RSMA assisted high energy efficient beamforming precoder design method for nf-isac system

Abstract

This invention relates to a high-efficiency beamforming precoder design method for RSMA-assisted NF-ISAC systems, belonging to the field of near-field communication and sensing integration technology. It includes constructing an RSMA-assisted integrated communication and sensing system model within the near-field range, establishing a joint optimization problem for beamforming precoder design, communication, and sensing resource allocation with the goal of maximizing system energy efficiency, and iteratively solving the reconstructed problem using Dinkelbach's algorithm and SCA algorithm to obtain the optimal system energy efficiency. This invention effectively suppresses multi-user interference and introduces additional degrees of freedom, significantly improving system performance and maximizing the energy efficiency of NF-ISAC systems under normal conditions. It also rigorously guarantees the QoS performance of communication users and the sensing and detection capabilities of radar targets within the system, providing a new technical solution for energy efficiency optimization of NF-ISAC systems, future green and sustainable communication systems, and robust transmission design.

Description

Technical Field

[0001] This invention belongs to the field of Near Field Integrated Sensing and Communication (NF-ISAC) technology, specifically relating to a design method for a high-efficiency beamforming pre-encoder for an RSMA-assisted NF-ISAC system. Background Technology

[0002] Integrated Sensing and Communication (ISAC) technology, through spectrum sharing and joint signal processing, integrates wireless communication and radar sensing capabilities, and is widely recognized as one of the key enabling technologies for the 6G era. This technology aims to integrate the traditionally separate functions of wireless communication and radar sensing into a unified dual-function architecture, thereby significantly improving the utilization efficiency of spectrum, energy, and hardware resources. With the continuous evolution of wireless communication technology, 5G and future 6G networks are developing towards greater intelligence and integration, especially in complex communication environments where the deep integration of sensing and communication has become a critical requirement. This dual-function characteristic is crucial for emerging application scenarios such as autonomous vehicles, smart cities, and industrial automation. Furthermore, these emerging fields also have stringent requirements for high-resolution sensing, high-reliability communication, and high energy efficiency (EE).

[0003] Meanwhile, with the widespread application of ELAA and UHF bands in 6G systems, Rayleigh distances can be extended to tens or even hundreds of meters, and the propagation mode of electromagnetic waves has changed from plane waves in the far-field (FF) to spherical waves in the near-field (NF). Against this backdrop, NF-ISAC systems have become feasible and highly promising. However, this transformation introduces unique channel characteristics in the NF region, making NF beamforming extremely complex. In NF-ISAC systems, compared to FF plane waves which rely solely on the angle dimension, spherical waves introduce an additional distance dimension, posing new challenges to NF beamforming design. Furthermore, in NF-ISAC systems employing complex NF channels, the joint optimization of spherical wave propagation and communication and sensing functions must be considered simultaneously, further increasing the complexity of NF beamforming and significantly exacerbating the difficulty of system interference management. At the same time, the ELAA arrays used in NF-ISAC systems (which may contain hundreds or even thousands of antennas) inevitably generate significant power consumption, making system-level energy efficiency a crucial and non-negligible design metric. Existing ISAC schemes developed based on FF channel models exhibit severe performance degradation when directly applied to NF scenarios. Therefore, the beamforming design scheme must be redesigned for the NF-ISAC system.

[0004] ISAC technology combines wireless communication with environmental sensing capabilities. By sharing spectrum resources, it enables both efficient data transmission and precise environmental monitoring. This integration not only improves spectrum resource utilization efficiency but also reduces system latency and enhances real-time performance, making it particularly suitable for applications requiring both communication and sensing capabilities, such as autonomous driving, smart manufacturing, and smart cities. ISAC's core advantage lies in the fact that the fusion of communication and sensing allows the same spectrum resource to serve multiple functions, improving overall network performance. For example, in autonomous driving applications, ISAC ensures that vehicles maintain stable communication connections while traveling at high speeds, while simultaneously sensing changes in the surrounding environment in real time, guaranteeing driving safety and accurate path planning. This collaborative approach significantly improves the system's adaptability to environmental changes and reduces latency caused by improper resource allocation.

[0005] Furthermore, the ISAC system is designed to perform both communication and radar sensing tasks simultaneously, leading to more complex interference scenarios in NF-ISAC systems. This interference exists not only between communication users but also between communication beams and radar beams. Common interference management methods to address this issue include space-division multiple access (SDMA) and non-orthogonal multiple access (NOMA). However, SDMA treats all interference as noise, while NOMA completely decodes it. Both methods lack flexibility and fail to provide effective interference management, limiting system performance. To achieve more flexible and efficient interference management, RSMA emerged as a powerful framework. RSMA splits the messages a user wants to send into public and private parts. At the receiving end, it first treats the private stream and interference as noise, then removes the public stream using Successive Interference Cancellation (SIC) technology, and finally decodes the user's private stream as noise. This technique offers greater flexibility compared to SDMA, which treats interference as noise, and NOMA, which completely decodes interference, significantly improving system performance. Studies of traditional FF (Front-Off) multi-antenna systems have shown that RSMA (Reverse-Side Magnetic Multi-Analog) significantly outperforms SDMA (Short-Side Magnetic Multi-Analog) and NOMA (Non-Front-Off) technologies in terms of spectral efficiency, energy efficiency, user fairness, and robustness. However, when the operating region shifts from FF mode to NF (Non-Front-Off) mode, the propagation characteristics of electromagnetic waves undergo a fundamental change, with the spherical wave model replacing the plane wave model of FF. In this new electromagnetic environment, the applicability and potential advantages of RSMA in NF-ISAC (Non-Front-Off) systems still lack in-depth research, which is one of the core motivations of this invention.

[0006] In summary, combining NF-ISAC with RSMA provides an innovative solution for next-generation communication-sensing fusion. ISAC, by integrating sensing and communication functions and sharing spectrum and hardware resources, achieves complementarity between sensing and communication, improving the overall system capability and resource utilization. RSMA further optimizes spectral and resource efficiency through flexible rate splitting and dynamic resource allocation. Although these technologies have made some progress in far-field environments where electromagnetic wave propagation is plane wave, they remain a challenge in complex and energy-intensive near-field environments. Therefore, studying the energy efficiency of NF-ISAC using large-scale antenna arrays and exploring high-efficiency beamforming design for RSMA-assisted NF-ISAC systems in near-field environments is of great significance, as it can meet the requirements for the green and sustainable development of future wireless networks. Summary of the Invention

[0007] This invention provides a design method for a high-efficiency beamforming pre-encoder for an RSMA-assisted NF-ISAC system, which solves the problems of high power consumption caused by the introduction of near-field ELAA in existing technologies, as well as the complexity of interference management in NF-ISAC systems.

[0008] The technical solution adopted by this invention includes the following steps:

[0009] Step 1: Construct an RSMA-assisted NF-ISAC system model based on a very large-scale antenna array (ELAA) with multiple users, multiple inputs, and single outputs (MU-MISO), including: a communication model, a near-field spherical wave channel model, a radar sensing model, and an energy efficiency model.

[0010] Step 2: Under the conditions of satisfying the QoS constraints of system communication users and the CRB constraints of radar sensing performance, a joint optimization problem is constructed for the proposed NF-ISAC system with the goal of maximizing system energy efficiency. This optimization problem also designs the beamforming precoder and the allocation of communication and sensing resources.

[0011] Step 3: Using the idea of ​​Dinkelbach's algorithm, introduce auxiliary variables to equivalently reconstruct the proposed highly nonconvex fractional optimization problem into an easily tractable equivalent form;

[0012] Step 4: In the iterative solution of the optimization problem, the low-complexity continuous convex approximation algorithm SCA is used for convex approximation processing. By optimizing the beamforming matrix of the system and the allocation of user common flow rates, the maximum energy efficiency of the NF-ISAC system and the corresponding optimal beamforming precoder are obtained.

