as follows (9): Formula (9); Among them, F k For users k The transmit precoding matrix; For users k The symbol for synergistic signal transmission; S represents the transmitted signal matrix; Therefore, users k The user-end received signal model (or optimal receiver equalizer u) adopted k The following formula (10) is used: Formula (10); in, For noise power; F k For users k The transmit precoding matrix; From base station to user k The equivalent channel matrix; I is the identity matrix; q This refers to any user index and is often used in summation to iterate through all users. This is the conjugate transpose of the matrix.

[0050] S3013. Based on the received signal model, calculate the signal-to-interference-plus-noise ratio for each user, and calculate the total communication rate of the system based on the signal-to-interference-plus-noise ratio for each user.

[0051] Based on the above formula (10), end users k The signal-to-interference-plus-noise ratio (SINR) is given by the following formula (11): Formula (11); in, For users k The signal-to-interference-plus-noise ratio;K It is the total number of users per antenna; F q It is all other users (except user) k The transmit precoding matrix of ) means that these signals are interference to the user.

[0052] final, K The total communication rate achievable by a user can be calculated using the following formula (12): Formula (12); in, This represents the total communication rate. For users k The ratio of signal to interference plus noise.

[0053] (2) The ISAC system in this invention can use the transmitted unified signal to detect the potential target state in a specified spatial direction in order to obtain the sensing beam gain in the ISAC system in the spatial direction associated with the sensing target. Specifically, its sensing beam gain is as follows (13). Formula (13); in, The trace of the matrix; for N Each sensing direction response matrix; For norm; F k For users k The transmit precoding matrix; To sense beam gain.

[0054] Furthermore, N Individual sensing direction response matrix The formula is as follows (14): Formula (14); in, It corresponds to the first r Perceptual components in each direction, at the same time r Take 1~ N .

[0055] S302. Introduce the sensing adjustment coefficient and combine it with the weighted operation of the total communication rate and the sensing beam gain to construct the target optimization function.

[0056] In this embodiment of the invention, considering the base station transmit power constraint, the transmit beam can be jointly optimized by using a hybrid RIS structure to assist the ISAC system in utilizing the maximum transmit power. Finally, the target optimization function set by the present invention is as follows (15). Formula (15); Where st represents the condition that satisfies the following constraints ( ); Among them, the synesthesia adjustment coefficient The inductive dual-mode function for balancing systems It is the total achievable communication rate. It is the maximum communication rate. It is the total amount of information that can be perceived. It is the maximum value that can be perceived in total quantity. This is the maximum transmit power of the base station. It can be seen that both inductive and asymmetric values ​​have been normalized, ensuring consistency in their ranges in the dimensionless case, which is beneficial. The adjustment settings are used to adjust the synesthetic mode operation of the ISAC system; F k For users k The transmit precoding matrix; For the first k A complex reflection coefficient matrix.

[0057] After obtaining the target optimization function, the step of "jointly optimizing the reflection beam parameters of the RIS and the transmission beam parameters of the base station under the base station transmit power constraint" can be performed; specifically, the reflection beam parameters are the complex reflection coefficient matrix of the RIS; the transmission beam parameters are the transmission precoding matrix of the base station.

[0058] At the communication level, the transmit precoding matrix can use digital signal processing to shape the data streams sent to different users, ensuring that the signal energy is precisely directed to their respective target users in space. This enhances the signal strength of the target users and suppresses or even eliminates interference to other users (achieving spatial division multiple access). At the sensing level, the shape of the transmit beam also determines the signal energy and direction illuminating the sensing target (such as vehicles or drones). A well-optimized transmit beam can concentrate energy to illuminate the area or target to be sensed, just like a radar beam, thereby improving the quality of the echo signal and ultimately enhancing the sensing accuracy.

[0059] It should be noted that since the signals received by communication users and sensing targets are the result of the combined effect of the base station's transmit beam and the RIS reflection beam, changing either one will affect the channel state of the entire system. Therefore, the present invention adopts a joint optimization approach, which allows the transmit beam and the reflection beam to work together in a clever way, so that while the same transmit signal illuminates the sensing target, its path reflected by the RIS can also provide high-quality service to communication users.

