A method and apparatus for multi-base station cooperative sensing resource optimization based on generative AI

Abstract

This invention provides a method and apparatus for multi-base station cooperative sensing resource optimization based on generative AI. The method includes: acquiring first communication sensing-related information of a first multi-base station cooperative integrated communication sensing (ISAC) system; inputting the first communication sensing-related information into a trained AI resource optimization model to obtain first resource allocation information and / or first sensing mode of the first multi-base station cooperative ISAC system output by the AI ​​resource optimization model. This invention overcomes the limitations of traditional fixed resource allocation, can generalize from existing scenario information to new scenario requirements, dynamically adjust the sensing mode and resource allocation strategy, has a certain degree of robustness, and can adapt to dynamic environmental changes, solving the problems of low resource utilization efficiency and limited communication sensing performance caused by insufficient system flexibility.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of communication and sensing integration, and particularly relates to a multi-base station cooperative sensing resource optimization method based on generative AI. BACKGROUND

[0002] In future mobile communication networks, Integrated Sensing and Communication (ISAC) realizes the functions of communication and environmental sensing through the same platform, and is widely applied to scenes such as unmanned aerial vehicle monitoring, automatic driving, intelligent transportation systems, etc., significantly improving the overall efficiency of the network. For example, a base station can simultaneously provide communication services and perform target detection and tracking. With the complication of application scenarios, a single base station is difficult to support wide-area and multi-target sensing requirements due to limited coverage and resources. Therefore, the multi-base station cooperative ISAC system works jointly through multiple base stations to optimize resource allocation and coordination, and improve sensing accuracy and communication quality. Through methods such as joint beamforming, power allocation and resource scheduling, the base stations can realize cooperative optimization, expand the sensing coverage, balance the communication and sensing performance, and provide support for efficient operation in complex dynamic scenarios.

[0003] However, the existing multi-base station cooperative ISAC system has limitations in resource allocation and system optimization, mainly reflected in that the fixed sensing mode limits the flexibility of the system and leads to low resource efficiency.

[0004] Currently, the resource allocation optimization scheme for the multi-base station cooperative ISAC system usually adopts an active sensing or passive sensing mode (i.e., the fixed sensing mode is an active sensing mode or a passive sensing mode), and interference modeling and resource optimization are based on this. However, under different sensing modes, the signal propagation path, channel state and interference characteristics differ significantly. For example, in the active sensing mode, the base station continuously transmits sensing signals, resulting in large sensing echo power, and the weak echo signal of the adjacent base station is often regarded as interference; in the passive sensing mode, the system relies on low-power echo signals in the environment, and the channel state and interference characteristics are more susceptible to external environmental changes. Therefore, using a fixed sensing mode cannot adapt to the real-time changes of dynamic environments and task requirements according to the channel state and interference level. In high-interference or resource-limited scenarios, a fixed sensing mode may lead to unreasonable resource allocation, resulting in resource waste or insufficient sensing performance. In multi-target task scenarios, a fixed sensing mode also cannot combine the importance of different targets and the real-time changes of tasks to dynamically adjust resource scheduling and sensing strategies, thereby limiting the cooperative optimization effect of communication and sensing tasks and affecting the overall performance of the system.

[0005] In summary, the fixed sensing mode adopted in the existing multi-base station cooperative ISAC system has the problems of low resource utilization efficiency and limited communication sensing performance due to the difficulty in adapting to dynamic environment changes and the insufficient flexibility of the system. SUMMARY

[0006] The present application provides a multi-base station cooperative sensing resource optimization method and device based on generative AI to solve the problems of low resource utilization efficiency and limited communication sensing performance due to the difficulty in adapting to dynamic environment changes and the insufficient flexibility of the system in the existing multi-base station cooperative ISAC system.

[0007] To solve the above technical problems, the embodiments of the present application provide the following technical solutions:

[0008] In a first aspect, the embodiments of the present application provide a multi-base station cooperative sensing resource optimization method based on generative AI, comprising:

[0009] obtaining first communication sensing related information of a first multi-base station cooperative communication sensing integrated ISAC system;

[0010] inputting the first communication sensing related information into a trained artificial intelligence AI resource optimization model to obtain first resource allocation information of the first multi-base station cooperative ISAC system and / or first sensing mode of the first multi-base station cooperative ISAC system output by the AI resource optimization model;

[0011] The AI resource optimization model is used to indicate at least one of the following:

[0012] the correspondence between the communication sensing related information and the resource allocation information;

[0013] the correspondence between the communication sensing related information and the sensing mode.

[0014] Optionally, the method further comprises:

[0015] obtaining second communication sensing related information, second sensing mode and second resource allocation information of a second multi-base station cooperative ISAC system;

[0016] training using the second communication sensing related information and the second sensing mode, and / or using the second communication sensing related information and the second resource allocation information to obtain the AI resource optimization model.

[0017] Optionally, the communication sensing related information comprises at least one of the following:

[0018] a base station;

[0019] a communication user;

[0020] a sensing target;

[0021] resource constraint information;

[0022] communication sensing index information.

[0023] Optionally, the sensing mode comprises at least one of:

[0024] an active sensing mode;

[0025] a cooperative sensing mode.

[0026] Optionally, the second resource allocation information of the second multi-base-station cooperative ISAC system is obtained, comprising:

[0027] obtaining, in the active sensing mode, a downlink communication signal received by a communication user from a base station in the second multi-base-station cooperative ISAC system, a reflected echo signal of a sensing target received by the base station, and beam power allocation information of the base station;

[0028] generating, according to the reflected echo signal of the sensing target received by the base station, a first optimization objective for resource allocation optimization in the active sensing mode;

[0029] generating, according to the downlink communication signal received by the communication user from the base station and the beam power allocation information of the base station, a first optimization constraint condition for resource allocation optimization in the active sensing mode;

[0030] performing resource allocation optimization on the second multi-base-station cooperative ISAC system in the active sensing mode according to the first optimization objective and the first optimization constraint condition, to obtain the second resource allocation information in the active sensing mode.

[0031] Optionally, the first optimization objective for resource allocation optimization in the active sensing mode is generated according to the reflected echo signal of the sensing target received by the base station, comprising:

[0032] performing separation filtering processing on the reflected echo signal of the sensing target received by the base station to obtain a first filtered echo signal;

[0033] generating a Cramer-Rao lower bound (CRLB) of a sensing performance of the sensing target according to the first filtered echo signal;

[0034] obtaining a first global positioning performance of the sensing target according to the CRLB of the sensing target, a reflected echo azimuth angle (AoA) of the sensing target relative to the base station, and a reflected echo zenith angle (ZoA) of the sensing target relative to the base station;

[0035] generating the first optimization objective according to the first global positioning performance.

[0036] Optionally, the first optimization constraint condition for resource allocation optimization in the active sensing mode is generated according to a downlink communication signal received by the communication user from a base station and beam power allocation information of the base station, and the first optimization constraint condition comprises:

[0037] The first signal-to-interference ratio (SINR) received by the communication user is obtained according to the downlink communication signal received by the communication user from the base station.

[0038] The first condition that the first SINR is greater than or equal to a first preset SINR threshold is set.

[0039] The transmission power of the base station and the number of beams supported by the base station are obtained according to the power allocation information of the base station and the beam power allocation information of the base station.

[0040] The second condition that the transmission power of the base station is less than or equal to a first preset power threshold is set.

[0041] The third condition that the number of beams supported by the base station is less than or equal to a first preset number is set.

[0042] The fourth condition is generated according to a preset beam selection matrix.

[0043] The first optimization constraint condition is obtained according to the first condition, the second condition, the third condition and the fourth condition.

[0044] Optionally, the resource allocation optimization in the second multi-base station cooperative ISAC system in the active sensing mode is performed according to the first optimization target and the first optimization constraint condition, and the second resource allocation information in the active sensing mode is obtained, which comprises:

[0045] The fourth condition in the first optimization constraint condition is subjected to approximate convex relaxation processing to obtain a fifth condition.

[0046] The first optimization target is decomposed into a first sub-target and a second sub-target by using an initial beam selection matrix.

[0047] The numerator information and the denominator information in the first SINR received by the communication user are obtained according to the initial beam selection matrix.

[0048] The first condition in the first optimization constraint condition is converted into a sixth condition according to the numerator information and the denominator information.

[0049] The first optimization problem is generated according to the first sub-target, the sixth condition and the second condition in the first optimization constraint condition.

[0050] generate a second optimization problem according to the second sub-target, third and fifth conditions in the first and second optimization constraint conditions;

[0051] obtain the second resource allocation information in the active sensing mode by alternately solving the first and second optimization problems.

[0052] Optionally, the obtaining of the second resource allocation information of the second multi-base-station cooperative ISAC system includes:

[0053] obtaining, in the cooperative sensing mode, a transmission signal of a first base station in the second multi-base-station cooperative ISAC system, a reception signal of a first communication user in communication with the first base station, and a reflection signal received by the first base station;

[0054] obtaining a second SINR of the first communication user according to the transmission signal of the first base station and the reception signal of the first communication user;

[0055] obtaining a third SINR of a radar output signal of a sensing target at the first base station according to the reflection signal received by the first base station;

[0056] generating a second optimization target according to the second SINR;

[0057] generating a second optimization constraint condition for resource allocation optimization in the cooperative sensing mode according to the third SINR and power budget information of the first base station;

[0058] performing resource allocation optimization on the second multi-base-station cooperative ISAC system in the cooperative sensing mode according to the second optimization target and the second optimization constraint condition, to obtain the second resource allocation information in the cooperative sensing mode.

[0059] Optionally, the generating of the second optimization constraint condition for resource allocation optimization in the cooperative sensing mode according to the third SINR and the power budget information of the first base station includes:

[0060] taking the third SINR being greater than or equal to a second preset SINR threshold as a seventh condition;

[0061] taking the transmission power of the first base station being less than or equal to a second preset power threshold indicated by the power budget information of the first base station as an eighth condition;

[0062] generating a ninth condition according to a preset target scheduling matrix;

[0063] obtaining the second optimization constraint condition according to the seventh, eighth and ninth conditions.

[0064] Optionally, according to the second optimization objective and the second optimization constraint condition, resource allocation optimization is performed on the second multi-base station cooperative ISAC system in the cooperative sensing mode, and second resource allocation information in the cooperative sensing mode is obtained, including:

[0065] According to the transmit beamforming matrix and the target scheduling matrix, the second optimization objective is converted into a third optimization problem;

[0066] The third optimization problem is solved by using generalized Rayleigh entropy, and an optimized receive beamforming matrix is obtained;

[0067] The second optimization objective is converted into a third sub-objective by using the Dinkelbach method;

[0068] According to the third sub-objective and sixth and seventh conditions in the second optimization constraint condition, a fourth optimization problem is obtained;

[0069] The fourth optimization problem is solved by using the Lagrange dual method, and an optimized transmit beamforming matrix is obtained;

[0070] According to the optimized receive beamforming matrix, the optimized transmit beamforming matrix, and the second optimization objective, a target scheduling matrix is optimized, and the second resource allocation information in the cooperative sensing mode is obtained.

[0071] Optionally, the AI resource optimization model is obtained by training using the second communication sensing related information and the second sensing mode, and / or by training using the second communication sensing related information and the second resource allocation information, including:

[0072] The second communication sensing related information is feature extracted, and first extracted features are obtained;

[0073] The first extracted features are feature clustered, and a plurality of feature domains are obtained;

[0074] The extracted features in the same feature domain are weighted processed, and a first domain feature representation is obtained;

[0075] The extracted features in different feature domains are feature extracted and fused, and a second domain feature representation is obtained;

[0076] According to the first domain feature representation and the second domain feature representation, a target feature representation is obtained;

[0077] The AI resource optimization model is obtained by training using the target feature representation and the second sensing mode, and / or by training using the target feature representation and the second resource allocation information.

[0078] Optionally, the AI resource optimization model is obtained by training using the target feature representation and the second resource allocation information.

[0079] The first training resource allocation information is obtained according to the target feature representation by using a preset optimization manner.

[0080] The second training resource allocation information is obtained according to the target feature representation by using a preset initial AI model.

[0081] The loss function is obtained according to the first training resource allocation information and the second training resource allocation information.

[0082] The AI resource optimization model is obtained by adjusting model parameters in the initial AI model to reduce the loss function.

[0083] In a second aspect, an embodiment of the present application further provides a multi-base station cooperative sensing resource optimization device based on generative AI, comprising:

[0084] A first acquisition module is configured to acquire first communication-sensing related information of a first multi-base station cooperative integrated sensing and communication (ISAC) system.