[0013] The system model in step 1 of this invention is specifically as follows:

[0014] Consider an NF-ISAC system with a multi-antenna dual-function base station (BS). The BS simultaneously serves multiple single-antenna communication users and a radar target detection system. The radar and communication subsystems share antenna deployment. The base station's antenna array is a uniform linear array (ULA) with N antennas, positioned along the positive x-axis. The spacing between antenna elements is... The aperture of the antenna is The BS (Base Station) is aware of the system's channel state information and possesses perfect self-cancellation technology. During operation, the BS provides communication services to all communication users via RSMA access, while simultaneously sharing the base station's antennas to transmit detection signals to radar-sensing targets. The number of base station antennas is... , It is the index of the antenna element, the total number of communication users is K, and the user set is represented as... Meanwhile, it is assumed that all communication users and sensing targets are located within the near field, that is, the distance between the communication user or sensing target and the BS is... And satisfying, , The wavelength is the signal wavelength.

[0015] The communication model in step 1 of this invention is specifically as follows:

[0016] In the Time slot, base station to user The sent message is denoted as , Applying Layer 1 RSMA, information for each user is... Split into public message portion and private message section , Public messages for all users Common encoding as public data stream And each user's private messages Encoded separately as independent private data streams After encoding, the final transmitted data stream is represented as follows: The beamforming precoder at the transmitter uses express, , It is a public flow beamforming vector. It is the beamforming vector of user 1's private data stream. It is the beamforming vector of user 2's private data stream. This is the private stream beamforming vector of user K. The beamforming precoder corresponds to the transmitted data stream. The following assumptions are made: 1) All data streams are independent of each other, that is, the data streams satisfy... 2) The transmission power of the communication system is less than the total antenna power. , It is the total antenna power of the system;

[0017] No. The transmission signal of the time-slot base station is ,have:

[0018]

[0019] in, It is the beamforming vector of the public flow. User The private flow beamforming vector, where ; It is a public data stream, and at the same time It also serves as a system sensing signal. This represents the private data stream of user i;

[0020] user The received signal is:

[0021]

[0022] in Indicates the first Near-field channel vectors of each user , It is the conjugate transpose of the near-field channel vector; It is additive white Gaussian noise. It is the environmental noise of user k in the l-th time slot, which has a mean of 0 and a variance of . Additive white Gaussian noise, user The receiving end first decodes the public data stream. Treating the remaining private data streams as interference, and obtaining user data... The signal-to-interference-plus-noise ratio (SINR) for decoding the public stream is:

[0023]

[0024] user When decoding a private data stream, the previously decoded public data stream is first removed using Serial Interference Cancellation (SIC) technology. Then, the user's own private data stream is decoded. At this point, other users' private data streams are considered interference, thus obtaining the user's private data stream. The SINR of its private data stream is:

[0025]

[0026] in, It is the beamforming vector of user k's private data stream;

[0027] According to Shannon's formula, the first The public reachability and private reachability of each user are as follows:

[0028]

[0029]

[0030] All users need to successfully decode the public data stream. The actual achievable rate of the public stream is limited by the public data stream rate of the worst-performing user, i.e.:

[0031]

[0032] remember To be assigned to users If the public rate is given, then the public rate for each user needs to satisfy:

[0033]

[0034] So, the first The final reachability rate for each user is:

[0035]

[0036] Because communication models are typically characterized using system reachability and rate, the system's communication model is as follows: As shown.

[0037] The near-field spherical wave channel model in step 1 of this invention is specifically as follows:

[0038] If the ULA is placed along the positive x-axis, then the coordinates of its nth antenna element can be represented as: ,in Assume that any communication user or sensing target is located in the near field, and their distance relative to the BS is... The angle is Their coordinates are Therefore, the distance of these communication users or sensing targets from the nth antenna element in the ULA is ,have:

[0039]

[0040] in, For ULA Root antenna element, The distance between antenna elements;

[0041] Assume signal transmission uses free-space transmission loss, and the loss factor... So, for users The channel between the nth antenna element in the ULA can be represented as Then we have:

[0042]

[0043] in, Represents the complex gain of the channel. Because it is the signal wavelength, therefore, communication users Near-field channel vector for:

[0044]

[0045] in, It is a near-field device The array response vector of the root antenna ULA, its nth element is represented as:

[0046]

[0047] , and Representing users respectively The distance from the origin, the angle, and the complex gain between them, at this point, the base station and the user Near-field channel vectors between Represented as:

[0048]

[0049] in, It is the array response vector of communication user k;

[0050] make , and Let represent the distance and angle of the target from the origin, and the complex gain of the channel, respectively. Considering the base station operates in monostatic radar sensing mode, then the return channel matrix of the near-field sensing target is... for:

[0051]

[0052] in, It is the array response vector of the perceived target. It is its transpose;

[0053] For the communication subsystem, its near-field spherical wave channel model is as follows: As shown; for the sensing subsystem, its near-field spherical wave channel model is as follows: As shown.

[0054] The radar perception model in step 1 of this invention is specifically as follows:

[0055] Using CRB as a radar sensing indicator, the reflected echo signal obtained at the BS receiving antenna Represented as:

[0056]

[0057] in, The return channel matrix for sensing the target, This represents the background noise of the echo signal received by the base station, with a mean of 0 and a variance of . Additive white Gaussian noise, To obtain parameter information of the perceived target, the base station (BS) needs to estimate the target's parameters over the entire coherent time block using received echo signal samples. , yes The received echo signal accumulated within the time slot makes The signal transmitted within the time slot is , The noise accumulated at the receiver within the time slot is ,So All echo signals received within the time slot are:

[0058]

[0059] To obtain the CRB expression for distance and angle, the equation is... After vectorization, we have:

[0060]

[0061] in, yes The result of vectorization and ,definition If the unknown parameters are to be estimated, then the estimated parameters... The Fisher Information Matrix (FIM) is represented as follows:

[0062]

[0063] Since the CRB matrix expression is the inverse of the FIM matrix, the diagonal elements are the parameters to be estimated, which are then defined. , , For any Each FIM value is:

[0064]

[0065] Through calculation, we have:

[0066]

[0067] So, near-field sensing of target angle and distance information Represented as:

[0068]

[0069]

[0070] Among them, in the formula It is the covariance matrix of the signal;

[0071] The radar perception model is represented by a CRB that has information on the angle and range of the perceived target, i.e., equation Japanese style .

[0072] The energy efficiency model in step 1 of this invention is specifically as follows:

[0073] System energy efficiency is defined as the system's achievable sum rate. Total system power consumption The ratio of , let the system's energy consumption be . Then we have:

[0074]

[0075] in, It is the efficiency coefficient of the base station amplifier; For the static circuit loss power of the base station and the user, there are ,in It is the power consumption for activating the radio frequency link, which is a dynamic power loss; It is static power loss, which is consumed by the cooling system and power supply. It is also the dynamic power loss during transmission, among which This reflects the dynamic loss per unit rate; therefore, the energy efficiency of the entire system is expressed as... Then we have:

[0076]

[0077] Command Divide the numerator and denominator by the same amount. ,have to:

[0078]

[0079] when hour, It is a monotonically decreasing function, therefore, maximizing Equivalent to minimizing Therefore, the denominator of the expression is:

[0080]

[0081] And because It is a positive number, therefore the expression Equivalent to minimizing , The formula is reconstructed as follows:

[0082]

[0083] Therefore, according to the definition of energy efficiency, which is the system's achievable sum rate divided by the system's total power consumption, the system energy efficiency model is as follows: As shown.

[0084] The optimization problem model established in step 2 of this invention is expressed as follows:

[0085] Under constraints of limited resources, communication user QoS, target angle CRB, and target distance CRB, a beamforming precoder design for an RSMA-assisted NF-ISAC system is established with maximizing system energy efficiency as the optimization objective. The optimization problem model for communication and sensing resource allocation is as follows: :

[0086]

[0087] in, It is the set of common flow rates for all communication users. That is the maximum transmission power. These are the QoS requirements of communication users. and Let CRB and CRB represent the threshold values ​​for angle and distance, respectively. (The question is incomplete.) The objective function in the equation represents the ratio of the sum of the system's total speeds to the system's total power consumption, and the constraints... It is a constraint on the system's transmission power; constraint This constraint limits the public data rate for each user, ensuring that the final public message can be successfully decoded by all users; It is the common rate constraint required by RSMA; constraint These are the QoS performance constraints for communication users, which guarantee the performance of the communication subsystem; constraints It is the CRB constraint for the perceived target angle. It is the distance CRB constraint for the perceived target. and This ensured the performance of the radar subsystem.

