Precoding method and system based on CI in ISAC system

By constructing a non-convex optimization problem based on CI in the MIMO-ISAC system, and optimizing radar lighting power using the SCA method, the problem of high CI-SLP calculation complexity is solved, and more efficient precoder design and system performance improvement is achieved.

CN119051699APending Publication Date: 2024-11-29XI AN JIAOTONG UNIV
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411093430.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-09
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

The calculation complexity of CI-SLP in the existing MIMO-ISAC system is too high and difficult to be accepted in the actual system, resulting in a long optimization time.

Method used

The SCA method is used to construct non-convex optimization problems under communication CI constraints and daily line power constraints. The objective function is the maximum total lighting power of the radar for the target of interest. By simplifying the optimization of the objective function and using the continuous convex approximation method, CI-SLP is reduced. The computational complexity of the , and the precoding scheme is extended from the symbol level to the block level, providing a unified optimization form.

Benefits of technology

It significantly shortens the execution time of CI-SLP, reduces computational complexity, and improves the performance of ISAC systems, especially in terms of bit error rate and beam pattern optimization, providing a more efficient precoder update frequency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119051699A_ABST
    Figure CN119051699A_ABST
Patent Text Reader

Abstract

The invention provides a CI-based precoding method and system in an MIMO-ISAC system, and the method comprises the following steps: constructing a CI-based non-convex optimization problem under the communication CI constraint and daily linear power constraint according to a communication model and a radar model, and enabling the target function of the non-convex optimization problem to be the maximum total illumination power of a radar for an interested target; and solving the non-convex optimization problem by adopting an SCA method to obtain an optimal emission signal matrix. In the CI-SLP ISAC system, the illumination power of the target is used as the target function, the illumination power of the target is maximized, the detection probability is as high as possible, the complexity of the constructed non-convex optimization problem is greatly reduced, and after the target function is simply processed, the optimal solution of the optimization problem can be obtained only through dozens of iterations. And the calculation complexity of the CI-SLP scheme is greatly reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of communication perception integration, and in particular to a CI-based precoding method and system in a MIMO-ISAC system. Background Art

[0002] From vehicles to drones, from monitoring facilities in cities to farm tools in rural areas, in the 6G era, promoting the gradual integration of the physical world and the digital world is considered a reality. To this end, the wireless industry suggests that future networks should be able to "see" the physical world, which is considered one of the unique features of the next generation of networks, rather than the traditional communication-only function. This idea has been prompting researchers to rethink the design of current communication infrastructure and terminals. In addition, spectrum resources are becoming increasingly scarce, and spectrum sharing between radar and communication systems is consistent with the idea of ​​building a gradual integration of the physical world and the digital world, which has triggered recent research in the field of "ISAC". In this model, through the same hardware platform, common spectrum, joint signal processing strategies, and even a unified control framework, sensing and communication functions may be jointly designed, optimized, and scheduled to share resources or assist each other.

[0003] There are many radar metrics and communication metrics in Integrated Sensing and Communication (ISAC) systems. Many researchers focus on the transmit precoding design of Multiple-Input Multiple-Output (MIMO) ISAC systems, optimizing the transmit precoding matrix through different radar and communication metrics. Typical radar metrics include the radar receiver's signal-to-noise ratio (SNR), beam direction mean square error (MSE), Cramer-Rao bound, and the similarity between the designed beamformer and the beamformer of the reference radar system only. Meanwhile, widely used communication metrics include achievable rate, signal-to-interference-to-noise ratio (SINR) of the communication users, and multi-user interference (MUI). Under the system resource conditions, the waveform design of the ISAC system needs to take into account both radar performance and communication performance. Therefore, the transmit waveform should be carefully designed to balance the requirements of these two functions to obtain better system performance.

[0004] The above transmit precoding designs are all traditional block-level precoding (BLP), which aims to eliminate interference between users. In fact, research shows that multi-user interference does not need to be completely eliminated. Unlike traditional BLP, symbol-level precoding (SLP) is a nonlinear and symbol-dependent method that optimizes each instantaneous transmission vector based on the specific symbol to be transmitted, rather than simply eliminating MUI at the symbol level. From a communication perspective, SLP can use symbol information to convert harmful MUI into constructive interference (CI) to reduce the symbol error rate (SER) and achieve more reliable multi-user communication. From a radar perspective, the instantaneous transmit beam pattern of each time slot can be carefully designed to ensure a well-formed beam pattern with a limited number of waveform samples.