[0060] Furthermore, this application employs an alternating optimization approach to jointly optimize the transmit precoding matrix and the complex reflection coefficient matrix. The specific process is as follows: (1) Optimize the complex reflection coefficient matrix; First, we need to define the complex reflection coefficient matrix. It is a complex diagonal matrix, when the precoding matrix is ​​transmitted. When fixed, the following optimization steps are involved: Step H1: Initialize the complex reflection coefficient matrix and set its elements to satisfy the unit modulus constraint, i.e., follow the following formula (16): Formula (16); Where st represents the condition that satisfies the following constraints ( ); It is a special set called a manifold, and here it specifically refers to a complex circular manifold.

[0061] In this step, the objective optimization function is transformed into the objective function in formula (16). The following is based on Optimize the complex reflection coefficient matrix for the objective function; Step H2: Calculate the Euclidean gradient of the transformed objective function with respect to the complex reflection coefficient matrix; The objective function Embedded After spatial arrangement, relative to the reflection coefficient matrix The Euclidean gradient is given by the following formula (17): Formula (17); in, The Euclidean gradient; Step H3: Project the Euclidean gradient onto the tangent space of the unit circular manifold to obtain the Riemann gradient; When optimizing a unit circular manifold, the Euclidean gradient must be projected onto the current point. The tangent space of the manifold is represented by the following formula (18): Formula (18); in, For element-wise multiplication; for taking the real part of a complex number; For projection operators; vector The conjugate transpose of; To take the real part of the complex number; = All are represented as Riemann gradients.

[0062] Step H4: Calculate the update amount of the complex reflection coefficient matrix based on the Riemann gradient and the Riemann-Hesse operator-determined Riemann-Newton direction. The first t The complex-valued reflection matrix at the nth iteration is denoted as . Then, the update according to the gradient descent of the Riemannian manifold is as follows (19): Formula (19); in, To make the first t The complex-valued reflection matrix at +1 iterations is denoted as follows: This refers to the learning rate or step size during iteration; At point Riemann gradient at the location; This is an exponential mapping.

[0063] Step H5: Adaptively adjust the optimization step size based on the norm of the Riemann gradient and the norm of the Riemann-Hesse operator; To improve the efficiency of manifold search, second-order gradient information is introduced to adaptively adjust the step size and capture the local curvature of the objective function. For the objective function, the Riemann-Hessian function on the complex circular manifold is expressed by the following formula (20): Formula (20); in, The Riemann-Hessian matrix; Furthermore, in each iteration, the update direction is adjusted by the Riemann gradient and the curvature scaling step size to achieve perceptual optimization. Thus, the... t The update rule for the +1st iteration can be extended to the following formula (21): Formula (21); in, This indicates that the solution is sought in the Riemann-Newtonian direction; This indicates that the search direction is predetermined using curvature; Ultimately, step length The adaptive adjustment rule is as follows (22): Formula (22); Among them, the square of the Riemann gradient norm Used to measure the current gradient; when the gradient is large, it is passed through... Avoid oscillations by reducing the step size too much.

[0064] Riemann-Hesse operator norm Used to measure local curvature; when the curvature is large, the curvature is optimized by reducing the step size.

[0065] It should be noted that one of the hyperparameters The overall step size is used to control the adaptive step size as follows (23): Formula (23); in, For the first t This adaptive hyperparameter is used in the next iteration; These are the initial hyperparameters for the adaptive hyperparameters; T This represents the maximum number of convergence iterations. This represents the attenuation rate.

[0066] Due to another adaptive hyperparameter To balance the importance of gradient and curvature, an adaptive formula (24) is used: Formula (24): in, Let be the norm of the current Riemann gradient. Let be the norm of the initial Riemann gradient; It is a very small positive number; For the first t This adaptive hyperparameter is used in the next iteration; These are the initial hyperparameters for the adaptive hyperparameters; Step H6: Iteratively update the complex reflection coefficient matrix according to the update amount and step size until the convergence condition is met, so as to output the jointly optimized complex reflection coefficient matrix.