[0088] Step 3 of this invention equivalently reconstructs the proposed highly nonconvex fractional optimization problem into an easily tractable equivalent form as follows:

[0089] question This is a highly nonconvex fractional problem. We can utilize the idea of ​​Dinkelbach's algorithm and introduce auxiliary variables. , and The problem is refactored into an equivalent form, making it easier to handle:

[0090]

[0091] in, The problem is solved using Dinkelbach's algorithm. The result of reconstruction , and These are constraints generated after reconstructing the original objective function using Dinkelbach's algorithm;

[0092] After refactoring the original problem, the originally highly non-convex fractional objective function has become a convex function, although the original function has been removed. The nonconvexity of the original objective function, the problem It remains non-convex, and its non-convexity is mainly due to the constraints. , , , , and This is caused by the following, and we will use the ideas of SCA and Schur complementary relaxation to deal with these nonconvex terms.

[0093] Step 4 of this invention employs the SCA algorithm to solve the optimization problem in the joint optimization model. To obtain the system's maximum energy efficiency, the steps are as follows:

[0094] 1) To handle non-convex terms , Introducing auxiliary variables To represent the user's private flow rate, then, constraints Represented as:

[0095]

[0096] To handle nonconvex terms Introducing auxiliary variables Let 1 represent the signal-to-interference-plus-noise ratio (SINR) of each communication user's private stream. It becomes:

[0097]

[0098] in, The conversion is based on ,based on Formula calculation Then introduce auxiliary variables , This represents the interference and noise that each user experiences when decoding their own private stream. The expression is converted to:

[0099] After the above steps, the original Use the processed item This can be expressed as:

[0100]

[0101] The same method is used to handle non-convex constraints. Introducing variables , and Let represent the public stream rate of each communication user, the sum of the public stream's signal-to-interference-plus-noise ratio (SINNR) and the interference plus noise experienced by the communication user when decoding the public stream, respectively. The original non-convex constraint... Equivalent to:

[0102]

[0103] The result obtained after the above processing and It remains non-convex; we apply the idea of ​​SCA approximation to process... , and Three non-convex terms;

[0104] for On the left side, in each iteration, a first-order Taylor expansion is used to approximate the lower bound linear function. Alternatively, there are:

[0105]

[0106] in, It is generated in the m-th iteration value;

[0107] for and To the left of the point, the same first-order Taylor expansion approximation method is used at the point. and Use its linear lower bound function instead:

[0108]

[0109]

[0110] For non-convex terms Application processing Auxiliary variables introduced at the time Transform it into:

[0111]

[0112] For the angle CRB and distance CRB constraints of the perceived target, the Schur complementary relaxation method is used to... and Reconstruct it into a convex form that is easy to handle. and Then reconstruct it into the formula Japanese style :

[0113]

[0114]

[0115] At this point, all non-convex terms have been transformed into convex terms, and the entire optimization problem is now transformed into... :

[0116]

[0117] in, The problem The result obtained after performing a non-convex term transformation is: , and These are the equations obtained by transformation. , and , It is a set of convex terms obtained by processing non-convex terms;

[0118] 2) Problem The global optimal solution is found using the interior-point method. This optimization problem is solved using the CVX toolbox, and the problem is solved in each iteration of the SCA algorithm. Assuming the number of iterations is represented by m, then the value from the previous round is used. , , , , , , , Solve the problem to obtain the optimal value for this round. , , , , , , , Then use the optimal value obtained in this round to update , , , , , , , Finally, after m iterations, the maximum energy efficiency is output. ;

[0119] For the initial beamformer It is achieved by finding a constraint that satisfies the constraint. , and Reconstruction and This yields a feasible solution, but to ensure fairness among communication users... Initialize the public flow rate of communication users by using a uniform distribution method. get, , , , , , Calculate using the corresponding formulas respectively. It is the covariance matrix of the signal, through get,

[0120] The objective function is a function that increases rather than decreases with the number of iterations, and it always satisfies... This ensures convergence. The beneficial effects of this invention are:

[0121] 1. This invention, for the first time, uses system energy efficiency as an optimization target in the near field integrated sensing and communication (NF-ISAC) technology field assisted by Rate Splitting Multiple Access (RSMA). Through a rate splitting mechanism, it separates the public and private parts of communication user messages, utilizing the public stream to simultaneously serve communication and sensing functions (without requiring additional dedicated sensing signals), significantly improving the flexibility of interference management and overall system performance. Simultaneously, this invention maximizes system energy efficiency, overcoming the high power consumption problem associated with Extremely Large Antenna Arrays (ELAA). Furthermore, RSMA overcomes the limitations of traditional Space-Division Multiple Access (SDMA) (treating all interference as noise) and Non-Orthogonal Multiple Access (NOMA) (fully decoding interference), providing a smooth transition from viewing interference as noise to fully decoding interference. This achieves more efficient resource utilization and user fairness in complex NF-ISAC interference scenarios, enabling flexible interference management within the NF-ISAC system. Compared with existing ISAC schemes, this invention fully considers near-field channel characteristics (spherical wave model, angle-range coupling) and avoids performance degradation caused by the far-field plane wave assumption. Simultaneously, through flexible interference management via RSMA, it effectively addresses complex scenarios in NF-ISAC such as interference between communication users and interference between communication and sensing beams, improving the system's robustness and practicality in high-frequency bands and ELAA environments.

[0122] 2. This invention, under constraints of transmission power budget, sensing performance CRB threshold, and communication service quality (QoS), jointly optimizes the beamformer and common rate allocation to maximize system energy efficiency (EE). Simulation results show that this scheme improves energy efficiency by approximately 14% compared to traditional SDMA-assisted NF-ISAC systems, and by approximately 64% compared to ISAC schemes based on far-field channel models. Simultaneously, this invention maintains high sensing accuracy and communication QoS requirements, significantly reduces the power consumption burden caused by ELAA, and achieves an excellent trade-off between spectral efficiency, energy efficiency, and sensing performance.

[0123] 3. This invention addresses highly non-convex fractional optimization problems by innovatively employing the Dinkelbach algorithm to process the fractional form of EE, transforming the original problem into an easily tractable equivalent form. Furthermore, it combines this with the SCA algorithm to efficiently solve beamforming and rate allocation for each subproblem within the inner layer. This iterative framework ensures the algorithm's convergence, maintains controllable computational complexity, and is suitable for practical engineering deployments.

[0124] 4. This invention provides an efficient solution for key applications in 6G wireless networks (such as autonomous vehicles, smart cities, and industrial automation), which require simultaneous high-resolution sensing, highly reliable low-latency communication, and high energy efficiency. This technical approach not only improves resource utilization and system sustainability but also lays the foundation for further integration of RSMA and NF-ISAC (such as combining RIS and hybrid beamforming), demonstrating significant theoretical innovation and broad engineering application prospects. Attached Figure Description

[0125] Figure 1 This is a flowchart illustrating the implementation of the present invention;

[0126] Figure 2 This is a system model diagram of the present invention;

[0127] Figure 3 This invention proposes an algorithm that combines the Dinkelbach algorithm and the SCA algorithm;

[0128] Figure 4 This is a schematic diagram showing the convergence of the system proposed in this invention compared with traditional algorithms under different dynamic transmit power consumption conditions;

[0129] Figure 5 This is a schematic diagram showing the EE of the system proposed in this invention compared with traditional algorithms under different transmit antenna conditions;

[0130] Figure 6 This is a schematic diagram illustrating the results of comparing the reachable EE boundaries of different communication users according to the present invention;

[0131] Figure 7 This is a visual diagram showing the results of verifying the sensing performance of this invention;

[0132] Figure 8 This is a graph showing the achievable EE of the present invention compared with traditional algorithms and traditional far-field range systems. Detailed Implementation

[0133] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the present invention.