[0085] A first processing module is configured to input the first communication-sensing related information into a trained artificial intelligence (AI) resource optimization model to obtain first resource allocation information of the first multi-base station cooperative ISAC system and / or a first sensing mode of the first multi-base station cooperative ISAC system output by the AI resource optimization model.

[0086] The AI resource optimization model is configured to indicate at least one of the following:

[0087] A correspondence between the communication-sensing related information and the resource allocation information;

[0088] A correspondence between the communication-sensing related information and the sensing mode.

[0089] In a third aspect, an embodiment of the present application further provides a multi-base station cooperative sensing resource optimization device based on generative AI, comprising a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program is executed by the processor to implement the steps in the multi-base station cooperative sensing resource optimization method based on generative AI according to any one of the first aspect.

[0090] In a fourth aspect, an embodiment of the present application further provides a readable storage medium, wherein the readable storage medium stores a program, and the program is executed by a processor to implement the steps in the multi-base station cooperative sensing resource optimization method based on generative AI according to any one of the first aspect.

[0091] In a fifth aspect, the embodiments of the present application also provide a computer program product comprising computer instructions for implementing the steps of the method for optimizing sensing resources of a multi-base station cooperative system based on generative AI according to any one of the first aspect when executed by a processor.

[0092] The present application has the following beneficial effects:

[0093] The method for optimizing sensing resources of a multi-base station cooperative system based on generative AI provided by the present application trains an AI resource optimization model, obtains first resource allocation information of a first multi-base station cooperative ISAC system output by the AI resource optimization model by using the correspondence between the communication and sensing related information and the resource allocation information indicated in the AI resource optimization model, and / or obtains a first sensing mode of the first multi-base station cooperative ISAC system output by the AI resource optimization model by using the correspondence between the communication and sensing related information and the sensing mode indicated in the AI resource optimization model, thereby breaking through the limitations of traditional fixed resource allocation, generalizing to new scene requirements according to existing scene information, dynamically adjusting the sensing mode and resource allocation strategy, having certain robustness, and being able to adapt to dynamic environmental changes, solve the problems of low resource utilization efficiency and limited communication and sensing performance caused by the lack of flexibility of the system. BRIEF DESCRIPTION OF DRAWINGS

[0094] Figure 1 A flowchart of the method for optimizing sensing resources of a multi-base station cooperative system based on generative AI provided by the embodiments of the present application is shown;

[0095] Figure 2 A schematic diagram of multi-base station cooperative communication and sensing in an active sensing mode provided by the embodiments of the present application is shown;

[0096] Figure 3 A flowchart of obtaining resource allocation information in an active sensing mode provided by the embodiments of the present application is shown;

[0097] Figure 4 A flowchart of obtaining resource allocation information in a cooperative sensing mode provided by the embodiments of the present application is shown;

[0098] Figure 5 A schematic diagram of multi-base station cooperative communication and sensing provided by the embodiments of the present application is shown;

[0099] Figure 6 A flowchart of the training process of the AI resource optimization model provided by the embodiments of the present application is shown;

[0100] Figure 7Fig. 1 shows a structural schematic diagram of a device for optimizing a sensing resource of a multi-base station cooperation based on a generative AI according to an embodiment of the present application.

[0101] Figure 8 Fig. 1 shows a structural schematic diagram of a device for optimizing a sensing resource of a multi-base station cooperation based on a generative AI according to an embodiment of the present application. DETAILED DESCRIPTION

[0102] To make the technical problems, technical solutions and advantages of the present application clearer, specific embodiments will be described in detail below with reference to the accompanying drawings. In the following description, specific details such as specific configurations and components are provided only to help a comprehensive understanding of embodiments of the present application. Therefore, it should be apparent to those skilled in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. In addition, descriptions of known functions and configurations have been omitted for clarity and conciseness.

[0103] It should be understood that the phrase "one embodiment" or "an embodiment" appearing throughout the specification means that a particular feature, structure, or characteristic related to an embodiment is included in at least one embodiment of the present application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. In addition, these particular features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0104] In various embodiments of the present application, it should be understood that the size of the serial number of the following processes does not mean the order of execution, and the execution order of the processes should be determined by their functions and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0105] In the embodiments provided in the present application, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined according to A. However, it should also be understood that determining B according to A does not mean that B is determined only according to A, but B can also be determined according to A and / or other information.

[0106] In the embodiments of the present application, the term "and / or" describes the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B can mean that A exists alone, A and B exist together, and B exists alone. The character " / " generally represents that the associated objects before and after it are in an "or" relationship.

[0107] To solve the problem that the fixed sensing mode adopted in the existing multi-base station cooperation ISAC system has difficulty in adapting to dynamic environment changes, the system lacks flexibility, resulting in low resource utilization efficiency and limited communication sensing performance, embodiments of the present application provide a method and device for optimizing a sensing resource of a multi-base station cooperation based on a generative AI.

[0108] As Figure 1 shown, the embodiment of the present application provides a multi-base station cooperative sensing resource optimization method based on generative AI, comprising:

[0109] Step 101: Obtain the first communication sensing related information of the first multi-base station cooperative integrated sensing and communication (ISAC) system.

[0110] The first multi-base station cooperative ISAC system can be any multi-base station cooperative ISAC system. The first multi-base station cooperative ISAC system is the system to be subjected to sensing resource optimization allocation.

[0111] The communication sensing related information includes at least one of the following:

[0112] The base station, wherein the communication sensing related information further includes location information, quantity information, etc. of the base station;

[0113] The communication user, wherein the communication sensing related information further includes location information, quantity information, type information, etc. of the communication user, wherein the type information of the communication user includes vehicles and user equipment (UE);

[0114] The sensing target, wherein the communication sensing related information further includes location information, quantity information, etc. of the sensing target, and in this embodiment, the sensing target is taken as an example of a drone;

[0115] The resource constraint information, wherein the resource constraint information includes at least one of the following: SINR threshold, maximum number of beams that the base station can support, maximum transmit power that the base station can support;

[0116] The communication sensing index information, wherein the communication sensing index information includes at least one of the following: downlink communication signal received by the communication user from the base station, echo signal reflected by the sensing target received by the base station, power allocation information of the base station, beam power allocation information of the base station, transmit signal of the base station, received signal of the communication user in communication with the base station, and interference of the sensing target received by the base station.

[0117] That is, in this step, the first communication sensing related information in the first multi-base station cooperative ISAC system to be subjected to resource optimization allocation is obtained, and the first sensing communication related information is used for subsequent resource optimization allocation.

[0118] Step 102: input the first communication and perception related information into a trained artificial intelligence (AI) resource optimization model to obtain first resource allocation information of the first multi-base station cooperative ISAC system and / or first perception mode of the first multi-base station cooperative ISAC system output by the AI resource optimization model;

[0119] The AI resource optimization model is used to indicate at least one of the following:

[0120] A correspondence between the communication and perception related information and the resource allocation information;

[0121] A correspondence between the communication and perception related information and the perception mode.

[0122] That is, in this step, the AI resource optimization model is first trained using historical data, and the AI resource optimization model stores a correspondence between the communication and perception related information and the resource allocation information and / or a correspondence between the communication and perception related information and the perception mode.

[0123] The first communication and perception related information is input into the trained AI resource optimization model, and the first resource allocation information of the first multi-base station cooperative ISAC system output by the AI resource optimization model is obtained using the correspondence between the communication and perception related information and the resource allocation information, and / or the first perception mode of the first multi-base station cooperative ISAC system output by the AI resource optimization model is obtained using the correspondence between the communication and perception related information and the perception mode, i.e., the perception mode and resource allocation are determined by the model. This breaks the limitations of traditional fixed resource allocation, can generalize to new scene requirements according to existing scene information, dynamically adjusts the perception mode and resource allocation strategy, has a certain robustness, and can adapt to dynamic environmental changes, solve the problem of low resource utilization efficiency and limited communication and perception performance caused by the lack of system flexibility.

[0124] The perception mode includes at least one of the following:

[0125] Active perception mode, wherein the active perception mode is that in the multi-base station cooperative ISAC system, a same base station transmits a perception signal and receives a reflected echo signal of a perception target;

[0126] Cooperative perception mode, wherein the cooperative perception mode is that in the multi-base station cooperative ISAC system, one base station transmits a perception signal and another different base station receives a reflected echo signal of a perception target.

[0127] Further, the communication and perception related information includes:

[0128] Communication sensing related information in the active sensing mode;

[0129] Communication sensing related information in the cooperative sensing mode.

[0130] It should be noted that the active sensing mode can also be referred to as A-to-A mode, in which the base station is responsible for both transmitting signals and receiving echoes, achieving active sensing. This mode is suitable for high-precision sensing tasks within a single base station range. The key features of this mode are: signal modeling: the base station transmits sensing signals and receives echo signals reflected from the target. In this mode, the base station's signal modeling is based on the Cramer-Rao lower bound (CRLB), and the sensing accuracy is optimized by minimizing the CRLB; beam and power allocation: the base station performs beamforming and power allocation while transmitting signals to ensure accurate sensing of the target, and through CRLB optimization, the base station can more efficiently utilize resources in the sensing task; application scenarios: suitable for tasks that require high-precision sensing, such as precise positioning and tracking of unmanned aerial vehicle targets. In this mode, the base station can independently complete the sensing task without relying on the cooperation of other base stations.

[0131] The cooperative sensing mode can also be referred to as A-to-B mode, in which multiple base stations cooperate to complete the sensing task. Base station A is responsible for transmitting sensing signals, and base station B receives and processes the echo signals to achieve sensing of the target. The key features of this mode are: signal modeling: in the A-to-B mode, base station A transmits communication and sensing signals, and base station B is responsible for receiving the sensing signals reflected from the target. Through multi-base station cooperation, the sensing coverage is expanded, and the quality of the sensing signals is improved; beam and power allocation: due to the cooperation of multiple base stations, the system needs to reasonably allocate power and beam direction among multiple base stations to avoid mutual interference between base stations and ensure the overall performance of the system is optimal. Beamforming and power allocation are more complex in this mode, requiring coordination of resources among multiple base stations; application scenarios: suitable for scenarios with wider coverage and higher target complexity, such as unmanned aerial vehicle clusters or large-scale monitoring tasks. The cooperation of multiple base stations can improve the sensing accuracy and anti-interference capability of the system for targets.

[0132] The training process of the AI resource optimization model provided by the embodiments of the present application will be described in detail below:

[0133] In an embodiment, the method further comprises:

[0134] Obtaining second communication sensing related information, a second sensing mode, and second resource allocation information of a second multi-base station cooperative ISAC system;

[0135] The number of the second multi-base station cooperative ISAC systems can be one or more, and the systems are used for model training.

[0136] It should be noted that the parameters in the second communication awareness related information and the first communication awareness related information are the same, and the specific values thereof are different.

[0137] The second resource allocation information is obtained according to the second notification awareness information, and the second awareness mode is set according to the application scenario of the multi-base station cooperative ISAC system.

[0138] The second communication awareness related information and the second awareness mode are used for training, and / or the second communication awareness related information and the second resource allocation information are used for training to obtain the AI resource optimization model.

[0139] Specifically, the second communication awareness related information and the second awareness mode are used for training to obtain a corresponding relationship between the communication awareness related information and the awareness mode in the AI resource optimization model, and the second communication awareness related information and the second resource allocation information are used for training to obtain a corresponding relationship between the communication awareness related information and the resource allocation information in the AI resource optimization model.

[0140] In an optional embodiment, the process of obtaining the communication awareness related information in the active awareness mode is specifically described.

[0141] In the active awareness mode, the base station is equipped with a massive multiple-input multiple-output (Massive MIMO, mMIMO) uniform planar array (Uniform Planar Array, UPA), and has the ability to perform awareness through the ISAC technology. The base station uses independent transmitting and receiving antennas, and can receive the awareness echo signal while maintaining the downlink communication.

[0142] For this mode, the embodiments of the application model the relationship between the downlink communication signal and the active awareness echo signal and the base station transmitting beam in the multi-base station cooperative scenario, and derive the signal-to-interference-and-noise ratio (SINR) expression of the ground communication user and the Cramer-Rao lower bound (CRLB) of the aerial unmanned aerial vehicle awareness performance. By minimizing the CRLB, the awareness accuracy is improved in the case where the position of the unmanned aerial vehicle (i.e., the awareness target) is approximately known.