[0005] In previous work on ISAC systems, the transmit vector of CI-SLP is optimized to minimize the square error between the obtained beam pattern and the desired beam pattern, subject to the communication CI constraint and power constraint. This scheme uses three methods: Penalty Dual Decomposition (PDD), Majorization-Minimization (MM), and Block Coordinate Descent (BCD) to solve the optimization problem, called PDD-MM-BCD. Simulation results show that the CI-SLP scheme for ISAC systems can achieve better instantaneous beam patterns and better SER performance compared to the traditional BLP scheme. However, due to the complex processing of the quartic function in the optimization objective and the need for thousands of iterations, the computational complexity of PDD-MM-BCD is too high to be accepted in practical systems. The high computational complexity of the CI-SLP format compared to the traditional BLP format is also a long-standing problem. Summary of the invention

[0006] In order to solve the above problems of the prior art, the present invention provides a CI-based precoding method and system in a MIMO-ISAC system, which greatly reduces the complexity of CI-SLP and significantly shortens the execution time required for CI-SLP.

[0007] The present invention is achieved through the following technical solutions:

[0008] A CI-based precoding method in an ISAC system comprises the following steps:

[0009] According to the communication model and the radar model, a non-convex optimization problem based on CI is constructed under the constraints of communication CI and power per antenna, wherein the objective function of the non-convex optimization problem is to maximize the total illumination power of the radar on the target of interest;

[0010] The SCA method is used to solve the non-convex optimization problem to obtain the optimal transmission signal matrix.

[0011] Preferably, the communication model is:

[0012] The signal received by the kth user in the lth time slot is:

[0013]

[0014] in, is the channel between ISAC-BS and user k, is the additive noise of user k in the lth time slot, is the noise power; x[l]=[x1[l],x2[l],…,x N [l]] T is the ISAC waveform corresponding to the lth time slot;

[0015] Based on the symbol scaling metric, the data symbol vector s of user k in the lth time slot is calculated along the two decision boundaries of QPSK modulation. k [l] is decomposed to obtain:

[0016]

[0017] Received noise-free signal The two decision boundaries along QPSK modulation decompose into:

[0018]

[0019] in, and is a non-negative scaling factor;

[0020] Define the set vector α of all user scaling factors E , and x E :

[0021]

[0022] The set vector α of all user scaling factors E By x E It is expressed as:

[0023] α E =Mx E

[0024] M is a coefficient matrix as follows:

[0025] M=[p1,p2,…p Ku ,q1,q2,q Ku ] T

[0026]

[0027] or The value of represents the impact of inter-user interference.

[0028] Furthermore, the communication CI constraint formula is:

[0029] Mx E ≥t

[0030] Where t = σ max Γ,σ max for The maximum value of , Γ is the SINR threshold.

[0031] Furthermore, the radar model is:

[0032] The event signal at the target is expressed as:

[0033]

[0034] in, ISAC-BS towards the Kth t The target guidance vector, is the corresponding azimuth, is the effective channel propagation coefficient; x[l]=[x1[l],x2[l],…,x N [l]] T is the ISAC waveform corresponding to the lth time slot;

[0035] For k t The illumination power of a target is expressed as:

[0036]

[0037] in, and x E is defined as:

[0038]

[0039] Furthermore, when the precoding is symbol-level precoding, the non-convex optimization problem is:

[0040]

[0041] Where D is defined as:

[0042]

[0043] Where p0 represents the transmit signal power budget for each time slot, x n Represents x E The nth element in .

[0044] Furthermore, when the precoding is block-level precoding, the non-convex optimization problem is expressed as:

[0045]

[0046] in, and Defined as:

[0047]

[0048] D is defined as:

[0049]

[0050] χ is a convex set defined as:

[0051]

[0052] In the formula, x E,l and The element M in l represent the x of the lth time slot respectively E and M,x n,l Represents x E,l The nth element in ; represents the Kronecker product.