[0134] This invention addresses the problems existing in the prior art by proposing a high-energy-efficiency beamforming precoder design method for a RateSplitting Multiple Access (RSMA)-assisted Near Field Integrated Sensing and Communication (NF-ISAC) system. This invention considers a base station equipped with an Extremely Large Antenna Array (ELAA) serving multiple communication users and sensing a single target, utilizing a common flow for sensing without introducing a dedicated sensing signal. It constructs a near-field channel model, considering the distance dimension introduced by spherical wave propagation; defines the signal-to-interference-plus-noise ratio (SINR) and sensing performance metrics for communication users, such as the Cramer-Rao Bound (CRB); and jointly optimizes the beamformer and common rate allocation under constraints of transmission power budget, sensing performance threshold, and communication user Quality-of-Service (QoS) to maximize system energy efficiency (EE). This optimization problem is a highly non-convex fractional problem. First, the Dinkelbach algorithm is used for reconstruction, and then the solution is obtained through a beamforming optimization algorithm based on Successive Convex Approximation (SCA). By employing the rate decomposition mechanism of RSMA, a smooth transition from completely viewing interference as noise to completely decoding interference is achieved, significantly improving the flexibility of interference management. Compared to traditional space-division multiple access (SDMA), this invention can improve system EE by approximately 14%; compared to far-field ISAC schemes, it improves EE by approximately 64%, while maintaining perception accuracy and communication QoS. This invention is applicable to 6G and future wireless networks, supporting application scenarios such as autonomous driving and smart cities, and has high practicality and promotional value.

[0135] See Figure 1 The steps include:

[0136] Step 1: Construct an ISAC system model based on ELAA (Multi-User Multiple-Input Single-Output, MU-MISO) with RSMA assistance in a near-field environment, including a communication model, a near-field spherical wave channel model, a radar sensing model, and an energy efficiency model:

[0137] System Model:

[0138] See Figure 2 Consider an NF-ISAC system with multiple dual-function base stations (BS). Each BS simultaneously serves multiple single-antenna communication users and a radar-detected target. The radar and communication subsystems share antenna deployment. The base station's antenna array is a uniform linear array (ULA) with N antennas, positioned along the positive x-axis. The spacing between antenna elements is... Therefore, the achievable aperture of the antenna is... The Base Station (BS) is aware of the system's channel state information and possesses perfect self-cancellation technology. During operation, the BS provides communication services to all communication users via RSMA access, while simultaneously sharing the base station's antennas to transmit detection signals to radar-sensing targets. The number of base station antennas is... , This is the index of the antenna element. Furthermore, the total number of communication users is K, meaning the user set is represented as... Meanwhile, it is assumed that all communication users and sensing targets are located within the near field, i.e., the distance between the communication user or sensing target and the BS is... And satisfying, , The wavelength is the signal wavelength.

[0139] Communication model:

[0140] In the Time slot, base station to user The sent message is denoted as , Applying Layer 1 RSMA, messages from each user are... Decomposed into users Public information section and private message section ,Right now All public messages Common encoding as public data stream And each user's private messages Encoded separately as independent private data streams After encoding, the final transmitted data stream can be represented as follows: The beamforming pre-encoder at the transmitter uses... express, , It is a public flow beamforming vector. It is the beamforming vector of user 1's private data stream. It is the beamforming vector of user 2's private data stream. This is the private stream beamforming vector of user K. That is, the precoder corresponds to the transmitted data stream. This invention makes the following assumptions: 1) Each data stream is independent, i.e., the data stream satisfies... 2) The transmission power of the communication system is less than the total antenna power. , It is the total power of the system antenna.

[0141] Therefore, the first The transmitted signal of each time slot base station is Then we have:

[0142]

[0143] in, It is the beamforming vector of the public flow. User The private flow beamforming vector, where, ; It is a public data stream, and at the same time It also serves as a system sensing signal. This represents the private data stream of user i.

[0144] user The received signal is:

[0145]

[0146] in Indicates the first Near-field channel vectors of each user , It is the conjugate transpose of the near-field channel vector; It is additive white Gaussian noise. It is the environmental noise of user k in the l-th time slot, which has a mean of 0 and a variance of . Additive white Gaussian noise. (User) The receiving end first decodes the public data stream. By treating the remaining private data streams as interference, the user's... The signal-to-interference-plus-noise ratio (SINR) for decoding the public stream is:

[0147]

[0148] user When decoding a private data stream, firstly, the previously decoded public data stream is removed using Successive Interference Cancellation (SIC). Then, the user's own private data stream is decoded, treating other users' private data streams as interference. The SINR of its private data stream is:

[0149]

[0150] in, It is the beamforming vector of user k's private data stream.

[0151] According to Shannon's formula, the first Public reachability for individual users and private reachability for:

[0152]

[0153]

[0154] Since all users need to successfully decode the public data stream, the actual public rate under the entire RSMA system is:

[0155]

[0156] remember Indicates user Given the actual public rate, the public rate for each user needs to satisfy:

[0157]

[0158] So, the first The final reachability rate for each user is:

[0159]

[0160] Because communication models are typically characterized using system reachability and rate, the system's communication model is as follows: As shown.

[0161] Channel model:

[0162] Due to the use of ELAA and high-frequency bandwidth, the Rayleigh distance has been extended to tens or even hundreds of meters, causing a near-field effect in the communication system. The propagation mode of electromagnetic waves changes from a plane wave in the far field to a spherical wave in the near field. For example, when the BS operates with 65 antennas at 28 GHz, the antenna aperture... When m, the Rayleigh distance of the system can be extended to =46.73 m. As described in the system model of claim 2, the ULA is placed along the positive half of the x-axis, then the coordinates of its nth antenna element can be expressed as... ,in Assume that any communication user (or sensing target) is located within the near field, and their distance relative to the BS is... The angle is Therefore, their coordinates are: Therefore, the distance between these communication users (or sensing targets) and the nth antenna element in the ULA is . Then we have:

[0163]

[0164] in, For ULA Root antenna element, This represents the distance between antenna elements.

[0165] Assume signal transmission uses free-space transmission loss, and the loss factor... So, for users... The channel between the nth antenna element in the ULA can be represented as Then we have:

[0166]

[0167] in, Represents the complex gain of the channel. This is the signal wavelength. Therefore, communication users... Near-field channel vector for:

[0168]

[0169] in, It is a near-field device The array response vector of the root antenna ULA, its nth element can be expressed as:

[0170]

[0171] We let , and Representing users respectively Given the distance from the origin, the angle, and the complex gain between them, then the distance between the base station and the user... Near-field channel vectors between It can be represented as:

[0172]

[0173] in, It is the array response vector of communication user k.

[0174] Compared to communication, radar sensing requires base stations to collect the echo signals from the sensed target; therefore, the round-trip channel needs to be considered. , and Let represent the distance and angle of the target from the origin, and the complex gain of the channel, respectively. Considering the base station operates in monostatic radar sensing mode, then the return channel matrix of the near-field sensing target is... for:

[0175]

[0176] in, It is the array response vector of the perceived target. It is its transpose.

[0177] Therefore, for the communication subsystem, its near-field spherical wave channel model is as follows: As shown; for the sensing subsystem, its near-field spherical wave channel model is as follows: As shown.