[0143] That is, in the optional embodiment, the second resource allocation information of the second multi-base station cooperative ISAC system is obtained, including:

[0144] The downlink communication signal received by the communication user from the base station in the second multi-base station cooperative ISAC system, the echo signal reflected by the awareness target and received by the base station, the power allocation information of the base station, and the beam power allocation information of the base station in the active awareness mode are obtained.

[0145] Specifically, in the active sensing mode, all base stations (ISAC BSs) provide sensing and communication services simultaneously in a multi-beam mode. In terms of communication services, ground communication users, such as vehicles, user equipment (UEs), with different quality of service (QoS) requirements for downlink communication are considered. In terms of sensing services, each ISAC BS receives the echo of its own transmitted signals reflected by unmanned aerial vehicles (UAVs). Through analysis of the reflected signals, each ISAC BS uploads its positioning results to a fusion center for data integration. Similar to ground communication users, UAVs exhibit different threat levels in different areas, thereby having different sensing QoS requirements.

[0146] In the active sensing mode, please refer to Figure 2 , Figure 2 FIG. 1 is a schematic diagram of a multi-base station cooperative communication and sensing mode in the active sensing mode, in which three adjacent sectors of three adjacent BSs are regarded as a cooperative communication and sensing integrated unit, i.e., a cooperative ISAC unit (areas with the same color in the figure), Figure 2 The non-cooperative UAV shown in FIG. 2 is a sensing target. In the cooperative ISAC unit, the sensing targets include K UAVs, and the communication users include M single-antenna UEs. Each base station is equipped with N t transmit antennas and N r receive antennas, which are both uniform planar arrays (UPAs). The base stations transmit multiple beams to the UAVs or UEs through beamforming (BF) technology. Each UAV can reflect the beams from one or more BSs, and after receiving the echo of the UAV by each base station, the base stations perform sensing signal level fusion at one base station through data interaction between the base stations to obtain the sensing result; meanwhile, the base stations transmit multiple beams to perform downlink data transmission with the UEs.

[0147] The baseband signal of the nth base station is defined as:

[0148]

[0149] where J = K + M represents the number of UAVs and UEs, and the ISAC signal after beamforming is:

[0150]

[0151] where represents a transmit beamforming matrix, represents a beam selection vector, and a binary variable g j,n takes 1 when the base station n allocates a beam pointing to the sensing target j, and takes 0 when the base station n does not allocate a beam pointing to the sensing target j.

[0152] For the communication UE, the downlink communication signal received by the mth UE from the nth base station is:

[0153]

[0154]

[0155] where N represents the total number of cooperative base stations, represents the antenna gain, M represents the set of UEs, and m,n represents the path loss between the mth UE and the nth base station, p m,n represents the beam power allocated by the nth base station to the mth UE, p k,n represents the beam power allocated by the nth base station to the kth UAV, p m,n and p k,n are beam power allocation information of the base station (the beam power allocation information of the base station includes p m,n and p k,n ), and θ m,n represent the azimuth angle and the elevation angle of the mth UE relative to the nth base station, respectively, represents additive white Gaussian noise with variance represents the steering vector of the transmitting antenna, and the calculation method is:

[0156]

[0157] where N y and N z represent the number of rows and columns of the UPA antenna, respectively.

[0158] The echo signal received by the nth base station from all the UAVs is:

[0159]

[0160] where represents the antenna gain, β k,n , μ k,n and τ k,n represent the reflection coefficient, the Doppler frequency, and the time delay of the kth UAV relative to the nth base station, respectively, is Gaussian white noise with variance and θ k,n represent the azimuth angle of arrival (AoA) and the zenith angle of arrival (ZoA) of the echo signal of the kth UAV relative to the nth base station, respectively, represents the steering vector of the receiving antenna.​​

[0161] And the beamformer is designed as follows:

[0162]

[0163] wherein and denote the estimated elevation and azimuth angles.

[0164] According to the echo signal of the sensing target reflected received by the base station, a first optimization objective for resource allocation optimization in the active sensing mode is generated, that is, the first optimization objective is generated according to the formula (4) described above.

[0165] According to the downlink communication signal received by the base station from the communication user and the beam power allocation information of the base station, a first optimization constraint condition for resource allocation optimization in the active sensing mode is generated. According to

[0166] The first optimization objective and the first optimization constraint condition obtain the modeled optimization problem.

[0167] According to the first optimization objective and the first optimization constraint condition, resource allocation optimization is performed on the second multi-base station cooperative ISAC system in the active sensing mode, and second resource allocation information in the active sensing mode is obtained.

[0168] Optionally, according to the echo signal of the sensing target reflected received by the base station, the first optimization objective for resource allocation optimization in the active sensing mode is generated, including:

[0169] The echo signal of the sensing target reflected received by the base station is separated and filtered to obtain a first filtered echo signal, specifically:

[0170] According to formula (4), the echoes from different UAVs can be separated by a space division filter, and the filtered echo is:

[0171]

[0172] wherein w k,n denotes the receiving beamforming vector of the nth base station in the direction of the kth UAV, denotes an additive white Gaussian noise.

[0173] According to the first filtered echo signal, a Cramer-Rao lower bound (CRLB) of the sensing performance of the sensing target is generated.

[0174] Specifically, the embodiments of the present application use CRLB to evaluate the performance of UAV multi-base station cooperative sensing and derive its expression. On this basis, the multi-base station cooperative ISAC system beam and power allocation optimization problem is proposed, and then the multi-base station collaborative beam and power allocation (MCBPA) algorithm is proposed to solve this problem.

[0175] The derivation process of the UAV's sensing CRLB under multi-base station cooperation is as follows:

[0176] When the UAV position is approximately known as prior information, the estimation performance can be improved by minimizing the CRLB. As the lower bound of the variance of various unbiased estimators, CRLB is the inverse of the fisher information matrix (FIM), which can be written as:

[0177]

[0178] where is the true value of the UAV distance, azimuth angle and pitch angle, J k is given by:

[0179]

[0180] where r′ k,n is the echo without noise (i.e., the first filtered echo signal), J k,n is a 3x3 matrix, the (i,j) element is denoted as J k,n (i,j), thus the sensing CRLB of the kth UAV is as follows:

[0181]

[0182] where α1, α2 and α3 depend on the base station frequency, bandwidth and antenna number and other parameter configurations, κ k,n is given by:

[0183]

[0184] Then, the UAV differentiated QoS sensing performance optimization problem is modeled:

[0185] In optimizing multi-objective sensing performance, the existing technology usually adopts the method of minimizing the total CRLB or minimizing the maximum CRLB. However, this may cause resource allocation to focus on a single target. Under such standards, the system tends to allocate resources to the target with the highest potential gain, even if the target does not need such high sensing accuracy. Considering this, the present embodiments use an exponential utility function to represent the sensing performance of UAVs with different QoS requirements.

[0186] Specifically, without loss of generality, the FIM of d k is used as the performance indicator of the k-th UAV's perception, denoted as:

[0187] ξ k = J k (1,1) (11)

[0188] If AoA and ZoA are considered, it can be extended to a weighted FIM of three parameters. The global positioning performance of all UAVs (i.e., the first global positioning performance) can be denoted as:

[0189]

[0190] where η k denotes the expected positioning CRLB of the k-th UAV. Due to the characteristics of the exponential utility function, when the variable ξ k is greater than η k , the growth rate of the function φ is much slower than when the variable ξ k is less than η k . This characteristic shows that allocating resources to UAVs that have not yet reached the expected sensing performance will generate more benefits.

[0191] Therefore, the first optimization objective in the optimization problem can be modeled as maximizing the first global positioning performance. Denoted as:

[0192]

[0193] Optionally, according to the downlink communication signal received by the communication user from the base station, and the beam power allocation information of the base station, a first optimization constraint condition for resource allocation optimization in the active perception mode is generated, including:

[0194] According to the downlink communication signal received by the communication user from the base station, a first signal-to-interference ratio (SINR) received by the communication user is obtained.

[0195] Specifically, according to the above formula (1)-formula (3), the SINR of the m-th UE is:

[0196]

[0197] Taking the first SINR being greater than or equal to a first preset SINR threshold as a first condition, wherein the resource constraint information includes the first preset SINR threshold, and the first condition is denoted as:

[0198] SINR mm ≥ γ m , m = 1, 2,..., M,

[0199] Where, γ m This represents the SINR threshold (i.e., the first preset power threshold) for the m-th UE.

[0200] Based on the power allocation information and beam power allocation information of the base station, the transmit power of the base station and the number of beams supported by the base station are obtained.

[0201] The second condition is that the base station's transmit power is less than or equal to a first preset power threshold. The second condition is expressed as follows:

[0202]

[0203] Where, P = [p1, p2, ..., p n ] T It is the power allocation matrix (i.e., the beam power allocation information of the base station), p n =[p 1,n ,p 2,n ,...,p j,n ] T It is the beam power allocation vector of the nth base station, G=[g1,g2,...,g n ] T It is the beam selection matrix. The maximum transmit power of the nth base station (i.e., the first preset power threshold) is included in the resource constraint information, which includes the maximum transmit power of the base station.

[0204] The third condition is that the number of beams supported by the base station is less than or equal to a first preset number. The third condition is expressed as follows:

[0205]

[0206] Where G = [g1, g2, ..., g n ] T It is the beam selection matrix. This indicates the maximum number of beams supported by the nth base station (i.e., the first preset number), where the resource constraint information includes the maximum number of beams supported by the base station.

[0207] The fourth condition is generated based on the preset beam selection matrix;

[0208] Specifically, the fourth condition is expressed as follows:

[0209]

[0210] Where G = [g1, g2, ..., g n ] T It is the beam selection matrix.

[0211] According to the first condition, the second condition, the third condition and the fourth condition, the first optimization constraint condition is obtained.

[0212] In summary, according to the first optimization target and the first optimization constraint condition, the optimization problem is represented as follows:

[0213]

[0214] Where P=[p1,p2,...,p n ] T is a power allocation matrix, p n =[p 1,n ,p 2,n ,...,p j,n ] T is the beam power allocation vector of the nth base station, G=[g1,g2,...,g n ] T is a beam selection matrix, γ m represents the received SINR threshold of the mth UE, and represent the maximum transmit power and the maximum number of supported beams of the nth base station, respectively.

[0215] Optionally, according to the first optimization target and the first optimization constraint condition, resource allocation optimization is performed on the second multi-base station cooperative ISAC system in the active sensing mode, to obtain second resource allocation information in the active sensing mode, including:

[0216] The fourth condition in the first optimization constraint condition is approximately convexly relaxed to obtain a fifth condition.

[0217] Specifically, due to the coupling and discontinuity of the optimization variables, direct solution of the optimization problem is challenging. Therefore, first, an approximate convex relaxation is made by replacing g j,n ∈{0,1} with the inequality constraint 0≤g j,n ≤1.

[0218] The first optimization target is decomposed into a first sub-target and a second sub-target using an initial beam selection matrix, specifically, the original optimization problem is decomposed into two sub-problems, P and G are alternately optimized, the first sub-target is to maximize P, and the second sub-target is to maximize G.

[0219] According to the initial beam selection matrix, the numerator information and the denominator information in the first SINR received by the communication user are obtained.

[0220] Where, through the initial beam selection matrix G (initialized as an all-one matrix), a m (p) and bm (p) represents the numerator and denominator in the SINR expression received by the UE.

[0221] According to the molecule information and the denominator information, the first condition in the first optimization constraint condition is converted into a sixth condition, and specifically, the sixth condition is expressed as follows:

[0222] a m (p) > y m b m (p), m = 1,..., M,

[0223] According to the first sub-target, the sixth condition and the second condition in the first optimization constraint condition, a first optimization problem is generated.

[0224] That is, the original optimization problem can be written as sub-problem 1 (i.e., the first optimization problem), as follows:

[0225]

[0226] According to the second sub-target, the sixth condition, the third condition in the first optimization constraint condition and the fifth condition, a second optimization problem is generated.

[0227] Sub-problem 2 (i.e., the second optimization problem) related to the beam selection matrix is as follows:

[0228]

[0229] By alternately solving the first optimization problem and the second optimization problem, the second resource allocation information in the active sensing mode is obtained.

[0230] Sub-problem 1 is a convex problem and can be solved by a convex optimization solver. Then, the optimization result P * As a fixed variable in the original optimization problem, sub-problem 2 about the beam selection matrix G is obtained, and similarly, sub-problem 2 is convex and can be solved by a convex optimization solver. Then, the optimization result G * is used as a fixed variable to continue to solve sub-problem 1. By alternately solving the two sub-problems, the objective function is finally converged. The resource allocation information in the converged system is the second resource allocation information in the active sensing mode.