[0053] Furthermore, the SCA method is used to solve the non-convex optimization problem P2:

[0054] The objective function is expressed as:

[0055]

[0056] Define ξ as the feasible region of P2, by The first-order Taylor expansion near the target function approximates the target function, and we get:

[0057]

[0058] in, represents the gradient of f(·) and:

[0059]

[0060] The following optimization problem is solved at the mth iteration of the SCA solver:

[0061]

[0062] In the formula, in is the optimal solution after (m-1) algorithm iterations;

[0063] By solving problem P3, we get the optimal solution, which is expressed as The term generates the descent direction of each iteration, by moving the variable in the descent direction with a step size of λ,

[0064]

[0065] Among them, the step size λ is obtained by Armijo search or exact line search, The SCA solver iterates until convergence, according to Get the transmitted signal matrix X.

[0066] Furthermore, a parallel algorithm is used to solve problem P3:

[0067] Defined by problem P3:

[0068]

[0069] Set the initial feasible point and For any vector in the set χ, set the initialization vector Set loop iteration index i = 1. When i ≤ maximum number of iterations, perform the following steps:

[0070] renew Where α>0 is a constant parameter;

[0071] renew

[0072] Update average value

[0073] i=i+1;

[0074] if Then terminate the loop;

[0075] Finally get

[0076] Further updates Adopt the ADMM framework to:

[0077] Set the initialization feasible point z to be any vector in the set χ, and set the initialization vector y 1= 0, set loop iteration index loop = 1, when loop ≤ maximum number of iterations, perform the following steps:

[0078] renew

[0079] renew

[0080] renew

[0081] loop=loop+1;

[0082] if Then terminate the loop;

[0083] Finally get

[0084] Among them, ρ>0 is the penalty parameter,

[0085] The present invention also provides a precoding system based on CI in an ISAC system, comprising:

[0086] A problem construction module is used to construct a CI-based non-convex optimization problem according to a communication model and a radar model under a communication CI constraint and a per-antenna power constraint, wherein the objective function of the non-convex optimization problem is to maximize the total illumination power of the radar on the target of interest;

[0087] The solution module is used to solve the non-convex optimization problem by using the SCA method to obtain the optimal transmission signal matrix.

[0088] Compared with the prior art, the present invention has the following beneficial effects:

[0089] In the CI-SLP ISAC system, the present invention uses the illumination power of the target as the objective function to maximize the illumination power of the target and make the detection probability as high as possible, so that the complexity of the constructed non-convex optimization problem is greatly reduced. After a simple processing of the objective function, only dozens of iterations are required to obtain the optimal solution of the optimization problem, which greatly reduces the computational complexity of the CI-SLP solution. Since the optimization objective function is not a convex function, the present invention uses the Succesive Convex Approximation (SCA) method to solve the proposed optimization problem, and the optimal transmission signal matrix can be obtained. Compared with the existing CI-SLP solution in the ISAC system, the execution time required for CI-SLP in the present invention is significantly shortened.

[0090] Furthermore, the present invention extends the proposed precoding scheme from symbol-level precoding CI-SLP to block-level precoding CI-BLP, and obtains a unified problem form of symbol-level precoding and block-level precoding in the MIMO-ISAC system. CI-BLP can provide better bit error rate performance; at the same time, compared with the CI-SLP method that optimizes the precoding matrix (or precoding signal) at the symbol level, the CI-BLP scheme uses a constant precoding matrix for an entire symbol block, and only one optimization is required for each block of symbols, which greatly reduces the update frequency of the precoder and leaves more computing resources for each optimization. Compared with the existing traditional BLP scheme and CI-SLP scheme, the ISAC performance of CI-BLP in the present invention is significantly improved.

[0091] Furthermore, the optimization problem of each iteration in the SCA method is a convex optimization problem. In order to further improve the solution efficiency, the present invention adopts a simple parallel algorithm and the Alternate Direction Multiplier Method (ADMM) to solve this sub-optimization problem. BRIEF DESCRIPTION OF THE DRAWINGS

[0092] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0093] Figure 1 Schematic diagram of the ISAC system scenario of the present invention, a multi-antenna ISAC BS serves multiple communication users and sends radar detection waveforms to the target of interest at the same time.

[0094] Figure 2 It is a communication model of the present invention, and a geometrical schematic diagram of a CI measurement method based on symbol scaling under QPSK modulation.

[0095] Figure 3 1 and 2. The communication SER performance (a) and radar beam pattern simulation result (b) of Example 1 of the present invention are shown.

[0096] Figure 4 1 and 1. The communication SER performance (a) and radar beam pattern simulation result (b) of Example 2 of the present invention.