[0067] Based on the above rules, the algorithm iteratively updates the objective function until it considers a sufficiently good solution to have been found, at which point it stops. The convergence condition can be gradient convergence (e.g., when the norm of the Riemann gradient reaches a preset threshold, the convergence condition is considered satisfied); or convergence due to changes in the solution (e.g., when the solution changes very little between two consecutive iterations, the convergence condition is considered satisfied); or the iteration loop terminates when the maximum number of iterations has been reached. After the algorithm converges, the final iteration yields... What we get is the final complex reflection coefficient matrix.

[0068] As can be seen, the embodiments of this invention address a hybrid constraint problem faced in the optimization of hybrid RIS: passive elements are constrained by unit mode (non-convex), while active elements are constrained by amplitude (convex). The Riemann-Hesse algorithm is used to adaptively optimize the RIS phase shift matrix. Through the mathematical framework of manifold optimization, the non-convex unit mode constraint of passive elements and the amplitude constraint of active elements are unified and processed within a single geometric space, fundamentally ensuring the feasibility and accuracy of the optimized solution. The key lies in treating the phase of all elements as a holistic variable for collaborative optimization. The Riemann gradient and Riemann-Hesse matrix, which capture global correlations, guide the algorithm's update direction, thereby achieving deep collaboration and global performance optimization between active and passive element beams, rather than simple functional superposition. Simultaneously, the adaptive step-size mechanism introduced by the algorithm dynamically adjusts according to local curvature, balancing convergence speed and stability in complex optimization terrains.

[0069] As can be seen, this invention, through dynamic configuration of a hybrid RIS structure, enables active units to provide gain on critical links to overcome path loss and noise, while passive units achieve wide-area beamforming with low power consumption. This simultaneously avoids the channel noise coupling problem of a fully passive RIS and the limited reflection gain problem of an active RIS at the system level. At the same time, the configuration strategy based on importance scores and the Riemannian manifold optimization algorithm avoids the non-convexity and relaxation problems of random phase RIS, improving the convergence rate and optimization accuracy.

[0070] (2) When optimizing the transmission precoding matrix, i.e. when the reflection coefficient matrix When fixed, embodiments of the present invention use the Lagrange multiplier method to apply the transmit precoding matrix. The optimization process includes the following steps: Step G1: Introduce Lagrange multipliers and construct an optimization objective equation for the transmission precoding matrix based on the mean square error of the total communication rate and the objective optimization function; By introducing Lagrange multipliers and the mean square error of the total communication rate, formula (12) is rewritten to obtain the following formula (25): Formula (25); in, These are the optimal weighting coefficients, used as optimization variables for the weighted mean square error; Mean square error; For users k The ratio of signal to interference plus noise.

[0071] Specifically, Defined as follows (26): Formula (26); and, The corresponding optimal weighting coefficient is: Formula (27); Therefore, the objective function of formula (15) and the constraint of base station transmit power can be rewritten as formula (28) as follows: Formula (28); Next, The objective equation for optimization is defined as follows (29): Formula (29); Step G2: Take the partial derivative of the Lagrange function with respect to the transmission precoding matrix and set it to zero to obtain the optimal solution expression of the transmission precoding matrix with respect to the Lagrange multipliers; Considering the base station transmit power constraint, Lagrange multipliers are introduced. Formula (29) is used to calculate the transmit precoding matrix. The partial derivatives are used to obtain the following formula (30): Formula (30); If the result of equation (30) is 0, then: Formula (31); For the sake of brevity, matrices A and b are introduced into formula (31). k ,right Eigenvalue decomposition yields the following formula (32): Formula (32); Among them, diagonal array A unitary matrix is ​​an eigenvalue matrix; a unitary matrix is ​​an eigenvector matrix. It is the eigenvector matrix. It is an eigenvector of the identity norm.

[0072] Therefore, the solution can be obtained using formula (31). Finally, the optimal solution expression of the transmit precoding matrix with respect to the Lagrange multipliers is obtained as follows: Formula (33) Formula (33); Step G3: Substitute the optimal solution expression into the optimization objective equation to construct the power equation for the Lagrange multipliers; Next, substituting the result of formula (33) into formula (28), we obtain the power equation for the Lagrange multipliers as follows: formula (34): Formula (34); Formula (34) above can reflect the actual transmission power of the base station.