[0231] The following will be specifically described in combination with Figure 3 , the process of obtaining the resource allocation information in the active sensing mode provided by the embodiment of the application:

[0232] Prior location information of UAVs and UEs is acquired, a signal modeling analysis is performed on the cooperative sensing unit, communication performance metrics and sensing performance metrics are calculated according to the signal modeling, an optimization problem is established according to the communication QoS and sensing QoS requirements, the optimization problem is solved, and each base station beam selection vector and power allocation vector are obtained, and the base station performs beamforming and power regulation.

[0233] For this mode, the embodiments of the application model the relationship between downlink communication signals and active sensing echo signals and base station transmit beams in a multi-base station cooperative scenario, and derive the signal-to-interference-and-noise ratio (SINR) expression of the ground communication user and the Cramer-Rao lower bound (CRLB) of the aerial UAV sensing performance, respectively. By minimizing the CRLB, the sensing accuracy is improved under the condition that the approximate location of the UAV is known.

[0234] In terms of resource allocation optimization, to solve the problem of excessive concentration of resources on a single target caused by traditional methods, the embodiments of the application design a global sensing performance evaluation function based on an exponential utility function for multiple UAV targets with different sensing QoS requirements. Taking this as the optimization objective, an optimization problem is constructed in combination with the base station beam capability, transmit power constraint and ground user communication QoS requirement. The optimization problem is decomposed into two sub-optimization problems by an approximate convex relaxation method, and an alternating iteration algorithm is used to solve them, and finally efficient allocation of multi-base station cooperative beams and power is realized.

[0235] By reasonably allocating resources, the sensing performance is significantly improved, the multi-target requirements are taken into account, and the transmit power of the base station is reduced.

[0236] In an optional embodiment, the process of acquiring communication and sensing related information in the cooperative sensing mode is specifically described:

[0237] As shown in Figure 4 , in the cooperative sensing mode considered, the ISAC scenario is composed of J synchronous base stations, K communication users (CUs) and I UAV sensing targets, denoted as and Each BS is equipped with a unified planar array (UPA), the transmit antenna array is configured as M t =M tx ×M ty , and the receive antenna array is configured as M r =M rx ×M ry . In order to realize the ISAC capability, each BS can communicate with multiple CUs while sensing multiple targets. In addition, it is assumed that each CU is pre-assigned to the relevant BS. Denote as the CU set related to BSj, and k jdenotes the k-th CU associated with BS j. For sensing target scheduling, define a binary target scheduling variable (i.e., target scheduling matrix, is pre-defined) a j,i ∈ {0, 1}. If a j,i = 1, it means that BS j receives the reflection signal of sensing target i for processing, otherwise a j,i = 0 means that BS j does not receive the reflection signal of sensing target i for processing.

[0238] In an optional embodiment, the second resource allocation information of the second multi-base station cooperative ISAC system is obtained, including:

[0239] The transmission signal of the first base station in the second multi-base station cooperative ISAC system, the reception signal of the first communication user in communication with the first base station, and the reflection signal received by the first base station in the cooperative sensing mode are obtained; the first base station is any base station in the second multi-base station cooperative ISAC system; the first communication user is any communication user in communication with the first base station;

[0240] First, for the downlink signal model of the base station, the transmission signal of BS j (i.e., the first base station) can be represented as:

[0241] x j′ = W j′ s j′ (17)

[0242] wherein denotes the transmission beamforming matrix of the communication user served by BS j (is pre-defined), denotes the communication symbol vector. Without loss of generality, it is assumed that the communication symbols from different CUs are uncorrelated, that is, for the communication user for other communication users

[0243] Therefore, based on the signal model shown in the above formula (17), the CU k j′ (the first communication user) in communication with BS j' can be constructed as shown in the following formula (18):

[0244]

[0245] wherein denote the communication channels from BS to CU k j′ , respectively. The first part of formula (18) is the expected communication channel from BS j' to CU k j′The communication signal received by BS j is composed of three parts: the first part is the desired signal from the first communication user, the second part is the co-cell interference from other CUs, and the third part is the inter-cell interference from other BSs. The last part is the CU k j′ Additive White Gaussian Noise (AWGN) at the receiving end, where

[0246] Considering that the general UAV target moving speed is low, the sensing target motion state change is relatively small compared to the sensing signal transmission and processing capability, the accurate direction of the target can be estimated in advance to design the transmitted signal to detect the target. The reflected signal received by BS j can be represented as:

[0247]

[0248] where n r [l] is the receiving noise of the receiving BS j, represents the direct channel from BS j' to BS j, if j'≠j, then G' j ,j=1, otherwise 0. β j′,j,i is the reflection coefficient from the transmitting BS j' through the sensing target to the receiving BS j, including the round trip path loss and RCS σ rcs According to the radar equation, formula (20) is obtained:

[0249]

[0250] where λ represents the wavelength, d1, and d2 represent the distance from the transmitting BS to the sensing target and the distance from the sensing target to the receiving BS. Assuming that the antenna spacing of the BS is half a wavelength, the response vectors of the antenna array to the azimuth angle and the elevation angle are respectively represented as:

[0251]

[0252]

[0253] where m tx ∈[1,M tx ] and m ty ∈[1,M ty ] are antenna indexes.

[0254] According to the transmitting signal of the first base station and the receiving signal of the first communication user, a second SINR of the first communication user is obtained.

[0255] According to formula (17) and formula (18), the following is obtained:

[0256] The SINR (i.e., the second SINR) of the CU k j′ communicating with BS j' can be represented as:

[0257]

[0258] wherein, and are the intra- and inter-cell interference, respectively, which is expressed as:

[0259]

[0260]

[0261] a third SINR of the radar output signal of the target perceived at the first base station is obtained according to the reflected signal received by the first base station;

[0262] In the ISAC scenario of cooperative perception, BS j receives the reflected perception signals from multiple BSs through multiple targets. In order to capture the desired signal of target i, a receive beamforming filter is applied on the received perception signal y Considering that the premise of perception is that the reflected signal can be detected, the perception signal strength is usually positively correlated with the perception accuracy and the detection probability. Therefore, the SINR of the received perception signal can be used as a measure of the perception performance.

[0263] Based on the above formula (17), the radar output signal SINR (third SINR) of target i perceived at BS j can be expressed as:

[0264]

[0265] wherein, In formula (26), Γ j,i represents the total interference of target i perceived at BS j, which can be expressed as:

[0266]

[0267] wherein, the first term represents the direct interference signal from other base stations, the second term is noise, and the third term is the interference signal from other perception targets.

[0268] According to the second SINR, a second optimization objective is generated, which is to maximize the perception signal SINR of the receiving base station under the principle of max-min fairness, i.e., to maximize the second SINR as the second optimization objective, which is expressed as:

[0269]

[0270] According to the third SINR and the power budget information of the first base station, a second optimization constraint condition for resource allocation optimization in the cooperative sensing mode is generated, and the second optimization constraint condition is to satisfy the communication SINR of the communication user and the power budget constraint of the base station.

[0271] According to the second optimization target and the second optimization constraint condition, resource allocation optimization is performed on the second multi-base station cooperative ISAC system in the cooperative sensing mode, and second resource allocation information in the cooperative sensing mode is obtained.

[0272] Optionally, according to the third SINR and the power budget information of the first base station, the second optimization constraint condition for resource allocation optimization in the cooperative sensing mode is generated, including:

[0273] The seventh condition is that the third SINR is greater than or equal to a second preset SINR threshold, and is represented as follows:

[0274]

[0275] Wherein, γ min is the minimum SINR threshold required by the communication user, that is, the second preset SINR threshold, and the minimum SINR threshold required by the communication user is included in the resource constraint information.

[0276] The eighth condition is that the transmission power of the first base station is less than or equal to a second preset power threshold indicated by the power budget information of the first base station, and is represented as follows:

[0277]

[0278] Wherein, is the transmission beamforming matrix of the base station, that is, the transmission power of the first base station, P0 is the maximum transmission power of each base station, that is, the second preset power threshold, and the maximum transmission power of each base station is included in the resource constraint information.

[0279] According to a preset target scheduling matrix, a ninth condition is generated, and is represented as follows:

[0280]

[0281] According to the seventh condition, the eighth condition and the ninth condition, the second optimization constraint condition is obtained.

[0282] Under the maximum-minimum fairness principle, the sensing signal SINR of the receiving base station is maximized while satisfying the communication SINR of the communication user and the power budget constraint of the base station. This problem is expressed as optimizing the target scheduling variable a j,i , the base station transmission beamforming matrix and receive beamforming vectors (receive beamforming matrix) The joint optimization problem is as follows:

[0283]

[0284]

[0285] where γ min is the minimum SINR threshold required by the communication users, and P0 is the maximum transmit power of each base station.

[0286] Due to the multiplication of the beamforming matrix and the target scheduling matrix in the objective function, the optimization problem is a non-convex mixed integer programming. In addition, the fractional structure of the objective function also leads to a non-convex optimization problem, which is difficult to solve. Therefore, the original problem needs to be converted. The conversion of the original problem includes the following steps:

[0287] Optionally, according to the second optimization objective and the second optimization constraint, the resource allocation optimization of the second multi-base station cooperation ISAC system in the cooperative sensing mode is performed to obtain the second resource allocation information in the cooperative sensing mode, including:

[0288] According to the transmit beamforming matrix and the target scheduling matrix, the second optimization objective is converted into a third optimization problem;

[0289] The third optimization problem is solved by using generalized Rayleigh entropy to obtain the optimized receive beamforming matrix;

[0290] Specifically, the receive beamforming optimization problem: in this sub-problem, the transmit beamforming matrix and the target scheduling matrix are given. Then, the original problem only depends on u j,i The converted problem (i.e. the third optimization problem) can be solved by generalized Rayleigh entropy.

[0291] The second optimization objective is converted into a third sub-objective by using the Dinkelbach method;

[0292] According to the third sub-objective and the sixth condition and the seventh condition in the second optimization constraint, a fourth optimization problem is obtained;

[0293] The fourth optimization problem is solved to obtain the optimized transmit beamforming matrix;

[0294] Specifically, the transmit beamforming optimization problem: in this sub-problem (fourth optimization problem), the fractional structure of the original problem is first converted by using the Dinkelbach method, and then the optimal transmit beamforming matrix is obtained by using the Lagrange dual method.

[0295] According to the optimized receiving beamforming matrix, the optimized transmitting beamforming matrix and the second optimization target, a target scheduling matrix is optimized to obtain second resource allocation information in the cooperative sensing mode.

[0296] That is, a target scheduling optimization problem: based on the optimized transmitting and receiving beamforming variables, the target scheduling variables are optimized according to the highest SINR.

[0297] The following specifically describes a receiving beamforming optimization sub-problem:

[0298] Given the transmitting beamforming matrix and the target scheduling matrix, the receiving beamforming matrix only depends on the target. Therefore, the receiving beamforming matrix is optimized by the following optimization criterion:

[0299]

[0300] Where B represents the direct channel interference from other targets and other base stations, Q represents the covariance matrix of the downlink ISAC signal, and can be represented as

[0301]

[0302]

[0303] As can be seen from the expression of formula (32), the sub-problem maximizes the minimum sensing signal-to-noise ratio between targets, thereby obtaining the optimal receiving beamforming solution. In the case of fixed transmitting beamforming and target scheduling, the optimization of u j,i does not affect the performance of other targets for the detection of target i. Therefore, the optimization of u j,i of each target can be carried out separately. On this basis, the maximum-minimization optimization problem formula (32) is equivalent to the following problem:

[0304]

[0305] The transformed problem formula (32) is a generalized Rayleigh quotient, and the optimal solution of u is:

[0306]

[0307] Where V max (Π) represents the eigenvector corresponding to the maximum eigenvalue of the matrix H.

[0308] Transmitting beamforming optimization sub-problem:

[0309] Since a j,i is a 0-1 integer variable, a j,i is first relaxed to a continuous variable aj,i ∈ [0, 1]. Given the optimal receive beamforming matrix and the target scheduling variable, the transmit beamforming subproblem can be expressed as:

[0310]

[0311]

[0312]

[0313] where C j,i ,D j,i ,E j,i can be expressed as follows:

[0314]

[0315]

[0316]

[0317] It can be seen that the subproblem P3 is a quadratic problem about transmit beamforming , and the fractional structure of the objective function leads to a non-convex problem, which is difficult to solve directly. Inspired by the Dinkelbach method, the non-convexity of the subproblem P3 can be converted into a solvable problem. The optimal solution of the subproblem P3 can be obtained when and only when the following equation holds.