[0097] Figure 5 This is a diagram of the simulation results of the execution time of different schemes of Example 2 of the present invention. DETAILED DESCRIPTION

[0098] The following describes the embodiments of the present invention through specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention.

[0099] It should be noted that the process equipment or devices not specifically specified in the following embodiments are all conventional equipment or devices in the art.

[0100] It should be noted that the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices. Moreover, unless otherwise specified, the numbering of each method step is only a convenient tool for identifying each method step, and is not intended to limit the order of arrangement of each method step or to define the scope of the present invention. Changes or adjustments in their relative relationships should also be regarded as the scope of the present invention without substantially changing the technical content.

[0101] The present invention considers a MIMO-ISAC system, in which a multi-antenna ISAC-BS sends communication symbols to multiple single-antenna users in the downlink, and at the same time sends a radar detection waveform to the target of interest. The ISAC-BS uses the same antenna array for transmission and reception in different time slots through a time division protocol. The present invention proposes a CI-based MIMO-ISAC system precoding scheme, and extends the proposed precoding scheme from the symbol level to the block level. The illumination power of the radar is maximized under the constraints of the communication CI and the power per antenna. Since the optimization objective function is not a convex function, the present invention uses the SCA method to solve the proposed optimization problem. The present invention uses a simple parallel algorithm and ADMM, which can significantly reduce the computational complexity to solve the sub-problems in the SCA method.

[0102] The CI-based precoding method in the ISAC system provided by the present invention comprises the following steps:

[0103] Step 1: Reference Figure 1 , a MIMO-ISAC system is constructed, where the multi-antenna ISAC-BS simultaneously sends radar detection waveforms to the target of interest and sends communication symbols to the downlink user. ISAC-BS uses a uniform linear array composed of N antennas, for K u Single antenna users provide services and detect K t The target of interest is usually K u ≤N,Kt ≤ N. Through the time division protocol, the same antenna array is used for transmission and reception in different time slots.

[0104] The transmit signal matrix X = [x[1], x[2], …, x[L]] is used as the ISAC waveform for radar and communication, where x[l] = [x1[l], x2[l], …, x N [l]] T is the ISAC waveform corresponding to the lth time slot. In addition, K u The data symbol vector of a user is expressed as The value is taken from the normalized QPSK constellation point. The data symbol vector of user k in the lth time slot is represented by s k [l].

[0105] Step 2: Based on the MIMO-ISAC system, obtain the channel state information from ISAC-BS to different users and the steering vector to the target. Use the acquired information to build the communication model and radar model.

[0106] The signal received by the kth user in the lth time slot is:

[0107]

[0108] in, is the channel between ISAC-BS and user k, is the additive noise of user k in the lth time slot, is the noise power.

[0109] Reference Figure 2 , the present invention describes a CI metric based on symbol scaling for the quarter case of a QPSK constellation. For simplicity, the present invention removes the time slot subscript [l] in the figure. Without loss of generality, the present invention assumes is the nominal constellation point of user k in the lth time slot:

[0110]

[0111] The noise-free received signal with interference can be expressed as:

[0112]

[0113] Can be considered as the sum of CI from other user flows.

[0114] Based on the symbol scaling metric, the two decision boundaries along the QPSK modulation Decompose it and get and

[0115]

[0116] in, and They represent the components of the nominal constellation on the left and right decision boundaries, that is, based on the symbol scaling metric, along the two decision boundaries of QPSK modulation s k [l] is decomposed to obtain:

[0117]

[0118] Following a similar process, the received noise-free signal It can also be decomposed along two decision boundaries into

[0119]

[0120] The received noise-free signal It can also be decomposed along two decision boundaries as:

[0121]

[0122] in, and is a non-negative scaling factor. The present invention can be observed that and The value of represents the influence of CI between users. or A larger value of means that the communication symbol can be pushed away from one of its decision boundaries.

[0123] Define the set vector α of all user scaling factors E , and x E :

[0124]

[0125] in, and Represents the real part and imaginary part of each element. The set vector α of all user scaling factors E Can be x E It is expressed as:

[0126] α E =Mx E ,

[0127] M is a coefficient matrix, calculated as follows:

[0128]

[0129] or The value of represents the impact of inter-user interference, or A larger value of means that the symbol is pushed away from one of its decision boundaries, which is beneficial for symbol detection and thus further improves the bit error rate performance of the MIMO communication system.