[0178] Radar perception model:

[0179] Traditionally, the distance and angle estimates of the perceived target are used. Angle value relative to actual distance The minimum mean square error between the two is used as a metric for sensing performance. and However, this problem is difficult to solve and not easily obtained. and The closed-form solution. Therefore, in this invention, we use CRB as the radar sensing index. CRB provides a theoretical lower bound for the variance of unbiased estimators, including target parameters such as range, velocity, and angle, and has become a fundamental standard for evaluating the accuracy and resolution of radar systems in recent years. By obtaining the smallest possible CRB, the system's ability to accurately estimate target features can be enhanced, ensuring robust and reliable sensing performance. The reflected echo signal obtained at the BS receiving antenna. Represented as:

[0180]

[0181] in, It is the target feedback channel matrix. This represents the background noise of the echo signal received by the base station, with a mean of 0 and a variance of . Additive white Gaussian noise, To obtain parameter information of the perceived target, the base station (BS) needs to estimate the target's parameters over the entire coherent time block using received echo signal samples. , yes The received echo signal accumulated within the time slot. Then, if we let... The signal transmitted within the time slot is , The noise accumulated at the receiver within the time slot is ,So All echo signals received within the time slot are:

[0182]

[0183] To obtain the CRB expression for distance and angle, we... After vectorization, we have:

[0184]

[0185] in, yes The result of vectorization and We define If the unknown parameters are to be estimated, then the estimated parameters... The Fisher Information Matrix (FIM) can be represented as:

[0186]

[0187] Since the CRB matrix expression is the inverse of the FIM matrix, the diagonal elements are the parameters we need to estimate. Then, we define... , , For any Each FIM value is,

[0188]

[0189] Through calculation, we have:

[0190]

[0191] So, near-field sensing of target angle and distance information It can be represented as:

[0192]

[0193]

[0194] Among them, in the formula It is the covariance matrix of the signal.

[0195] Therefore, the radar perception model is represented by a CRB that has information on the angle and range of the perceived target, i.e., equation Japanese style .

[0196] Energy efficiency model:

[0197] System energy efficiency is defined as the system's achievable sum rate. Total system power consumption The ratio of . Let the system's energy consumption be . Then we have:

[0198]

[0199] in, It is the efficiency coefficient of the base station amplifier; For the static circuit loss power of the base station and the user, there are ,in It is the power consumption for activating the radio frequency link, which is a dynamic power loss; It is static power loss, which is consumed by the cooling system, power supply, etc. It is the dynamic power loss during transmission, among which This reflects the dynamic loss per unit rate. Therefore, the energy efficiency of the entire system can be expressed as: Then we have:

[0200]

[0201] Command Divide the numerator and denominator by the same amount. We can obtain:

[0202]

[0203] when hour, It is a monotonically decreasing function. Therefore, maximizing Equivalent to minimizing The denominator of the expression. Therefore, we have:

[0204]

[0205] And because It is a positive number, therefore the expression Equivalent to minimizing , The formula is reconstructed as follows:

[0206]

[0207] Therefore, according to the definition of energy efficiency, which is the system's achievable sum rate divided by the system's total power consumption, the system energy efficiency model is as follows: As shown.

[0208] Step 2: Under the conditions of satisfying the system communication user quality of service (QoS) constraints and radar sensing performance Cramér-Rao Bound (CRB) constraints, for the proposed NF-ISAC system, a joint optimization problem is constructed with the goal of maximizing system energy efficiency. This optimization problem also designs the beamforming precoder and the allocation of communication and sensing resources.

[0209] In a near-field environment, an RSMA-assisted MU-MISO-based ISAC system model is established to optimize the design of a high-efficiency beamforming precoder and the allocation of sensing resources, under constraints of limited resources, transmit power, communication user QoS, and radar sensing estimation angle and range CRB, in order to maximize the system's energy efficiency.

[0210] Under constraints of limited resources, communication user QoS, target angle CRB, and target distance CRB, a beamforming precoder design for an RSMA-assisted NF-ISAC system is established with maximizing system energy efficiency as the optimization objective. The optimization problem model for communication and sensing resource allocation is as follows: :

[0211]

[0212] in, It is the set of common flow rates for all communication users. That is the maximum transmission power. These are the QoS requirements of communication users. and These represent the threshold values ​​for angle CRB and distance CRB, respectively. (Question) The objective function in the equation represents the ratio of the sum of the system's total speeds to the system's total power consumption, and the constraints... It is a constraint on the system's transmission power; constraint This constraint limits the public data rate for each user, ensuring that the final public message can be successfully decoded by all users; It is the common rate constraint required by RSMA; constraint These are the QoS performance constraints for communication users, which guarantee the performance of the communication subsystem; constraints It is the CRB constraint for the perceived target angle. It is the distance CRB constraint for the perceived target. and This ensured the performance of the radar subsystem.

[0213] See Figure 3 Step 3: Using the idea of ​​Dinkelbach's algorithm, introduce auxiliary variables to equivalently reconstruct the proposed highly nonconvex fractional optimization problem into an easily tractable equivalent form;

[0214] Dinkelbach's algorithm is a classic method for solving fractional programming problems. Its core idea is to transform the original fractional programming problem into a series of parameterized subproblems. Through iterative updates, the algorithm gradually approximates the optimal function value. This algorithm exhibits superlinear convergence (and even local quadratic convergence) and has been widely validated in beamforming, power control, and energy efficiency optimization in wireless communication. It can efficiently transform fractional problems into equivalent non-fractional optimization problems, facilitating subsequent processing.

[0215] The steps are as follows:

[0216] question This is a highly nonconvex fractional problem. We can utilize the idea of ​​Dinkelbach's algorithm and introduce auxiliary variables. , and The problem is refactored into an equivalent form, making it easier to handle:

[0217]

[0218] in, The problem is solved using Dinkelbach's algorithm. The result of reconstruction , and These are constraints generated after reconstructing the original objective function using Dinkelbach's algorithm.

[0219] After refactoring the original problem, the originally highly non-convex fractional objective function has become a convex function. Although the original problem was removed... The nonconvexity of the original objective function, the problem It remains non-convex, and its non-convexity is mainly due to the constraints. , , , , and This is caused by [the following]. We will use the concepts of SCA and Schur complementary relaxation to handle these nonconvex terms.

[0220] Step 4: In the iterative solution of the optimization problem, the low-complexity continuous convex approximation algorithm SCA is used for convex approximation processing. By optimizing the beamforming matrix of the system and the allocation of the user common flow rate, the maximum energy efficiency of the NF-ISAC system and the corresponding optimal beamforming precoder are obtained.

[0221] SCA (Solving Problems with Local Convexity) is an iterative optimization method that transforms a non-convex problem into a convex one by locally convexizing the non-convex objective function or constraints in each iteration, thus gradually approximating the optimal solution. The basic idea of ​​SCA is to approximate the objective function or constraints at the current iteration point, typically using Taylor expansion or other local linearization techniques to transform it into a convex function. Then, in each iteration, this locally convex optimization problem is solved, and the current optimal solution is used as the starting point for the next iteration. SCA is particularly suitable for optimization problems that are globally non-convex but locally convex. The advantages of SCA are that it can progressively optimize non-convex problems through local convexity methods, is suitable for handling problems with strong local convexity, and can flexibly handle different types of constraints and objective functions. Furthermore, SCA can effectively handle high-dimensional complex problems and continuously improves the quality of the solution during the iteration process.

[0222] This invention, for the first time, uses energy efficiency as an optimization metric in an RSMA-assisted NF-ISAC system. It combines Dinkelbach's algorithm, SCA algorithm, and Schur complement algorithm to solve the optimization problem in the joint optimization model, obtaining the maximum energy efficiency of the system, as detailed below:

[0223] 1) To handle non-convex terms , Introducing auxiliary variables This represents the user's private stream rate. Therefore, the constraints... It can be represented as:

[0224]

[0225] To handle nonconvex terms Introducing auxiliary variables Let 1 represent the signal-to-interference-plus-noise ratio (SINR) of each communication user's private stream. It can be transformed into:

[0226]

[0227] in, The conversion is based on .based on Formula calculation Then introduce auxiliary variables , This represents the interference and noise that each user experiences when decoding their own private stream. Thus, The expression is converted to:

[0228]

[0229] After the above steps, the original The processed item can be used To represent. That is, equivalent to:

[0230]

[0231] The same method is used to handle non-convex constraints. Introducing variables , and Let represent the public stream rate of each communication user, the sum of the public stream's signal-to-interference-plus-noise ratio (SINNR) and the interference plus noise experienced by the communication user when decoding the public stream, respectively. Then, the original non-convex constraint... This can be equivalent to:

[0232]

[0233] The result obtained after the above processing and It remains nonconvex. We apply the idea of ​​SCA approximation to handle... , and Three non-convex terms.