[0318]

[0319] where η ＊ is the optimal non-negative parameter, and denote the numerator and denominator of the objective function in P3, respectively.

[0320]

[0321]

[0322] The proof of the above conclusion can be obtained from the theory of generalized fractional programming. Therefore, the problem P3 can be converted to the following problem:

[0323]

[0324] By introducing an auxiliary variable γ, the problem P4 can be converted to:

[0325]

[0326]

[0327] Although P5 is more solvable, it remains a non-convex quadratic problem. To obtain the optimal solution, we use the Lagrange duality method to analyze the optimal solution of the P5 problem, which possesses some strong duality conditions. The Lagrange function can be expressed as:

[0328]

[0329] Where, δ={δ j,i} j∈J,i∈I , and ζ={ζ j′} j′∈J It is a dual variable.

[0330] Π is represented as:

[0331]

[0332] Therefore, the dual function is expressed as:

[0333]

[0334] To ensure that the Lagrange dual function formula (49) is not unbounded, it is obvious that Π≥0 and Therefore, the dual problem can be expressed as:

[0335]

[0336]

[0337]

[0338] For convex optimization problems, strong duality is always guaranteed under Slater conditions, but this is more difficult for non-convex optimization problems. However, strong duality can also be achieved in non-convex problems under certain conditions. The following theorem clarifies the relationship between the existence of saddle points and strong duality.

[0339] Grange function saddle point The following conditions must be met:

[0340]

[0341] Theorem 1: Let p ＊ ,d ＊ Let be the optimal values ​​of the nonconvex quadratic problem and the dual problem, respectively. yes The saddle point then exhibits strong duality d. ＊ =p ＊ It is true. Conversely, if d ＊ Finite, and (δ) ＊ ,α ＊ ,ζ＊ ) = argmax δ,α,ζ≥0 g(δ, α, ζ), the original problem has an optimal solution at , then is a saddle point of

[0342] Theorem 2: The condition for the existence of a saddle point is that there exists (δ ＊ , α ＊ , ζ ＊ ) such that is convex, and has a minimum value at

[0343]

[0344]

[0345]

[0346]

[0347]

[0348]

[0349] The proof of this theorem can be easily generalized from existing technology to multiple constraints.

[0350] Therefore, in order to prove the strong duality of non-convex quadratic optimization problems and obtain the optimal solution by the dual method, it is necessary to find a saddle point that satisfies Theorem 2.

[0351] Since the dual problem P6 is convex with respect to the Lagrange multiplier, its optimal solution (δ ＊ , α ＊ , ζ ＊ ) can be obtained by standard convex optimization techniques or interior point methods. Considering that Π is defined as a positive definite matrix in (50) to ensure the finiteness of the dual function formula (49), the Lagrangian function formula (49) is proved to be convex on and γ. Therefore, it is only necessary to find:

[0352]

[0353] If Π > 0, it means that Π is a full rank matrix, whose eigenvalues are all positive, so the null space of Π is zero vector, is designed to minimize formula (49). If Π ≯ 0, there is an eigenvalue of 0. Therefore, Π is not a full rank matrix, with a non-zero null space. From this derivation, the null space of Π spans the unit norm eigenvector u Πwhose eigenvalues are zero (i.e., Πu п. = 0). Thus, which can be chosen to minimize the formula (49), can be computed from the formula (60).

[0354] Based on this, Υ ＊ can also be obtained from the condition (56), i.e.,

[0355]

[0356] is equivalent to:

[0357]

[0358] Objective scheduling optimization problem sub-problems:

[0359] Observing the Lagrangian function (49), a j,i is only related to the first term. Therefore, given the strong duality proved in the foregoing, the Lagrangian function can be decomposed into multiple independent functions and solved independently:

[0360]

[0361] Therefore, the independent sub-problems can be expressed as:

[0362]

[0363] To minimize each optimal a j,i can be expressed as:

[0364]

[0365] Considering that the BS scheduling decisions for each target are independent of each other, the objective scheduling problem can be decoupled into independent decisions for each target. This significantly simplifies the problem, allowing us to independently schedule each target based on the highest achievable SINR. This approach not only simplifies the calculation process, but also complies with the max-min fairness criterion, ensuring that even the target under the worst perceived condition can obtain sufficient resources.

[0366] By iteratively computing the above three sub-problems, the multi-base station cooperative perception resource allocation result can be obtained.

[0367] This part proposes a multi-base station cooperative sensing and communication integrated (ISAC) resource scheduling method based on active and passive cooperation. Multiple base stations simultaneously perform multi-target sensing and communication. By using downlink communication signals for sensing, combined with receive beamforming, a mathematical model of target sensing and downlink communication signal-to-interference-and-noise ratio (SINR) is established. The optimization problem is formulated as a joint optimization of multi-target scheduling, base station transmit beamforming, and receive beamforming, with the goal of maximizing the minimum radar sensing SINR. Given the transmit beamforming and target scheduling, the optimal receive beamforming matrix is obtained by solving the generalized Rayleigh entropy problem; the Dinkelbach method is used to handle the fractional structure of the original problem, and then the Lagrange dual method is used to derive the optimal transmit beamforming matrix; based on the optimization of transmit and receive beamforming variables, the target scheduling variable is optimized according to the highest SINR to ensure multi-target sensing fairness. Through iterative calculation of the three sub-algorithms, the final resource optimization result is obtained, as shown in the flowchart of Figure 5 .

[0368] The base station obtains the prior position information of the unmanned aerial vehicle (UAV) and the UE, models the active and passive cooperative sensing signal, i.e., signal modeling in cooperative sensing mode, optimizes the receive beamforming, optimizes the transmit beamforming, and optimizes the sensing target scheduling.

[0369] Considering the reality of the environment, the traditional optimization algorithm is fixed, and it is difficult to select the sensing mode (active sensing and cooperative sensing) according to the actual environment, resulting in low resource allocation efficiency. In addition, the execution of the above resource allocation algorithm requires iterative optimization calculation, so it is difficult to adapt to the rapidly changing motion state of the sensing target, and further causes time lag in resource allocation, affecting the real-time performance of sensing. Therefore, based on the idea of data knowledge double driving, a resource allocation method based on generative AI data driving is designed in this part, which can calculate the sensing mode selection and resource allocation result according to the real-time sensing demand. The knowledge-driven method driven by the traditional optimization algorithm in the previous part is combined to constrain the data-driven method, ensuring the accuracy of the data-driven method. Specifically, the generative AI includes a feature extraction module, a feature clustering module, a cross-domain invariant feature extraction module, and a domain feature module.

[0370] In an optional embodiment, the AI resource optimization model is obtained by training using the second communication sensing related information and the second sensing mode, and / or using the second communication sensing related information and the second resource allocation information.

[0371] The second communication sensing related information is subjected to feature extraction to obtain first extracted features;

[0372] Among them, the main purpose of the feature extraction module is to vector embed the time value data such as various resource constraints, communication awareness (comprehensive awareness) index requirements and network topology in the communication awareness related information, and then splice and combine to provide high-quality input features for the subsequent generative AI model.

[0373] Specifically, the feature vector construction process is as follows:

[0374] Different constraints and indexes in the system are set, including resource constraints, network topology and comprehensive awareness index requirements. Each case s i The feature vector construction process is as follows:

[0375] Resource constraint vector (R i ):

[0376] The resource constraints can include power limits. Assuming there are M resource constraints, the resource constraint vector is:

[0377] R i =[R i1 ,R i2 ,…,R iM ]

[0378] Comprehensive awareness index vector (C i ):

[0379] The comprehensive awareness index includes the signal-to-interference-and-noise ratio (SINR). Assuming there are P comprehensive awareness indexes, the comprehensive awareness index vector is:

[0380] C i =[C i1 ,C i2 ,...,C iP ]

[0381] Base station vector (B i ):

[0382] The base station vector describes the positions of the J base stations. Assuming there are Q base stations, the base station vector is

[0383] B i =[B i1 ,B i2 ,...,B iQ ]

[0384] Communication user vector (U i ):

[0385] The communication user vector describes the positions of the K communication users. Assuming there are R users, the communication user vector is:

[0386] U i =[U i1 ,U i2 ,...,UiR ]

[0387] perception target vector (G i ):

[0388] The perception target vector describes the positions of I UAV perception targets. Assuming there are S target features, the perception target vector is:

[0389] G i =[G i1 , G i2 ,..., G iS ]

[0390] Through these steps and formulas, we can construct the feature extraction module in detail, effectively vectorize and embed the data required by various resource constraints, network topologies, and sensory indicators, and provide reliable input features for the resource allocation algorithm of generative AI.

[0391] Vector embedding:

[0392] Embed the above vectors to obtain the corresponding embedding vectors. Assuming that the embedding function is φ, the embedding vectors are represented as

[0393] E R = φ R (R i ), E C = φ C (C i ), E B = φ B (B i ),

[0394] E U = φ U (U i ), E G = φ G (G i )

[0395] Concatenate and merge the embedding vectors of resource constraints, sensory indicators, base stations, communication users, and perception targets to form the final feature vector (i.e., the first extracted feature):

[0396] F i = [E R ; E C ; E B ; E U ; E G ]

[0397] Where [;] represents the concatenation operation of vectors.

[0398] Feature clustering is performed on the first extracted feature to obtain multiple feature domains;

[0399] The goal of the feature clustering module is to aggregate similar cases into a domain, thereby leveraging the similarity of these cases in the feature space to share feature information.

[0400] Commonly used clustering algorithms include K-means, Hierarchical Clustering, and DBSCAN (Density-Based Spatial Clustering of Applications with Noise). In this approach, the choice of a suitable clustering algorithm depends on the properties of the feature vectors and the size of the data.

[0401] K-means is a distance-based clustering algorithm suitable for most datasets. Its goal is to divide the data into K clusters such that the distance from each data point to the center of its cluster is minimized.

[0402] Hierarchical clustering does not require pre-specifying the number of clusters; by constructing a tree-like hierarchical structure, the clustering level can be selected as needed.

[0403] Assuming the K-means clustering algorithm is chosen, the specific process is as follows:

[0404] Feature matrix construction: Feature vectors F of all cases i Composition of characteristic matrix F:

[0405]

[0406] in, n is the number of cases, and d is the dimension of each feature vector.

[0407] Randomly select K initial cluster centers μ1, μ2, ..., μ K Based on the distance between each feature vector and the cluster center, the feature vector is assigned to the nearest cluster center. Let the feature vector F... i To the cluster center μ k The distance is d(F) i μ k If the i-th feature vector belongs to a cluster, then the cluster label is: c i =argmin k d(F i μ k ).

[0408] Calculate the new center for each cluster, which is the mean of all data points assigned to that cluster:

[0409]

[0410] where C k is the set of the k-th cluster, |C k | is the number of data points in it. Iteration: repeat the clustering step until the cluster centers no longer change or the maximum number of iterations is reached.

[0411] By clustering, similar cases are aggregated into a feature domain, and cases in each domain share feature information. Let the k-th cluster domain be D k , and the cases contained be The feature vector of each domain is represented as:

[0412]

[0413] The extracted features in the same feature domain are weighted to obtain the first domain feature representation.

[0414] Specifically, the domain feature module further processes the features in each domain to adapt to the specific domain task requirements. Specifically, the attention mechanism and fully connected layer are used for feature processing, and the final feature representation is combined with the cross-domain invariant feature. The specific steps are as follows:

[0415] In-domain feature processing:

[0416] For each feature in the domain, the attention mechanism is used for weighted processing, and then the domain feature representation (i.e. the first domain feature representation) is obtained through the fully connected layer.

[0417] Attention mechanism:

[0418] The attention mechanism is applied to each case in each domain to calculate the attention weight. Let the input feature be F i , and the attention weight be α i :

[0419]

[0420] where W a is the parameter matrix of the attention weight.

[0421] Weighted feature representation:

[0422] The attention weight is applied to the feature to obtain the weighted feature representation:

[0423] F i，att = α i F i

[0424] Domain feature fusion:

[0425] All features processed by the attention mechanism in the domain are fused to obtain the domain feature representation:

[0426]

[0427] Fully connected layer:

[0428] Further process the domain features through the fully connected layer to obtain the final intra-domain feature representation (i.e., the first domain feature representation):

[0429] F d = W d F d + b d

[0430] where W d and b d are the parameters of the fully connected layer.