[0130] Based on the above description, the communication CI constraint formula is:

[0131] Mx E ≥t

[0132] Where t = σ max Γ. Among them, σ max Represents the maximum noise power value among all users, which is The maximum value of , Γ is the SINR threshold.

[0133] The event signal at the target is expressed as:

[0134]

[0135] in, ISAC-BS towards the Kth t The guidance vector of the target, j represents θ kt is the corresponding azimuth, is the effective channel propagation coefficient.

[0136] The illumination power of the radar on the target of interest is expected to be as high as possible to ensure the detection probability of the target. t The illumination power of a target is expressed as:

[0137]

[0138] in, and is defined as:

[0139]

[0140] Step 3: Construct a CI-based non-convex optimization problem in the ISAC system to maximize the total illumination power of the radar to the target of interest while satisfying the communication CI constraint and the power constraint of each antenna, thereby improving the radar detection probability. The CI-SLP optimization problem in the ISAC system can be expressed as:

[0141]

[0142] Where p0 represents the transmit signal power budget for each time slot, x n Represents x E The nth element in . The objective function can be written in standard quadratic form:

[0143]

[0144] Where D is defined as:

[0145]

[0146] Considering the entire symbol block, the CI-BLP optimization problem in the ISAC system under the constraint of each antenna power can be expressed as:

[0147]

[0148] In the formula, x E,l and M l represent the x of the lth time slot respectively E and M,x n,l Represents x E,l The nth element in .

[0149] Similarly, It can be expressed more concisely as:

[0150]

[0151] in, and Defined as:

[0152]

[0153] Here, I L represents an L×L identity matrix, represents the Kronecker product. χ is a convex set defined as:

[0154]

[0155] Obviously, when L=1, CI-SLP is a special case of CI-BLP. Therefore, the subsequent algorithm proposed for CI-BLP is also applicable to CI-SLP. Defined as P2, it is a unified optimization problem form of CI-SLP and CI-BLP.

[0156] Step 4: Solve the non-convex optimization problem P2 using the SCA method. In order to solve the problem P2 using the SCA solver, the objective function is expressed as

[0157]

[0158] Define ξ as the feasible domain of P2, which is convex. The present invention can be solved by the objective function at a given point The first-order Taylor expansion near the target function is used to approximate the target function, and we get

[0159]

[0160] in, represents the gradient of f(·) and

[0161]

[0162] Note that the constant term in the approximate objective function can be ignored. Therefore, the following optimization problem is solved at the mth iteration of the SCA solver:

[0163]

[0164] In the formula in The optimal solution for the algorithm iteration. P3 is a convex problem and can be solved directly using the CVX tool. By solving problem P3, the present invention obtains the optimal solution, which is expressed as here, The term generates the descending direction of each iteration. By moving the variable in the descending direction with a step size of λ, the present invention has

[0165]

[0166] The step size λ can be obtained by Armijo search or exact line search. It is worth noting that since ξ is a convex set, The SCA solver iterates until convergence, according to The construction process of can obtain the transmitted signal matrix X.

[0167] The CI-based precoding design algorithm proposed in the present invention comprises the following steps:

[0168] Set the initial feasible point Initialize the parameters as described above. and t, set loop iteration index m = 1, and repeat the following steps:

[0169] Computing Gradients

[0170] The optimal solution is obtained by P3

[0171] renew where λ can be obtained by Armijo search or exact line search;

[0172] m=m+1;

[0173] when Or when m>maximum number of iterations, the repetition is terminated;

[0174] From the obtained and The transmitted signal matrix X is calculated by constructing .

[0175] Step 5: For P3, a simple parallel algorithm can be used to solve it more effectively. The present invention can first be defined by optimizing problem P3

[0176]

[0177] This simple parallel algorithm consists of the following steps:

[0178] Set the initial feasible point and For any vector in the set χ, set the initialization vector Set loop iteration index i = 1. When i ≤ maximum number of iterations, perform the following steps:

[0179] renew Where α>0 is a constant parameter;

[0180] renew

[0181] Update average value

[0182] i=i+1;

[0183] if The loop is terminated.