[0234] for On the left side, in each iteration, a first-order Taylor expansion is used to approximate the lower bound linear function. Alternatively, there are:

[0235]

[0236] in, It is generated in the m-th iteration value.

[0237] for and To the left of the point, the same first-order Taylor expansion approximation method is used at the point. and Use its linear lower bound function instead:

[0238]

[0239]

[0240] For non-convex terms Application processing Auxiliary variables introduced at the time Transform it into:

[0241]

[0242] For the angle CRB constraint and distance CRB constraint of the perceived target, we use the Schur complementary relaxation concept to... and The purpose of the reconstruction is to transform it into a convex form that is easier to handle. and Then reconstruct it into the formula Japanese style :

[0243]

[0244]

[0245] At this point, all non-convex terms have been transformed into convex terms, and the entire optimization problem is now transformed into... :

[0246]

[0247] in, The problem The result obtained after performing a non-convex term transformation is: , and These are the expressions obtained by transformation. , and . It is the set of convex terms obtained by processing non-convex terms.

[0248] 2) Problem The global optimal solution can be found using the interior-point method, and this optimization problem can be solved using the CVX toolbox. The problem is solved in each iteration of the SCA algorithm. Assuming the number of iterations is represented by m, then the value from the previous round is used. , , , , , , Solve the problem to obtain the optimal value for this round. , , , , , , , Then use the optimal value obtained in this round to update , , , , , , , Finally, after m iterations, the maximum energy efficiency is output. .

[0249] For the initial beamformer It is achieved by finding a constraint that satisfies the constraint. , and Reconstruction and A feasible solution was obtained. To ensure fairness among communication users, Initialize the public flow rate of communication users by using a uniform distribution method. get. , , , , , They can be calculated using the corresponding formulas. It is the covariance matrix of the signal, through get.

[0250] Furthermore, since this invention uses a first-order lower bound approximation, the objective function is a function that increases rather than decreases with the number of iterations, and it always satisfies... In other words, the algorithm in this invention is guaranteed to converge.

[0251] In summary, the algorithm combining the Dinkelbach algorithm and SCA was used to achieve the maximum energy efficiency in the NF-ISAC system.

[0252] The technical effects of the present invention will be further explained in detail below with reference to simulation experiments: The near-field sensing integrated system based on rate splitting multiple access proposed in this invention, namely RSMA-assisted NF-ISAC, will be compared with existing solutions.

[0253] (1) Simulation parameter settings

[0254] In the simulation experiment, the simulation system includes a dual-function base station with a uniform linear array of antennas. Let the antenna aperture be If the carrier frequency is 28 GHz, then the Rayleigh distance is... Number of users The QoS requirement for each communication user, i.e., the minimum communication rate, is... The number of radar targets is The distance and direction of the radar target to BS are: And the angle CRB threshold and the distance CRB threshold are respectively and The power amplifier efficiency is Maximum transmission power The static power consumption is 30 dBm, and the dynamic power consumption is 20 dBm; the convergence accuracy of the algorithm is... To avoid loss of generality, we assume unit noise variance and unit bandwidth. As a basic comparison scheme, we use common flow... Setting it to zero degrades RSMA to traditional SDMA. For far-field contrast schemes, the channel model considered is a plane wave model:

[0255]

[0256] in, For the signal wavelength, The distance between the antenna elements in the ULA array. The number of antenna elements. This indicates that a transpose operation is being performed.

[0257] (2) Simulation content and result analysis

[0258] Figure 4 This is a convergence graph of the proposed RSMA-based beamforming optimization algorithm and SDMA-based beamforming optimization algorithm in EE optimization of the NF-ISAC system under different dynamic power consumption levels. Figure 4 As can be observed, both schemes exhibit extremely fast convergence speeds, typically achieving convergence within 5–10 iterations, and are unaffected by the dynamic power consumption coefficient, fully validating the efficiency of the proposed algorithm. When the dynamic power consumption is 20 dBm, the EE of the RSMA scheme stabilizes at approximately 2.32 bits / J / Hz, while the SDMA scheme converges to 2.04 bits / J / Hz, with the relative gain of RSMA being approximately 13.75%. With the dynamic power consumption... Increasing the dynamic power consumption (EE) from 10 dBm to 20 dBm significantly reduced the EE of both schemes, but the EE value of the RSMA scheme was consistently better than that of the SDMA scheme. This trend is in line with expectations; a larger dynamic power consumption significantly increases circuit power consumption, thereby increasing the denominator of the EE index, while the achievable sum rate is not affected by this coefficient. Therefore, as the dynamic power consumption coefficient increases, the overall EE tends to decrease.

[0259] Figure 5This is a graph showing the EE (Electrical Energy) variation trend of the RSMA-assisted and SDMA-assisted NF-ISAC systems proposed in this invention with the number of transmit antennas, where the dynamic power consumption is fixed at 20 dBm. From... Figure 5 It can be observed that as the number of transmitting antennas increases... From 17 to 33 and then to 65, the EE (Effective Energy) of both schemes showed a downward trend. Specifically, the EE of the RSMA scheme decreased from approximately 3.73 bits / J / Hz to 2.32 bits / J / Hz, and further to 1.39 bits / J / Hz. This trend was in line with expectations, mainly due to dynamic power consumption. The spectral efficiency increases linearly with the number of antennas. Although additional antennas improve spectral efficiency by enhancing interference management and near-field beam focusing capabilities, the logarithmic characteristics that enable rate gain cannot fully offset the linear increase in total power consumption. Therefore, the overall energy efficiency (EE) decreases, highlighting the inherent trade-off between spectral efficiency and energy efficiency in large-scale antenna NF-ISAC systems. This result further validates the superiority of the proposed RSMA scheme across different antenna sizes. Although the EE decreases linearly with the number of antennas... While power consumption increases and decreases, RSMA consistently maintains a significant advantage over SDMA, demonstrating its unique value in managing complex near-field interference and efficiently utilizing resources. This trend also underscores the necessity of comprehensively considering power consumption models and beamforming design in ELAA deployments, providing an important reference for optimizing practical 6G NF-ISAC systems.

[0260] Figure 6 This is a view of the implementable EE region for communication users in the system designed in this invention. The Pareto boundary is obtained through a weighted sum method, where the weight coefficients are fixed. and in the weighting coefficient Variations within a range. From Figure 6 As can be observed, the RSMA scheme significantly outperforms the SDMA benchmark across the entire EE plane. Specifically, the RSMA scheme achieves an overall EE improvement of approximately 14%. These results demonstrate that employing RSMA in NF-ISAC systems provides system designers with greater flexibility to better balance multi-user performance with overall energy efficiency.

[0261] Figure 7 This invention presents a normalized three-dimensional MUSIC spatial spectrum diagram of an RSMA-assisted NF-ISAC system. The common flow obtained through rate splitting is directly multiplexed as the sensing signal, eliminating the need for any dedicated sensing signal overhead. This verifies the effectiveness of maintaining high-precision sensing performance while maximizing EE. The base station is located at the origin of the coordinate system. The search grid is defined in and Within range. The actual target location is... That is, Cartesian coordinates are (because The minimum communication rate constraint is set to 0.5 bits / s / Hz, and the number of antennas... Dynamic power consumption .from Figure 7 It can be clearly observed that the color gradient from blue to red indicates that the normalized spectral value gradually increases. A sharp spectral peak appears at this location, which corresponds to the polar coordinates of the true target. Precise alignment. This result demonstrates that the proposed RSMA-assisted NF-ISAC beamforming design achieves high-precision target localization while maximizing EE, without requiring additional dedicated sensing signal overhead, thus verifying the efficiency and feasibility of effectively integrating communication and sensing functions.