[0431] Feature extraction and fusion are performed on the extracted features in different feature domains to obtain a second domain feature representation.

[0432] Domain-invariant feature processing:

[0433] The goal of the cross-domain invariant feature extraction module is to extract consistent features from data in different domains, so that these features can be shared and applied across domains. This ensures that the generative Al model has consistency and robustness when processing unseen data. Here are the specific steps and related formulas:

[0434] Feature extraction:

[0435] Input the feature vector of each domain into the cross-domain feature extraction model to extract cross-domain invariant features. Assume that the input feature vector is F i , and the cross-domain invariant feature representation is F inv .

[0436] F i = [F Ri ; F Ti ; F Ci ; F Bi ; F Ui ; F Gi ]

[0437] Use a deep neural network (DNN) to extract cross-domain invariant features. Let the cross-domain feature extraction network be ψ, then:

[0438] F inv = ψ(F i )

[0439] Here, ψ can be a multi-layer perceptron (MLP) composed of multiple fully connected layers and activation functions:

[0440] F inv = ψ(F i ) = f L(W L (f L-1 (…f1(W1F i +b1)…)+b L-1 )+b L )

[0441] where L denotes the number of layers of the network, W l and b l are the weight matrix and bias vector of the l-th layer, respectively, and f l is the activation function of the l-th layer (such as ReLU, Sigmoid, etc.).

[0442] Fuse all case features:

[0443] To obtain cross-domain invariant features, fuse the features of all cases. Assuming there are n cases, the fused cross-domain invariant feature representation (i.e., the second domain feature representation) is:

[0444]

[0445] This step ensures consistency between different domains and fuses the features of all cases.

[0446] According to the first domain feature representation and the second domain feature representation, obtain a target feature representation;

[0447] Specifically, the cross-domain feature combination:

[0448] Concatenate or add the intra-domain feature representation and the cross-domain invariant feature to form the final feature representation (i.e., the target feature representation).

[0449] F final =[F d ;F inv ]

[0450] F final =F d +F inv

[0451] Using the target feature representation and the second perception mode, and / or using the target feature representation and the second resource allocation information to train the AI resource optimization model.

[0452] In an optional embodiment, training the AI resource optimization model using the target feature representation and the second resource allocation information includes:

[0453] Using a preset optimization method, obtaining first training resource allocation information according to the target feature representation;

[0454] The preset optimization mode is an existing traditional optimization algorithm, such as a convex optimization, a sequential approximation or a mathematical method.

[0455] The second training resource allocation information is obtained according to the target feature representation by using a preset initial AI model.

[0456] The loss function is obtained according to the first training resource allocation information and the second training resource allocation information.

[0457] The AI resource optimization model is obtained by adjusting model parameters in the initial AI model to reduce the loss function.

[0458] Specifically, the solution obtained by using the traditional optimization algorithm is used as a label, and a mean square error (MSE) loss is used to optimize the solution of the loss function.

[0459] Supposing that the solution (i.e., the training resource allocation information) obtained by using the traditional optimization algorithm is R opt , and the output of the generative AI model is R gen (i.e., the second training resource allocation information), the loss function is:

[0460]

[0461] The model parameters are updated by using gradient descent or other optimization algorithms to minimize the loss function, and the parameters of the generative AI model are trained so that the output of the resource allocation result is as close as possible to the solution of the traditional optimization algorithm.

[0462] Through the above steps and formulas, the cross-domain invariant feature extraction module can effectively process and combine feature information of different domains, the model can automatically select a perception mode according to data features, and the accuracy and efficiency of the generative AI model in resource allocation are improved. The generated model only needs to input data from the domain-invariant feature module, and the model can decide the perception mode and resource allocation.

[0463] The training process of the AI resource optimization model provided in the embodiments of the present application will be specifically described below with reference to the accompanying drawings. Figure 6

[0464] ​Feature extraction is performed on various resource constraints, communication awareness (communication awareness) indicators and network topology in the communication awareness related information to obtain a feature vector (i.e., first extracted feature). The feature vector is normalized and clustered, and an attention mechanism is introduced to obtain a first domain feature representation in the domain feature module and a second domain feature representation in the cross-domain invariant feature module. According to the first domain feature representation and the second domain feature representation, a target feature representation is obtained. The target feature representation is normalized, and a second training resource allocation information is obtained by a linear layer. A first training resource allocation information is obtained by using a traditional optimization algorithm. A loss function is obtained according to the second training resource allocation information and the first training resource allocation information. By adjusting the model parameters in the initial AI model, the loss function is reduced, and the direct correspondence between the communication resource related information and the training resource allocation information in the AI resource optimization model is obtained.

[0465] It should be noted that the training process of the correspondence between the communication resource related information and the sensing mode in the AI resource optimization model is basically the same as the above-mentioned direct correspondence between the communication resource related information and the training resource allocation information.

[0466] The embodiments of the present application realize intelligent management and adaptive resource allocation of two sensing modes by introducing generative AI. The generative AI can dynamically adjust the beam and power allocation strategy of the system according to environmental changes, the characteristics of the sensing target and the communication and sensing QoS requirements. The generative AI can extract intra-domain features and extra-domain features through training of historical data, and can learn and predict the sensing and communication requirements in different scenarios in real time. The AI model will select A-to-A or A-to-B mode according to the scene, and automatically allocate beam direction and power resources. In different scenarios, the AI autonomously determines which sensing mode to use. When there are more communication tasks around the base station, the AI will preferentially select the A-to-B mode to reduce the burden of the single base station; when the accuracy requirement is high, the A-to-A mode is selected. The AI ensures the best balance of communication and sensing tasks under resource competition through optimization algorithm. Multi-objective optimization: when processing multiple sensing targets, the AI combines the QoS requirements of different targets and optimizes through an exponential utility function to ensure that high-threat or high-priority targets obtain more resources, while low-priority targets maintain basic sensing capabilities.

[0467] The application provides a multi-base station beam and power allocation technology suitable for different communication and perception QoS requirements. The scheme is based on the base station deployment in the current cellular network, by introducing A-to-A and A-to-B two perception modes, using generative AI to adaptively learn and optimize the environment, dynamically switching the perception mode according to the scene, and autonomously allocating beam and power resources according to different communication and perception requirements, to realize efficient integrated sensing and communication (ISAC) resource management. Specifically, first, a multi-base station resource optimization scheme is designed for the A-to-A active perception mode and the A-to-B active and passive cooperative perception mode, and the condition information and resource allocation results of the two optimization schemes are used as the training data of the generative AI model. Second, a generative AI model based on the transformer architecture is constructed, which includes a domain-invariant feature extraction module and a cross-domain feature extraction module, can effectively extract general feature information and specific condition feature information under the two perception modes, and realize reliable resource allocation under unknown environmental conditions and precise resource allocation under known environmental conditions.

[0468] The feature extraction module can effectively embed various resource constraints, network topology, communication and perception indicators (such as SINR), base station location, user location, and perception target location information into a unified feature vector, ensuring standardized and quantitative processing of data in multi-base station cooperative communication and perception tasks. This module provides standardized input features for the generative AI model, ensuring the accuracy and consistency of resource allocation; the cross-domain invariant feature extraction module can identify and extract invariant features in different scenarios or domains, and embed these features into the feature representation of a specific resource optimization scenario, enhancing the generalization ability and robustness of the model. This module ensures that the generative AI can maintain the stability and consistency of resource allocation decisions when the environment changes dynamically, and realizes efficient communication and perception resource allocation in cross-domain scenarios; the feature fusion and connection module combines the cross-domain invariant features and the resource constraint features of a specific scenario through a deep neural network (DNN) to form a unified feature vector. Through multi-layer full connection structure and activation function processing, the effective combination of multi-domain features is ensured, which improves the resource allocation accuracy of the system in complex and dynamic scenarios, and ensures real-time resource scheduling optimization for communication and perception tasks; for the multi-base station A-to-A active perception system, an optimization problem modeling of differentiated QoS requirements is proposed. The resource allocation modeling is realized through an exponential utility function, which ensures the balance of perception accuracy and communication quality for different targets. This method can adapt to changes in requirements in different scenarios, and has high efficiency and flexibility.

[0469] As shown in Figure 7 The application also provides a multi-base station cooperative perception resource optimization device based on generative AI, which includes:

[0470] The first obtaining module 701 is configured to obtain first communication and sensing related information of a first multi-base-station cooperative integrated sensing and communication (ISAC) system.

[0471] The first processing module 702 is configured to input the first communication and sensing related information into a trained artificial intelligence (AI) resource optimization model to obtain first resource allocation information of the first multi-base-station cooperative ISAC system and / or a first sensing mode of the first multi-base-station cooperative ISAC system output by the AI resource optimization model.

[0472] The AI resource optimization model is configured to indicate at least one of the following:

[0473] A correspondence between the communication and sensing related information and the resource allocation information;

[0474] A correspondence between the communication and sensing related information and the sensing mode.

[0475] Optionally, the apparatus further includes:

[0476] The second obtaining module is configured to obtain second communication and sensing related information, a second sensing mode, and second resource allocation information of a second multi-base-station cooperative ISAC system.

[0477] The second processing module is configured to train the AI resource optimization model by using the second communication and sensing related information and the second sensing mode, and / or by using the second communication and sensing related information and the second resource allocation information.

[0478] Optionally, the communication and sensing related information includes at least one of the following:

[0479] A base station;

[0480] A communication user;

[0481] A sensing target;

[0482] Resource constraint information;

[0483] Communication and sensing index information.

[0484] Optionally, the sensing mode includes at least one of the following:

[0485] An active sensing mode;

[0486] A cooperative sensing mode.

[0487] Optionally, the second obtaining module includes:

[0488] The first obtaining unit is configured to obtain, in the active sensing mode, a downlink communication signal received by a communication user from a base station in a second multi-base-station cooperative ISAC system, a reflected echo signal of a sensing target received by the base station, and beam power allocation information of the base station;

[0489] The first processing unit is configured to generate, according to the reflected echo signal of the sensing target received by the base station, a first optimization target for resource allocation optimization in the active sensing mode;

[0490] The second processing unit is configured to generate, according to the downlink communication signal received by the communication user from the base station and the beam power allocation information of the base station, a first optimization constraint condition for resource allocation optimization in the active sensing mode;

[0491] The third processing unit is configured to perform resource allocation optimization on the second multi-base-station cooperative ISAC system in the active sensing mode according to the first optimization target and the first optimization constraint condition, to obtain second resource allocation information in the active sensing mode.

[0492] Optionally, the first processing unit is specifically configured to:

[0493] perform separation filtering processing on the reflected echo signal of the sensing target received by the base station, to obtain a first filtered echo signal;

[0494] generate a Cramer-Rao lower bound (CRLB) of sensing performance of the sensing target according to the first filtered echo signal;

[0495] obtain a first global positioning performance of the sensing target according to the CRLB of the sensing target, a reflected echo azimuth of arrival (AoA) of the sensing target relative to the base station, and a reflected echo zenith of arrival (ZoA) of the sensing target relative to the base station;

[0496] generate the first optimization target according to the first global positioning performance.

[0497] Optionally, the second processing unit is specifically configured to:

[0498] obtain a first signal-to-interference noise ratio (SINR) received by the communication user according to the downlink communication signal received by the communication user from the base station;

[0499] regard a first condition that the first SINR is greater than or equal to a first preset SINR threshold as the first condition;

[0500] obtain a transmission power of the base station and a number of beams supported by the base station according to the power allocation information of the base station and the beam power allocation information of the base station;

[0501] The second condition is that the base station's transmission power is less than or equal to a first preset power threshold.

[0502] The third condition is that the number of beams supported by the base station is less than or equal to the first preset number.

[0503] The fourth condition is generated based on the preset beam selection matrix;

[0504] The first optimization constraint is obtained based on the first condition, the second condition, the third condition, and the fourth condition.

[0505] Optionally, the third processing unit is specifically used for:

[0506] The fourth condition in the first optimization constraint is subjected to an approximate convex relaxation process to obtain the fifth condition;

[0507] Using the initial beam selection matrix, the first optimization objective is decomposed into a first sub-objective and a second sub-objective;

[0508] Based on the initial beam selection matrix, the numerator and denominator information in the first SINR received by the communication user are obtained;

[0509] Based on the numerator information and the denominator information, the first condition in the first optimization constraint is converted into the sixth condition;

[0510] Based on the first sub-objective, the sixth condition, and the second condition in the first optimization constraint, a first optimization problem is generated;

[0511] Based on the second sub-objective, the sixth condition, and the third and fifth conditions in the first optimization constraints, a second optimization problem is generated.