[0073] in, Will Transforming from the standard basis to the characteristic basis represents the user. k exist The amount on, User k Corresponding antenna m Power gain.

[0074] Step G4: Solve the power equation using an iterative algorithm, update the values ​​of the Lagrange multipliers, and substitute the updated Lagrange multipliers into the optimal solution expression to obtain the jointly optimized transmit precoding matrix.

[0075] Next, by differentiating formula (34), we can obtain the following formula (35): Formula (35); The Lagrange multipliers can then be obtained using Newton's iterative method. No. n+ The update for one iteration is given by the following formula (36): Formula (36); Finally, the optimal Lagrange multiplier is obtained by iterating continuously using Newton's method until it is found. Then, the optimal Lagrange multiplier is substituted into the above formula (33) to obtain the optimized emission precoding matrix.

[0076] As can be seen, this invention uses a set of alternating iterative mathematical algorithms to collaboratively design the base station's transmit beam and the RIS's reflect beam, enabling the overall performance of communication and sensing to reach global optimality under transmit power and hardware constraints, thereby achieving intelligent and dynamic control of the system's operating modes.

[0077] S400: The optimized reflected beam parameters and transmitted beam parameters are loaded into the ISAC system corresponding to the system model to drive the ISAC system to operate in the optimal synesthetic mode.

[0078] After obtaining the optimized transmit precoding matrix and complex reflection coefficient matrix, both are loaded into the ISAC system to complete the overall optimization of the hybrid RIS-assisted ISAC system. Generally, the optimized transmit precoding matrix is ​​loaded into the signal processing unit of the base station to configure the transmit link. The optimized complex reflection coefficient matrix is ​​sent to the RIS controller through the control link (such as a wired or dedicated wireless control channel). The RIS controller adjusts the state of each adjustable element according to the received parameters.

[0079] In a preferred embodiment, the induction adjustment coefficient The relative weights used to adjust the total communication rate and the sensing beam gain; When the value of the synesthesia adjustment coefficient is in the first range, the ISAC system is configured as a perception-first mode that prioritizes perception performance. When the value of the induction adjustment coefficient is in the second range, the ISAC system is configured as a communication priority mode that prioritizes communication performance. When the synesthesia adjustment coefficient is in the third range, the ISAC system is configured as a balanced mode with balanced synesthesia performance.

[0080] For example, once the parameters are loaded, the behavior of the entire ISAC system is set. The system will then optimize according to the parameters specified in the optimization objectives. The defined trade-off point leads to an optimal synesthetic mode: if When the value approaches 1, the system will operate in communication-priority mode, where the transmission precoding matrix and complex reflection coefficient matrix primarily focus on maximizing the communication rate; if When the values ​​approach 0, the system will operate in a perception-first mode, where the transmission precoding matrix and complex reflection coefficient matrix primarily focus on improving perception quality; if By taking the middle value, the system will operate in a state where communication and sensing performance are optimally balanced.

[0081] It should be noted that the first, second, and third numerical ranges mentioned above are mainly set by technical personnel based on system configuration, and therefore are not specifically limited in this embodiment of the invention. For example, the first numerical range can be 0.2±0.05, the second numerical range can be 0.8±0.05, and the third numerical range can be 0.5±0.05.

[0082] In practical applications, Table 1 shows the effect of the proposed method on the synesthesia adjustment coefficient. The signal-to-noise ratio (SNR) corresponds to the achievable total communication rate and the achievable sensing beam gain for each type of ISAC system under different settings. Consistent with the aforementioned principles, a higher SNR value improves both the total communication rate and the sensing beam gain for all three modes. For the sensing-dominant mode, the total communication rate is higher than the sensing beam gain corresponding to the communication-dominant mode at different SNRs; ​​for the communication-dominant mode, the sensing beam gain is higher than the total communication rate corresponding to the sensing-dominant mode at different SNRs; ​​for the sensing-balanced mode, a trade-off occurs between the two parameters.