[0262] Figure 8 This invention presents an energy efficiency diagram comparing RSMA and SDMA-assisted ISAC systems in a traditional far-field plane wave environment. To verify the advantages of NF spherical wave propagation and ensure a fair comparison, we kept all simulation parameters identical in both the far-field and near-field environments, differing only in channel modeling for NF and FF scenarios. Figure 8 We observed in the number of antennas Dynamic power consumption Under this configuration, the NF-RSMA scheme proposed in this invention achieves an EE of approximately 2.3 bits / J / Hz, which is about 13% higher than NF-SDMA (approximately 2.03 bits / J / Hz) and about 64% higher than FF-RSMA (approximately 1.4 bits / J / Hz). This superior performance stems from the rate splitting mechanism of RSMA, which effectively suppresses inter-user interference through a shared flow, combined with the significant array gain brought by near-field beamforming. Under far-field conditions, the EEs of FF-RSMA and FF-SDMA converge to lower levels of approximately 1.4 bits / J / Hz and 1.8 bits / J / Hz, respectively. This is because the plane wave assumption limits the beamforming accuracy, leading to higher energy consumption. It is particularly noteworthy that even NF-SDMA outperforms FF-RSMA in this configuration, further highlighting the inherent advantages of near-field propagation in ISAC systems.

[0263] In summary, Figure 4 and Figure 5 This invention effectively demonstrates the effectiveness and stability of the proposed algorithm under different system parameter configurations, and shows that the proposed algorithm can converge quickly within a relatively small number of iterations. Figure 6Furthermore, it is shown that the proposed RSMA-assisted NF-ISAC scheme consistently outperforms the traditional SDMA benchmark scheme within the user energy efficiency region. Under the same resource constraints, it can achieve more flexible and efficient joint allocation of communication and sensing resources, thereby improving the overall system performance and taking into account fairness among multiple users. Figure 7 The study verified that, under the premise of maximizing energy efficiency as the optimization objective, the RSMA-assisted NF-ISAC system proposed in this invention can still achieve high-precision target perception without introducing additional dedicated sensing signal overhead, demonstrating the deep integration capability of communication and sensing functions. Figure 8 Further verification demonstrates that the method of this invention exhibits significant performance advantages over traditional far-field ISAC schemes in near-field propagation environments, demonstrating higher energy efficiency and stronger system robustness. Therefore, it can be concluded that the proposed RSMA-assisted high-energy-efficiency beamforming precoding design method for near-field ISAC systems can achieve higher energy efficiency, better resource utilization, and more reliable communication sensing performance in complex near-field interference environments, effectively alleviating the problems of limited system power consumption and scarce spectrum resources, and significantly improving the overall performance of NF-ISAC systems.

[0264] This invention employs an ultra-large-scale antenna array (ELAA) and an ultra-high frequency band scenario. Through RSMA, flexible interference management can be achieved. By jointly optimizing the beamformer and the common rate allocation for users, the system energy efficiency can be maximized while meeting the constraints of transmission power, sensing performance Cramer-Rao boundary (CRB), and communication user quality of service (QoS).

Claims

1. A method for designing a high energy-efficient beamforming precoder for RSMA-assisted NF-ISAC systems, characterized in that, Includes the following steps: Step 1: Construct an RSMA-assisted NF-ISAC system model based on a very large-scale antenna array (ELAA) with multiple users, multiple inputs, and single outputs (MU-MISO), including: a communication model, a near-field spherical wave channel model, a radar sensing model, and an energy efficiency model. Step 2: Under the conditions of satisfying the QoS constraints of system communication users and the CRB constraints of radar sensing performance, a joint optimization problem is constructed for the proposed NF-ISAC system with the goal of maximizing system energy efficiency. This optimization problem also designs the beamforming precoder and the allocation of communication and sensing resources. Step 3: Using the idea of ​​Dinkelbach's algorithm, introduce auxiliary variables to equivalently reconstruct the proposed highly nonconvex fractional optimization problem into an easily tractable equivalent form; Step 4: In the iterative solution of the optimization problem, the low-complexity continuous convex approximation algorithm SCA is used for convex approximation processing. By optimizing the beamforming matrix of the system and the allocation of user common flow rates, the maximum energy efficiency of the NF-ISAC system and the corresponding optimal beamforming precoder are obtained.

2. The RSMA-assisted NF-ISAC system high energy-efficient beamforming precoder design method of claim 1, wherein, The system model in step 1 is specifically as follows: Consider a NF-ISAC system with a multi-antenna dual-functional base station BS, which serves multiple single-antenna communication users and a radar sensing target simultaneously. The radar and communication subsystems share the antenna deployment. The antenna array of the base station is a uniform linear array (ULA) with N antennas, and the ULA is placed along the positive half-axis of the x-axis with an inter-element spacing of . The aperture of the antenna is . The BS can know the channel state information of the system and has perfect self-cancellation technology. During the operation of the BS, all communication users are provided with communication services through the RSMA access mode, and the antennas of the base station transmit probe signals to the radar sensing target. The number of antennas of the base station is , is the index of the antenna element, the total number of communication users is K, and the user set is represented as . At the same time, it is assumed that all communication users and the sensing target are located within the near-field range, i.e., it is assumed that the distance between the communication users or the sensing target and the BS is , and it satisfies , is the signal wavelength.

3. The design method for a high-efficiency beamforming pre-encoder for an RSMA-assisted NF-ISAC system according to claim 1 or 2, characterized in that, The communication model in step 1 is as follows: In the Time slot, base station to user The sent message is denoted as , Applying Layer 1 RSMA, information for each user is... Split into public message portion and private message section , Public messages for all users Common encoding as public data stream And each user's private messages Encoded separately as independent private data streams After encoding, the final transmitted data stream is represented as follows: The beamforming precoder at the transmitter uses express, , It is a public flow beamforming vector. It is the beamforming vector of user 1's private data stream. It is the beamforming vector of user 2's private data stream. This is the private stream beamforming vector of user K. The beamforming precoder corresponds to the transmitted data stream. The following assumptions are made: 1) All data streams are independent of each other, that is, the data streams satisfy... 2) The transmission power of the communication system is less than the total antenna power. , It is the total antenna power of the system; No. The transmission signal of the time-slot base station is ,have: ； in, It is the beamforming vector of the public flow. User The private flow beamforming vector, where ; It is a public data stream, and at the same time It also serves as a system sensing signal. This represents the private data stream of user i; user The received signal is: ； in Indicates the first Near-field channel vectors of each user , It is the conjugate transpose of the near-field channel vector; It is additive white Gaussian noise. It is the environmental noise of user k in the l-th time slot, which has a mean of 0 and a variance of . Additive white Gaussian noise, user The receiving end first decodes the public data stream. Treating the remaining private data streams as interference, and obtaining user data... The signal-to-interference-plus-noise ratio (SINR) for decoding the public stream is: ； user When decoding a private data stream, the previously decoded public data stream is first removed using Serial Interference Cancellation (SIC) technology. Then, the user's own private data stream is decoded. At this point, other users' private data streams are considered interference, thus obtaining the user's private data stream. The SINR of its private data stream is: ； in, It is the beamforming vector of user k's private data stream; According to Shannon's formula, the first The public reachability and private reachability of each user are as follows: ； ； All users need to successfully decode the public data stream. The actual achievable rate of the public stream is limited by the public data stream rate of the worst-performing user, i.e.: ； remember To be assigned to users If the public rate is given, then the public rate for each user needs to satisfy: ； So, the first The final reachability rate for each user is: ； Because communication models are typically characterized using system reachability and rate, the system's communication model is as follows: As shown.

4. The design method for a high-efficiency beamforming pre-encoder for an RSMA-assisted NF-ISAC system according to claim 2, characterized in that, The near-field spherical wave channel model in step 1 is specifically as follows: If the ULA is placed along the positive x-axis, then the coordinates of its nth antenna element can be represented as: ,in Assume that any communication user or sensing target is located in the near field, and their distance relative to the BS is... The angle is Their coordinates are Therefore, the distance of these communication users or sensing targets from the nth antenna element in the ULA is ,have: ； in, For ULA Root antenna element, The distance between antenna elements; Assume signal transmission uses free-space transmission loss, and the loss factor... So, for users The channel between the nth antenna element in the ULA can be represented as Then we have: ； in, Represents the complex gain of the channel. Because it is the signal wavelength, therefore, communication users Near-field channel vector for: ； in, It is a near-field device The array response vector of the root antenna ULA, its nth element is represented as: ； , and Representing users respectively The distance from the origin, the angle, and the complex gain between them, at this point, the base station and the user Near-field channel vectors between Represented as: ； in, It is the array response vector of communication user k; make , and Let represent the distance and angle of the target from the origin, and the complex gain of the channel, respectively. Considering the base station operates in monostatic radar sensing mode, then the return channel matrix of the near-field sensing target is... for: ； in, It is the array response vector of the perceived target. It is its transpose; For the communication subsystem, its near-field spherical wave channel model is as follows: As shown; for the sensing subsystem, its near-field spherical wave channel model is as follows: As shown.