[0184] Finally get

[0185] In order to improve the efficiency of the algorithm, In the update of , the present invention adopts the ADMM framework to solve the corresponding optimization problem. In order to apply the ADMM framework, the present invention introduces the variable z and the indicator function as follows:

[0186]

[0187] renew The objective function of the sub-problem is defined as:

[0188]

[0189] in, Update The sub-problem can be equivalently written in a consensus form:

[0190]

[0191] For the optimization problem P4, its augmented Lagrangian function is expressed as:

[0192]

[0193] Where y is the dual vector and ρ>0 is the penalty parameter. In the ADMM framework, The update of z and y can be written as

[0194]

[0195] exist In the update, The sub-problem can be written as:

[0196]

[0197] Let the objective function in the above formula be The derivative of is zero, and the optimal for:

[0198]

[0199] In the update of z, the subproblem of z can be written as:

[0200]

[0201] The optimal z is expressed as

[0202]

[0203] For all can be realized by the following closed form

[0204]

[0205] The above ADMM algorithm includes the following steps:

[0206] Set the initialization feasible point z to be any vector in the set χ, and set the initialization vector y 1 = 0, set loop iteration index loop = 1, when loop ≤ maximum number of iterations, perform the following steps:

[0207] renew

[0208] renew

[0209] renew

[0210] loop=loop+1;

[0211] if The loop is terminated.

[0212] Finally get

[0213] So far, both the technical solution and the simulation results can verify the effective effect of the method of the present invention in ISAC.

[0214] The present invention provides a precoding system for implementing the above method, comprising:

[0215] A problem construction module is used to construct a CI-based non-convex optimization problem according to a communication model and a radar model under a communication CI constraint and a per-antenna power constraint, wherein the objective function of the non-convex optimization problem is to maximize the total illumination power of the radar on the target of interest;

[0216] The solution module is used to solve the non-convex optimization problem by using the SCA method to obtain the optimal transmission signal matrix.

[0217] Example 1

[0218] The transmit power budget for each time slot is set to p0 = 30 dBm, QPSK modulation is used, the BS is equipped with N = 10 antennas, and the number of users is K u =3, number of targets K t =3, respectively at θ1 = -40°, θ2 = 0° and θ3 = 40°, β1 = β2 = β3 = 1. The block length considered in the CI-BLP scheme is L = 6. Γ = 4dB, σ max =20dBm. Radar false alarm probability P FA =10 -7 The expected beam direction main lobe width of the SDR scheme and the PDD-MM-BCD scheme is Δd = 10°. Figure 3 , which is a simulation result diagram of the communication SER performance and radar beam pattern of Example 1 of the present invention. The SER performance of the SDR scheme is significantly worse than that of other schemes. This is because the SDR scheme is a traditional BLP scheme, which reflects the communication performance advantage of the CI-based precoding scheme. The SER performance of the CI-SLP scheme and the CI-BLP scheme proposed in the present invention are basically consistent with the PDD-MM-BCD scheme, indicating that the optimization problems of the three schemes all meet the CI constraints. During multi-target detection, the transmission beam pattern of the CI-BLP scheme proposed in the present invention is the best, whether it is the main lobe or the side lobe.

[0219] Example 2

[0220] The transmit power budget for each time slot is set to p0 = 30 dBm, QPSK modulation is used, the BS is equipped with N = 10 antennas, and the number of users is K u =3, number of targets K t =1, θ = 0°, β = 1. The block length considered in the CI-BLP scheme is L = 6. Γ = 4dB, σmax =20dBm. Radar false alarm probability P FA =10 -7 The expected beam direction main lobe width of the SDR scheme and the PDD-MM-BCD scheme is Δd = 10°. Figure 4 , which is a simulation result diagram of the communication SER performance and radar beam pattern of Example 2 of the present invention. The present invention notes that the SER results of multi-target detection and single-target detection are the same, which reflects that in the ISAC optimization scheme constrained by the communication function, the parameter change of the radar sensing function does not affect the communication performance. When detecting a single target, the transmit beam pattern of the CI-BLP scheme proposed by the present invention is still the best, whether it is the main lobe or the side lobe. Figure 5 , which is a simulation result diagram of the execution time of different schemes of Example 2 of the present invention. It can be seen that the execution time of the CI-SLP scheme and the CI-BLP scheme proposed in the present invention is significantly shorter than that of the PDD-MM-BCD scheme, which is due to the fact that the objective function proposed in this paper is simpler and the algorithm is more efficient.