[0083] Table 1 Examples of Experimental Results

[0084] Based on the above description, this application proposes an optimization method for an ISAC system with tunable inductive modality. This method first designs a hybrid RIS model composed of active and passive units, optimizes the index importance fraction matrix and decomposes the complex reflection coefficients, and constructs an integrated sensing and communication system that balances the amplification of active units and the low power consumption of passive units using inductive adjustment coefficients. For the multiple-input multiple-output communication system model, the Riemann-Hesse algorithm is used to adaptively optimize the complex reflection coefficient matrix, and the Lagrange multiplier method is used to solve for the optimal solution of the transmission precoding matrix. The algorithm processes for both types are derived. Compared with baseline methods such as passive RIS only, active RIS only, and random phase RIS, the proposed method can achieve tunable inductive modality in the ISAC system.

[0085] To further illustrate the performance advantages of the present invention, the proposed optimization method is compared below with random phase RIS, source-only RIS, passive-only RIS and non-RIS on the ISAC system.

[0086] The experimental environment consisted of an Intel i5-10300H CPU, 16 GB RAM, Windows 10, and MATLAB 2022a. The experimental parameters are shown in Table 2.

[0087] Table 2 Examples of Experimental Parameter Settings

[0088] Figure 5 The communication-sensing curves of the above methods were compared. When the communication rate is 5... At this time, the proposed method maintains approximately 4 dBi of sensing beam gain, while other baseline methods (such as random phase RIS, active RIS only, passive RIS only, and non-RIS) all show a significant decrease. When the total communication rate is greater than 5 At that time, the proposed method produced smoother and more stable boundaries, with high communication rates and sensing gain. It should be noted that... Figure 5 The horizontal axis represents the communication rate, and the vertical axis represents the sensing beam pattern gain. The proposed scheme is the scheme proposed in this invention. Random-phase RIS system is a random-phase RIS system, Active-only RIS system is an active-only RIS system, Passive-only RIS system is a passive-only RIS system, and Non-RIS system is a non-RIS system.

[0089] Figure 6 Communication-sensing curves were compared for different numbers of active RIS reflector elements. The proposed solution exhibits higher sensing gain at a given rate as the number of active RIS reflector elements increases.

[0090] Figure 7 The communication-sensing curves at different signal-to-noise ratios were compared. It is evident that as the transmit signal-to-noise ratio (SNR) increases, the proposed scheme can generate higher sensing beam gain for a given SNR.

[0091] The above description is merely a preferred embodiment of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention is not limited to the specific combination of the above-described technical features, but also includes other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in this invention.

Claims

1. A method for optimizing a synesthetic modality-tunable ISAC system, characterized in that, include: An ISAC system model is constructed, comprising a base station, a user, a sensing target, and a RIS-based system; wherein the RIS comprises reflective units with different reflective characteristics; the reflective units with different reflective characteristics include: active reflective units and passive reflective units; Based on the status information of the base station to RIS channel and the status information of the RIS to user channel, the importance score of each RIS unit is calculated, and the configuration strategy of active reflection unit and passive reflection unit is determined based on the importance score to form a hybrid RIS structure. Wherein, the importance score of the RIS unit ( i , j ) can be calculated by the following formula (1): Equation (1); in, For the RIS to user channel in the RIS unit ( i , j The coefficient of ) is the state information of the RIS to user channel; For the base station to RIS channel in the RIS unit ( i , j The coefficient of ) is the state information of the base station to RIS channel; For RIS unit ( i , j Importance score; The weighted sum of the communication rate and sensing beam gain of the ISAC system is used as the objective optimization function of the hybrid RIS unit, and the reflection beam parameters of the RIS and the transmission beam parameters of the base station are jointly optimized under the base station transmit power constraint. The optimized reflected beam parameters and transmitted beam parameters are loaded into the ISAC system corresponding to the system model to drive the ISAC system to operate in the optimal synesthetic mode; The configuration strategy for active and passive reflective units is determined based on the importance score, including: The importance scores of each RIS unit are sorted in descending order to obtain a priority index; Based on the priority index and the preset number of active reflection units, the RIS units corresponding to the index of the first preset number of active reflection units in the priority index vector are configured as active reflection units; the remaining RIS units in the priority index vector P are configured as passive reflection units.