5. The design method for a high-efficiency beamforming pre-encoder for an RSMA-assisted NF-ISAC system according to claim 1 or 2, characterized in that, The radar perception model in step 1 is specifically as follows: Using CRB as a radar sensing indicator, the reflected echo signal obtained at the BS receiving antenna Represented as: ； in, The return channel matrix for sensing the target, This represents the background noise of the echo signal received by the base station, with a mean of 0 and a variance of . Additive white Gaussian noise, To obtain parameter information of the perceived target, the base station (BS) needs to estimate the target's parameters over the entire coherent time block using received echo signal samples. , yes The received echo signal accumulated within the time slot makes The signal transmitted within the time slot is , The noise accumulated at the receiver within the time slot is ,So All echo signals received within the time slot are: ； To obtain the CRB expression for distance and angle, the equation is... After vectorization, we have: ； in, yes The result of vectorization and ,definition If the unknown parameters are to be estimated, then the estimated parameters... The Fisher Information Matrix (FIM) is represented as follows: ； Since the CRB matrix expression is the inverse of the FIM matrix, the diagonal elements are the parameters to be estimated, which are then defined. , , For any Each FIM value is: ； Through calculation, we have: ； So, near-field sensing of target angle and distance information Represented as: ； ； Among them, in the formula It is the covariance matrix of the signal; The radar perception model is represented by a CRB that has information on the angle and range of the perceived target, i.e., equation Japanese style .

6. The design method for a high-efficiency beamforming pre-encoder for an RSMA-assisted NF-ISAC system according to claim 1 or 2, characterized in that, The energy efficiency model in step 1 is as follows: System energy efficiency is defined as the system's achievable sum rate. Total system power consumption The ratio of , let the system's energy consumption be . Then we have: ； in, It is the efficiency coefficient of the base station amplifier; For the static circuit loss power of the base station and the user, there are ,in It is the power consumption for activating the radio frequency link, which is a dynamic power loss; It is static power loss, which is consumed by the cooling system and power supply. It is also the dynamic power loss during transmission, among which This reflects the dynamic loss per unit rate; therefore, the energy efficiency of the entire system is expressed as... Then we have: ； Command Divide the numerator and denominator by the same amount. ,have to: ； when hour, It is a monotonically decreasing function, therefore, maximizing Equivalent to minimizing Therefore, the denominator of the expression is: ； And because It is a positive number, therefore the expression Equivalent to minimizing , The formula is reconstructed as follows: ； Therefore, according to the definition of energy efficiency, which is the system's achievable sum rate divided by the system's total power consumption, the system energy efficiency model is as follows: As shown.

7. The design method for a high-efficiency beamforming pre-encoder for an RSMA-assisted NF-ISAC system according to claim 1, characterized in that, The optimization problem model established in step 2 is expressed as follows: Under constraints of limited resources, communication user QoS, target angle CRB, and target distance CRB, a beamforming precoder design for an RSMA-assisted NF-ISAC system is established with maximizing system energy efficiency as the optimization objective. The optimization problem model for communication and sensing resource allocation is as follows: : ； in, It is the set of common flow rates for all communication users. That is the maximum transmission power. These are the QoS requirements of communication users. and Let CRB and CRB represent the threshold values ​​for angle and distance, respectively. (The question is incomplete.) The objective function in the equation represents the ratio of the sum of the system's total speeds to the system's total power consumption, and the constraints... It is a constraint on the system's transmission power; constraint This constraint limits the public data rate for each user, ensuring that the final public message can be successfully decoded by all users; It is the common rate constraint required by RSMA; constraint These are the QoS performance constraints for communication users, which guarantee the performance of the communication subsystem; constraints It is the CRB constraint for the perceived target angle. It is the distance CRB constraint for the perceived target. and This ensured the performance of the radar subsystem.

8. The design method for a high-efficiency beamforming pre-encoder for an RSMA-assisted NF-ISAC system according to claim 1, characterized in that, Step 3 equivalently reconstructs the proposed highly nonconvex fractional optimization problem into an easily tractable equivalent form as follows: question This is a highly nonconvex fractional problem. We can utilize the idea of ​​Dinkelbach's algorithm and introduce auxiliary variables. , and The problem is refactored into an equivalent form, making it easier to handle: ； in, The problem is solved using Dinkelbach's algorithm. The result of reconstruction , and These are constraints generated after reconstructing the original objective function using Dinkelbach's algorithm; After refactoring the original problem, the originally highly non-convex fractional objective function has become a convex function, although the original function has been removed. The nonconvexity of the original objective function, the problem It remains non-convex, and its non-convexity is mainly due to the constraints. , , , , and This is caused by the following, and we will use the ideas of SCA and Schur complementary relaxation to deal with these nonconvex terms.

9. The design method for a high-efficiency beamforming pre-encoder for an RSMA-assisted NF-ISAC system according to claim 8, characterized in that, Step 4 employs the SCA algorithm to solve the optimization problem in the joint optimization model. To obtain the system's maximum energy efficiency, the steps are as follows: 1) To handle non-convex terms , Introducing auxiliary variables To represent the user's private flow rate, then, constraints Represented as: ； To handle nonconvex terms Introducing auxiliary variables Let 1 represent the signal-to-interference-plus-noise ratio (SIR) of each communication user's private stream. It becomes: ； in, The conversion is based on ,based on Formula calculation Then introduce auxiliary variables , This represents the interference and noise that each user experiences when decoding their own private stream. The expression is converted to: ; After the above steps, the original Use the processed item This can be expressed as: ； The same method is used to handle non-convex constraints. Introducing variables , and Let represent the public stream rate of each communication user, the sum of the public stream's signal-to-interference-plus-noise ratio (SINNR) and the interference plus noise experienced by the communication user when decoding the public stream, respectively. The original non-convex constraint... Equivalent to: ； The result obtained after the above processing and It remains non-convex; we apply the idea of ​​SCA approximation to process... , and Three non-convex terms; for On the left side, in each iteration, a first-order Taylor expansion is used to approximate the lower bound linear function. Alternatively, there are: ； in, It is generated in the m-th iteration value; for and To the left of the point, the same first-order Taylor expansion approximation method is used at the point. and Use its linear lower bound function instead: ； ； For non-convex terms Application processing Auxiliary variables introduced at the time Transform it into: ； For the angle CRB and distance CRB constraints of the perceived target, the Schur complementary relaxation method is used to... and Reconstruct it into a convex form that is easy to handle. and Then reconstruct it into the formula Japanese style : ； ； At this point, all non-convex terms have been transformed into convex terms, and the entire optimization problem is now transformed into... : ； in, The problem The result obtained after performing a non-convex term transformation is: , and These are the equations obtained by transformation. , and , It is a set of convex terms obtained by processing non-convex terms; 2) Problem The global optimal solution is found using the interior-point method. This optimization problem is solved using the CVX toolbox, and the problem is solved in each iteration of the SCA algorithm. Assuming the number of iterations is represented by m, then the value from the previous round is used. , , , , , , , Solve the problem to obtain the optimal value for this round. , , , , , , , Then use the optimal value obtained in this round to update , , , , , , , Finally, after m iterations, the maximum energy efficiency is output. ; For the initial beamformer It is achieved by finding a constraint that satisfies the constraint. , and Reconstruction and This yields a feasible solution, but to ensure fairness among communication users... Initialize the public flow rate of communication users by using a uniform distribution method. get, , , , , , Calculate using the corresponding formulas respectively. It is the covariance matrix of the signal, through get, The objective function is a function that increases rather than decreases with the number of iterations, and it always satisfies... This ensures convergence.

Images