[0512] By alternately solving the first optimization problem and the second optimization problem, the second resource allocation information under the active perception mode is obtained.

[0513] Optionally, the second acquisition module includes:

[0514] The second acquisition unit is used to acquire, in the cooperative sensing mode, the transmitted signal of the first base station in the second multi-base station cooperative ISAC system, the received signal of the first communication user communicating with the first base station, and the reflected signal received by the first base station.

[0515] The fourth processing unit is used to obtain the second SINR of the first communication user based on the transmitted signal of the first base station and the received signal of the first communication user;

[0516] a fifth processing unit, configured to obtain a third SINR of a radar output signal of the target perceived at the first base station according to the reflected signal received by the first base station;

[0517] a sixth processing unit, configured to generate a second optimization target according to the second SINR;

[0518] a seventh processing unit, configured to generate a second optimization constraint condition for resource allocation optimization in the cooperative perception mode according to the third SINR and power budget information of the first base station;

[0519] an eighth processing unit, configured to perform resource allocation optimization for the second multi-base-station cooperative ISAC system in the cooperative perception mode according to the second optimization target and the second optimization constraint condition, to obtain second resource allocation information in the cooperative perception mode.

[0520] Optionally, the seventh processing unit is specifically configured to:

[0521] take the third SINR greater than or equal to a second preset SINR threshold as a seventh condition;

[0522] take the transmission power of the first base station less than or equal to a second preset power threshold indicated by the power budget information of the first base station as an eighth condition;

[0523] generate a ninth condition according to a preset target scheduling matrix;

[0524] obtain the second optimization constraint condition according to the seventh condition, the eighth condition and the ninth condition.

[0525] Optionally, the eighth processing unit is specifically configured to:

[0526] convert the second optimization target into a third optimization problem according to a transmission beamforming matrix and a target scheduling matrix;

[0527] solve the third optimization problem by using generalized Rayleigh entropy to obtain an optimized receiving beamforming matrix;

[0528] convert the second optimization target into a third sub-target by using a Dinkelbach method;

[0529] obtain a fourth optimization problem according to the third sub-target and a sixth condition and a seventh condition in the second optimization constraint condition;

[0530] solve the fourth optimization problem by using a Lagrange dual method to obtain an optimized transmission beamforming matrix;

[0531] According to the optimized receiving beamforming matrix, the optimized transmitting beamforming matrix and the second optimization target, a target scheduling matrix is optimized to obtain second resource allocation information in the cooperative sensing mode.

[0532] Optionally, the second processing module comprises:

[0533] The ninth processing unit is configured to perform feature extraction on the second communication sensing related information to obtain first extracted features.

[0534] The tenth processing unit is configured to perform feature clustering on the first extracted features to obtain a plurality of feature domains.

[0535] The eleventh processing unit is configured to perform weighting processing on the extracted features in the same feature domain to obtain a first domain feature representation.

[0536] The twelfth processing unit is configured to perform feature extraction and fusion on the extracted features in different feature domains to obtain a second domain feature representation.

[0537] The thirteenth processing unit is configured to obtain a target feature representation according to the first domain feature representation and the second domain feature representation.

[0538] The fourteenth processing unit is configured to train the AI resource optimization model by using the target feature representation and the second sensing mode, and / or by using the target feature representation and the second resource allocation information.

[0539] Optionally, the fourteenth processing unit is specifically configured to:

[0540] obtain first training resource allocation information according to the target feature representation by using a preset optimization mode;

[0541] obtain second training resource allocation information according to the target feature representation by using a preset initial AI model;

[0542] obtain a loss function according to the first training resource allocation information and the second training resource allocation information;

[0543] obtain the AI resource optimization model by adjusting model parameters in the initial AI model to reduce the loss function.

[0544] It should be noted that the generative AI multi-base station cooperative sensing resource optimization apparatus provided by the embodiments of the present application is an apparatus capable of executing the above-mentioned generative AI multi-base station cooperative sensing resource optimization method, and all embodiments of the above-mentioned generative AI multi-base station cooperative sensing resource optimization method are applicable to the apparatus and can achieve the same or similar technical effects.

[0545] As Figure 8 shown, the embodiment of the present application also provides a multi-base station cooperative sensing resource optimization device of generative AI, comprising: a processor 801; and a memory 803 connected with the processor 801 through a bus interface 802, the memory 803 is used to store programs and data used by the processor 801 in executing operations, and the processor 801 calls and executes the programs and data stored in the memory 803.

[0546] Among them, the transceiver 804 is connected with the bus interface 802, used for receiving and sending data under the control of the processor 801, specifically, the processor 801 is used to read the program in the memory 803, and execute the following process:

[0547] Obtain the first communication sensing related information of the first multi-base station cooperative integrated sensing and communication (ISAC) system;

[0548] Input the first communication sensing related information into the trained artificial intelligence (AI) resource optimization model to obtain the first resource allocation information of the first multi-base station cooperative ISAC system and / or the first sensing mode of the first multi-base station cooperative ISAC system output by the AI resource optimization model;

[0549] Among them, the AI resource optimization model is used to indicate at least one of the following:

[0550] The correspondence between the communication sensing related information and the resource allocation information;

[0551] The correspondence between the communication sensing related information and the sensing mode.

[0552] Optionally, the processor 801 is also used for:

[0553] Obtain the second communication sensing related information, the second sensing mode and the second resource allocation information of the second multi-base station cooperative ISAC system;

[0554] Train the AI resource optimization model by using the second communication sensing related information and the second sensing mode, and / or by using the second communication sensing related information and the second resource allocation information.

[0555] Optionally, the communication sensing related information comprises at least one of the following:

[0556] Base station;

[0557] Communication user;

[0558] Sensing target;

[0559] Resource constraint information;

[0560] communication awareness index information.

[0561] Optionally, the awareness mode comprises at least one of:

[0562] active awareness mode;

[0563] cooperative awareness mode.

[0564] Optionally, the processor 801 is specifically configured to:

[0565] obtain, in the active awareness mode, a downlink communication signal received by a communication user from a base station in the second multi-base station cooperative ISAC system, a reflected echo signal of an awareness target received by the base station, and beam power allocation information of the base station;

[0566] generate, according to the reflected echo signal of the awareness target received by the base station, a first optimization objective for resource allocation optimization in the active awareness mode;

[0567] generate, according to the downlink communication signal received by the communication user from the base station and the beam power allocation information of the base station, a first optimization constraint condition for resource allocation optimization in the active awareness mode;

[0568] perform resource allocation optimization on the second multi-base station cooperative ISAC system in the active awareness mode according to the first optimization objective and the first optimization constraint condition, to obtain second resource allocation information in the active awareness mode.

[0569] Optionally, the processor 801 is specifically configured to:

[0570] perform separation filtering processing on the reflected echo signal of the awareness target received by the base station, to obtain a first filtered echo signal;

[0571] generate a Cramer-Rao lower bound (CRLB) of awareness performance of the awareness target according to the first filtered echo signal;

[0572] obtain a first global positioning performance of the awareness target according to the CRLB of the awareness target, a reflected echo azimuth angle (AoA) of the awareness target relative to the base station, and a reflected echo zenith angle (ZoA) of the awareness target relative to the base station;

[0573] generate the first optimization objective according to the first global positioning performance.

[0574] Optionally, the processor 801 is specifically configured to:

[0575] obtain a first signal-to-interference noise ratio (SINR) received by the communication user according to the downlink communication signal received by the communication user from the base station;

[0576] the first SINR being greater than or equal to a first preset SINR threshold as the first condition;

[0577] obtaining the transmission power of the base station and the number of supported beams of the base station according to the power allocation information of the base station and the beam power allocation information of the base station;

[0578] the transmission power of the base station being less than or equal to a first preset power threshold as the second condition;

[0579] the number of supported beams of the base station being less than or equal to a first preset number as the third condition;

[0580] generating a fourth condition according to a preset beam selection matrix;

[0581] obtaining the first optimization constraint condition according to the first condition, the second condition, the third condition and the fourth condition.

[0582] Optionally, the processor 801 is specifically configured to:

[0583] performing approximate convex relaxation processing on the fourth condition in the first optimization constraint condition to obtain a fifth condition;

[0584] decomposing the first optimization objective into a first sub-objective and a second sub-objective by using an initial beam selection matrix;

[0585] obtaining numerator information and denominator information in the first SINR received by the communication user according to the initial beam selection matrix;

[0586] converting the first condition in the first optimization constraint condition into a sixth condition according to the numerator information and the denominator information;

[0587] generating a first optimization problem according to the first sub-objective, the sixth condition and the second condition in the first optimization constraint condition;

[0588] generating a second optimization problem according to the second sub-objective, the sixth condition, the third condition in the first optimization constraint condition and the fifth condition;

[0589] obtaining the second resource allocation information in the active sensing mode by alternately solving the first optimization problem and the second optimization problem.

[0590] Optionally, the processor 801 is specifically configured to:

[0591] obtaining, in the cooperative sensing mode, a transmission signal of a first base station in the second multi-base station cooperative ISAC system, a reception signal of a first communication user in communication with the first base station, and a reflection signal received by the first base station;

[0592] obtaining a second SINR of the first communication user according to the transmission signal of the first base station and the reception signal of the first communication user;

[0593] obtaining a third SINR of a radar output signal of a sensing target at the first base station according to the reflection signal received by the first base station;

[0594] generating a second optimization target according to the second SINR;

[0595] generating a second optimization constraint condition for resource allocation optimization in the cooperative sensing mode according to the third SINR and power budget information of the first base station;

[0596] performing resource allocation optimization on the second multi-base station cooperative ISAC system in the cooperative sensing mode according to the second optimization target and the second optimization constraint condition, to obtain second resource allocation information in the cooperative sensing mode.

[0597] Optionally, the processor 801 is specifically configured to:

[0598] taking the third SINR being greater than or equal to a second preset SINR threshold as a seventh condition;

[0599] taking the transmission power of the first base station being less than or equal to a second preset power threshold indicated by the power budget information of the first base station as an eighth condition;

[0600] generating a ninth condition according to a preset target scheduling matrix;

[0601] obtaining the second optimization constraint condition according to the seventh condition, the eighth condition and the ninth condition.

[0602] Optionally, the processor 801 is specifically configured to:

[0603] transforming the second optimization target into a third optimization problem according to a transmission beamforming matrix and a target scheduling matrix;

[0604] solving the third optimization problem by using generalized Rayleigh entropy to obtain an optimized reception beamforming matrix;

[0605] transforming the second optimization target into a third sub-target by using a Dinkelbach method;

[0606] According to the third sub-target and a sixth condition and a seventh condition in the second optimization constraint condition, a fourth optimization problem is obtained;

[0607] By using a Lagrange dual method, the fourth optimization problem is solved to obtain an optimized transmit beamforming matrix;

[0608] According to the optimized receive beamforming matrix, the optimized transmit beamforming matrix and the second optimization target, a target scheduling matrix is optimized to obtain second resource allocation information in the cooperative sensing mode.

[0609] Optionally, the processor 801 is specifically configured to:

[0610] The second communication sensing related information is feature extracted to obtain first extracted features;

[0611] The first extracted features are feature clustered to obtain a plurality of feature domains;

[0612] The extracted features in the same feature domain are weighted processed to obtain a first domain feature representation;

[0613] The extracted features in different feature domains are feature extracted and fused to obtain a second domain feature representation;

[0614] According to the first domain feature representation and the second domain feature representation, a target feature representation is obtained;

[0615] By using the target feature representation and the second sensing mode, and / or by using the target feature representation and the second resource allocation information for training, the AI resource optimization model is obtained.

[0616] Optionally, the processor 801 is specifically configured to:

[0617] According to the target feature representation, first training resource allocation information is obtained by using a preset optimization mode;

[0618] According to the target feature representation, second training resource allocation information is obtained by using a preset initial AI model;

[0619] According to the first training resource allocation information and the second training resource allocation information, a loss function is obtained;

[0620] By adjusting model parameters in the initial AI model, the loss function is reduced to obtain the AI resource optimization model.