2. The optimization method for a tunable inter-sensory modality ISAC system according to claim 1, characterized in that, The weighted sum of the communication rate and sensing beam gain of the ISAC system is used as the objective optimization function of the hybrid RIS unit, including: Obtain the total communication rate of the ISAC system and the sensing beam gain in the ISAC system in the spatial direction associated with the sensing target. By introducing a sensing adjustment coefficient and combining it with a weighted calculation of the total communication rate and the sensing beam gain, a target optimization function is constructed.

3. The optimization method for a tunable ISAC system according to claim 2, characterized in that, Obtaining the total communication rate of the ISAC system includes: Based on the direct channel from base station to user, the base station to RIS channel, and the RIS to user channel, an equivalent communication channel model including direct links and reflection links is constructed. Based on the equivalent communication channel model and the base station transmitted signal, the user terminal received signal model is determined; Based on the received signal model, the signal-to-interference-plus-noise ratio (SNR) of each user is calculated, and the total communication rate of the system is calculated based on the SNR of each user.

4. The optimization method for a tunable ISAC system according to claim 3, characterized in that, The reflection beam parameters are the complex reflection coefficient matrix of the RIS; the transmission beam parameters are the transmission precoding matrix of the base station. Under base station transmit power constraints, the joint optimization of the reflection beam parameters of the RIS and the transmit beam parameters of the base station includes: When the transmission precoding matrix is ​​fixed Initialize the complex reflection coefficient matrix and set its elements to satisfy the unit modulus constraint; Calculate the Euclidean gradient of the transformed objective function with respect to the complex reflection coefficient matrix; The Euclidean gradient is projected onto the tangent space of the unit circular manifold to obtain the Riemann gradient; The update amount of the complex reflection coefficient matrix is ​​calculated based on the Riemann gradient and the Riemann-Hesse operator-determined Riemann-Newton direction. The optimization step size is adaptively adjusted based on the norm of the Riemann gradient and the norm of the Riemann-Hesse operator. The complex reflection coefficient matrix is ​​iteratively updated according to the update amount and the step size until the convergence condition is met, so as to output the jointly optimized complex reflection coefficient matrix.

5. The optimization method for a tunable ISAC system according to claim 4, characterized in that, Under base station transmit power constraints, the joint optimization of the reflection beam parameters of the RIS and the transmit beam parameters of the base station includes: When the complex reflection coefficient matrix is ​​fixed Lagrange multipliers are introduced, and an optimization objective equation for the transmit precoding matrix is ​​constructed based on the mean square error of the total communication rate and the objective optimization function. Taking the partial derivative of the Lagrange function with respect to the transmit precoding matrix and setting it to zero yields the optimal solution expression for the transmit precoding matrix with respect to the Lagrange multipliers. Substitute the optimal solution expression into the optimization objective equation to construct the power equation for the Lagrange multiplier; The power equation is solved by an iterative algorithm, the values ​​of the Lagrange multipliers are updated, and the updated Lagrange multipliers are substituted into the optimal solution expression to obtain the jointly optimized transmit precoding matrix.

6. The optimization method for a tunable ISAC system according to claim 4, characterized in that, The complex reflection coefficient matrix integrates the phase shift component shared by the active and passive reflection units, as well as the amplitude gain component possessed by the active reflection unit.

7. The optimization method for a tunable ISAC system according to claim 3, characterized in that, The inductive adjustment coefficient is used to adjust the relative weights of the total communication rate and the sensing beam gain; When the value of the sensing adjustment coefficient is within the first numerical range, the ISAC system is configured as a sensing-first mode with sensing performance as the primary consideration. When the value of the induction adjustment coefficient is within the second numerical range, the ISAC system is configured as a communication priority mode that prioritizes communication performance. When the value of the synesthesia adjustment coefficient is in the third numerical range, the ISAC system is configured as a balanced mode with balanced synesthesia performance.

Images