[0621] Wherein, in Figure 8In particular, the bus architecture can include any number of interconnected buses and bridges, specifically, various circuitry linking the processor(s) 801 and the memory represented by the memory 803. The bus architecture can also link various other circuitry, such as peripheral devices, voltage stabilizers, and power management circuitry, which are well known in the art and thus, not further described herein. The bus interface provides the user interface 805. The transceiver 804 can be a number of elements, including a transmitter that provides a means for communicating with various other apparatuses over a transmission medium and a receiver that provides a means for communicating with various other apparatuses over a transmission medium. The processor 801 is responsible for managing the bus architecture and general processing, and the memory 803 can store data used by the processor 801 in executing operations.

[0622] In addition, the embodiment of the present application further provides a readable storage medium, which stores a computer program, wherein the program is executed by a processor to realize the steps in the method for generating AI multi-base station cooperative sensing resource optimization provided in any one of the above embodiments.

[0623] In several embodiments provided in the present application, it should be understood that the disclosed method and device can be implemented in other manners. For example, the embodiment of the device described above is merely illustrative. For example, the division of the units is only a logical function division. There can be another division manner for the actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections can be indirect couplings or communication connections through some interfaces, devices or units, and can be electrical, mechanical or in other forms.

[0624] In addition, each function unit in the embodiments of the present application can be integrated into a processing unit, each unit can be physically present separately, or two or more units can be integrated into one unit. The above integrated unit can be implemented in the form of hardware or hardware plus software function unit.

[0625] The integrated unit in the form of software function unit can be stored in a computer readable storage medium. The software function unit is stored in a storage medium, and includes a plurality of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform part of steps of the resource selection method according to the embodiments of the present application, or to perform part of steps of the information sending method according to the embodiments of the present application. The storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various storage program code mediums.

[0626] The embodiment of the present application further provides a computer program product, which comprises computer instructions, and when the computer instructions are executed by a processor, each process of the method embodiment shown in the above Figure 1 The same technical effects can be achieved, and thus details are not described herein again.

[0627] The above is the preferred embodiment of the present application, and it should be noted that those skilled in the art can make several improvements and refinements without departing from the principles of the present application, and these improvements and refinements are also within the protection scope of the present application.

Claims

1. A method for multi-base station cooperative sensing resource optimization based on generative AI, characterized in that, The method comprises: obtaining first communication sensing related information of a first multi-base station cooperative integrated sensing and communication (ISAC) system; inputting the first communication sensing related information into a trained artificial intelligence (AI) resource optimization model to obtain first resource allocation information of the first multi-base station cooperative ISAC system and / or a first sensing mode of the first multi-base station cooperative ISAC system output by the AI resource optimization model; wherein the AI resource optimization model is used to indicate at least one of the following: a correspondence between the communication sensing related information and the resource allocation information; a correspondence between the communication sensing related information and the sensing mode; wherein the method further comprises: obtaining second communication sensing related information, a second sensing mode, and second resource allocation information of a second multi-base station cooperative ISAC system; training the AI resource optimization model using the second communication sensing related information and the second sensing mode, and / or using the second communication sensing related information and the second resource allocation information; wherein training the AI resource optimization model using the second communication sensing related information and the second sensing mode, and / or using the second communication sensing related information and the second resource allocation information comprises: extracting features from the second communication sensing related information to obtain first extracted features; performing feature clustering on the first extracted features to obtain a plurality of feature domains; performing weighting processing on the extracted features in the same feature domain to obtain a first domain feature representation; performing feature extraction and fusion on the extracted features in different feature domains to obtain a second domain feature representation; obtaining a target feature representation according to the first domain feature representation and the second domain feature representation; training the AI resource optimization model using the target feature representation and the second sensing mode, and / or using the target feature representation and the second resource allocation information; wherein training the AI resource optimization model using the target feature representation and the second resource allocation information comprises: obtaining first training resource allocation information according to the target feature representation using a preset optimization method; obtaining second training resource allocation information according to the target feature representation using a preset initial AI model; obtaining a loss function according to the first training resource allocation information and the second training resource allocation information; obtaining the AI resource optimization model by adjusting model parameters in the initial AI model to reduce the loss function.

2. The method of claim 1, wherein, The communication sensing related information comprises at least one of the following: a base station; a communication user; a sensing target; resource constraint information; communication sensing index information.

3. The method of claim 1, wherein, The sensing mode comprises at least one of the following: an active sensing mode; a cooperative sensing mode.

4. The method of claim 1, wherein, Obtaining the second resource allocation information of the second multi-base station cooperative ISAC system comprises: obtaining, in an active sensing mode, downlink communication signals received by a communication user from a base station, echo signals of a sensing target reflected by a base station, and beam power allocation information of the base station in the second multi-base station cooperative ISAC system. generate a first optimization target for resource allocation optimization in the active sensing mode according to echo signals of reflections of the sensing target received by the base station; generate a first optimization constraint condition for resource allocation optimization in the active sensing mode according to downlink communication signals received by the base station from the communication user and beam power allocation information of the base station; perform resource allocation optimization on the second multi-base-station cooperative ISAC system in the active sensing mode according to the first optimization target and the first optimization constraint condition, to obtain second resource allocation information in the active sensing mode.

5. The method of claim 4, wherein, generate a first optimization target for resource allocation optimization in the active sensing mode according to echo signals of reflections of the sensing target received by the base station, including: perform separation filtering processing on the echo signals of reflections of the sensing target received by the base station to obtain first filtered echo signals; generate a Cramer-Rao lower bound (CRLB) of sensing performance of the sensing target according to the first filtered echo signals; obtain a first global positioning performance of the sensing target according to the CRLB of the sensing target, an echo azimuth of arrival (AoA) of the sensing target relative to the base station, and an echo zenith of arrival (ZoA) of the sensing target relative to the base station; generate the first optimization target according to the first global positioning performance.

6. The method of claim 4, wherein, generate a first optimization constraint condition for resource allocation optimization in the active sensing mode according to downlink communication signals received by the base station from the communication user and beam power allocation information of the base station, including: obtain a first signal-to-interference noise ratio (SINR) received by the communication user according to the downlink communication signals received by the base station from the communication user; take a first condition that the first SINR is greater than or equal to a first preset SINR threshold; obtain a transmission power of the base station and a number of beams supported by the base station according to the power allocation information of the base station and the beam power allocation information of the base station; take a second condition that the transmission power of the base station is less than or equal to a first preset power threshold; take a third condition that the number of beams supported by the base station is less than or equal to a first preset number; generate a fourth condition according to a preset beam selection matrix; obtain the first optimization constraint condition according to the first condition, the second condition, the third condition, and the fourth condition.

7. The method of claim 4, wherein, perform resource allocation optimization on the second multi-base-station cooperative ISAC system in the active sensing mode according to the first optimization target and the first optimization constraint condition, to obtain second resource allocation information in the active sensing mode, including: perform approximate convex relaxation processing on the fourth condition in the first optimization constraint condition to obtain a fifth condition; decompose the first optimization target into a first sub-target and a second sub-target by using an initial beam selection matrix; obtain numerator information and denominator information in the first SINR received by the communication user according to the initial beam selection matrix; convert the first condition in the first optimization constraint condition into a sixth condition according to the numerator information and the denominator information; generating a first optimization problem according to the first sub-target, the sixth condition and a second condition in the first optimization constraint condition; generating a second optimization problem according to the second sub-target, the sixth condition, a third condition in the first optimization constraint condition and the fifth condition; obtaining the second resource allocation information in the active sensing mode by alternately solving the first optimization problem and the second optimization problem.

8. The method of claim 1, wherein, obtaining the second resource allocation information of the second multi-base-station cooperative ISAC system, comprising: obtaining, in the cooperative sensing mode, a transmission signal of a first base station in the second multi-base-station cooperative ISAC system, a reception signal of a first communication user in communication with the first base station, and a reflection signal received by the first base station; obtaining a second SINR of the first communication user according to the transmission signal of the first base station and the reception signal of the first communication user; obtaining a third SINR of a radar output signal of a sensing target at the first base station according to the reflection signal received by the first base station; generating a second optimization target according to the second SINR; generating a second optimization constraint condition for resource allocation optimization in the cooperative sensing mode according to the third SINR and power budget information of the first base station; performing resource allocation optimization on the second multi-base-station cooperative ISAC system in the cooperative sensing mode according to the second optimization target and the second optimization constraint condition, to obtain the second resource allocation information in the cooperative sensing mode.

9. The method of claim 8, wherein, generating a second optimization constraint condition for resource allocation optimization in the cooperative sensing mode according to the third SINR and power budget information of the first base station, comprising: taking the third SINR greater than or equal to a second preset SINR threshold as a seventh condition; taking the transmission power of the first base station less than or equal to a second preset power threshold indicated by the power budget information of the first base station as an eighth condition; generating a ninth condition according to a preset target scheduling matrix; obtaining the second optimization constraint condition according to the seventh condition, the eighth condition and the ninth condition.

10. The method of claim 9, wherein, performing resource allocation optimization on the second multi-base-station cooperative ISAC system in the cooperative sensing mode according to the second optimization target and the second optimization constraint condition, to obtain the second resource allocation information in the cooperative sensing mode, comprising: transforming the second optimization target into a third optimization problem according to a transmission beamforming matrix and a target scheduling matrix; solving the third optimization problem by using generalized Rayleigh entropy to obtain an optimized reception beamforming matrix; transforming the second optimization target into a third sub-target by using a Dinkelbach method; obtaining a fourth optimization problem according to the third sub-target and a sixth condition and a seventh condition in the second optimization constraint condition; solving the fourth optimization problem by using a Lagrange dual method to obtain an optimized transmission beamforming matrix; According to the optimized receiving beamforming matrix, the optimized transmitting beamforming matrix and the second optimization target, a target scheduling matrix is optimized to obtain second resource allocation information in the cooperative sensing mode. 11.A device for optimizing perception resources based on a generative AI and multi-base station cooperation, characterized in that, Comprise: The first acquisition module is used for acquiring first communication sensing related information of a first multi-base station cooperative integrated sensing and communication (ISAC) system; The first processing module is used for inputting the first communication sensing related information into a trained artificial intelligence (AI) resource optimization model to obtain first resource allocation information of the first multi-base station cooperative ISAC system and / or a first sensing mode of the first multi-base station cooperative ISAC system output by the AI resource optimization model; Wherein, the AI resource optimization model is used to indicate at least one of the following: The correspondence between the communication sensing related information and the resource allocation information; The correspondence between the communication sensing related information and the sensing mode; Wherein, the device further comprises: The second acquisition module is used for acquiring second communication sensing related information, a second sensing mode and second resource allocation information of a second multi-base station cooperative ISAC system; The second processing module is used for training the AI resource optimization model by using the second communication sensing related information and the second sensing mode, and / or by using the second communication sensing related information and the second resource allocation information; Wherein, the second processing module comprises: The ninth processing unit is used for extracting features from the second communication sensing related information to obtain first extracted features; The tenth processing unit is used for clustering the first extracted features to obtain a plurality of feature domains; The eleventh processing unit is used for weighting the extracted features in the same feature domain to obtain a first domain feature representation; The twelfth processing unit is used for extracting and fusing the extracted features in different feature domains to obtain a second domain feature representation; The thirteenth processing unit is used for obtaining a target feature representation according to the first domain feature representation and the second domain feature representation; The fourteenth processing unit is used for training the AI resource optimization model by using the target feature representation and the second sensing mode, and / or by using the target feature representation and the second resource allocation information; Wherein, the fourteenth processing unit is specifically used for: Obtaining first training resource allocation information according to the target feature representation by using a preset optimization method; Obtaining second training resource allocation information according to the target feature representation by using a preset initial AI model; Obtaining a loss function according to the first training resource allocation information and the second training resource allocation information; By adjusting the model parameters in the initial AI model to reduce the loss function, the AI resource optimization model is obtained.

12. A multi-base station cooperative sensing resource optimization device based on generative AI, characterized in that, Comprise: A processor, a memory and a program stored on the memory and executable on the processor, the program being executed by the processor to implement the steps in the generative AI based multi-base station cooperative sensing resource optimization method according to any one of claims 1 to 10.

13. A readable storage medium, characterized by, The readable storage medium has a program stored thereon, and the program, when executed by a processor, implements the steps in the multi-base station cooperative sensing resource optimization method based on generative AI in any one of claims 1 to 10.

14. A computer program product, characterised in that, The computer program product comprises computer instructions, and the computer instructions, when executed by a processor, implement the steps in the multi-base station cooperative sensing resource optimization method based on generative AI in any one of claims 1 to 10.

Images