[0221] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0222] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A CI-based precoding method in an ISAC system, characterized in that: The following steps are involved: According to the communication model and the radar model, a non-convex optimization problem based on CI is constructed under the constraints of communication CI and power per antenna, wherein the objective function of the non-convex optimization problem is to maximize the total illumination power of the radar on the target of interest; The SCA method is used to solve the non-convex optimization problem to obtain the optimal transmission signal matrix.

2. The CI-based precoding method in the ISAC system according to claim 1, characterized in that: The communication model is: The signal received by the kth user in the lth time slot is: in, is the channel between ISAC-BS and user k, is the additive noise of user k in the lth time slot, is the noise power; x[l]=[x1[l],x2[l],…,x N [l]] T is the ISAC waveform corresponding to the lth time slot; Based on the symbol scaling metric, the data symbol vector s of user k in the lth time slot is calculated along the two decision boundaries of QPSK modulation. k [l] is decomposed to obtain: Received noise-free signal The two decision boundaries along QPSK modulation decompose into: in, and is a non-negative scaling factor; Define the set vector α of all user scaling factors E , and x E : The set vector α of all user scaling factors E By x E It is expressed as: a E =Mx E M is a coefficient matrix as follows: or The value of represents the impact of inter-user interference.

3. The CI-based precoding method in the ISAC system according to claim 2, characterized in that: The communication CI constraint formula is: Mx E ≥t Where t = σ max Γ,σ max for The maximum value of , Γ is the SINR threshold.

4. The CI-based precoding method in the ISAC system according to claim 3, characterized in that: The radar model is: The event signal at the target is expressed as: in, ISAC-BS towards the Kth t The target guidance vector, is the corresponding azimuth, is the effective channel propagation coefficient; x[l]=[x1[l],x2[l],…,x N [l]] T is the ISAC waveform corresponding to the lth time slot; For k t The illumination power of a target is expressed as: in, and x E is defined as:

5. The CI-based precoding method in the ISAC system according to claim 4, characterized in that: When the precoding is symbol-level precoding, the non-convex optimization problem is: s.t.Mx E ≥t, Where D is defined as: Where p0 represents the transmit signal power budget for each time slot, x n Represents x E The nth element in .

6. The CI-based precoding method in the ISAC system according to claim 4, characterized in that: When the precoding is block-level precoding, the non-convex optimization problem is expressed as: in, and Defined as: D is defined as: χ is a convex set defined as: In the formula, x E,l and The element M in l represent the x of the lth time slot respectively E and M,x n,l Represents x E,l The nth element in ; represents the Kronecker product.

7. The CI-based precoding method in the ISAC system according to claim 6, characterized in that: Use SCA method to solve non-convex optimization problem P2: The objective function is expressed as: Define ξ as the feasible region of P2, by The first-order Taylor expansion near the target function approximates the target function, and we get: in, represents the gradient of f(·) and: The following optimization problem is solved at the mth iteration of the SCA solver: In the formula, in is the optimal solution after (m-1) algorithm iterations; By solving problem P3, we get the optimal solution, which is expressed as The term generates the descent direction of each iteration, by moving the variable in the descent direction with a step size of λ, Among them, the step size λ is obtained by Armijo search or exact line search, The SCA solver iterates until convergence, according to Get the transmitted signal matrix X.

8. The CI-based precoding method in the ISAC system according to claim 7, characterized in that: Use parallel algorithm to solve problem P3: Defined by problem P3: Set the initial feasible point and For any vector in the set χ, set the initialization vector Set loop iteration index i = 1. When i ≤ maximum number of iterations, perform the following steps: renew Where α>0 is a constant parameter; renew Update average value i=i+1; if Then terminate the loop; Finally get 9. The CI-based precoding method in the ISAC system according to claim 8, characterized in that: renew Adopt the ADMM framework to: Set the initialization feasible point z to be any vector in the set χ, and set the initialization vector y 1 = 0, set loop iteration index loop = 1, when loop ≤ maximum number of iterations, perform the following steps: renew renew renew loop=loop+1; if Then terminate the loop; Finally get Among them, ρ>0 is the penalty parameter, 10. A CI-based precoding system in an ISAC system, characterized in that: include: A problem construction module is used to construct a CI-based non-convex optimization problem according to a communication model and a radar model under a communication CI constraint and a per-antenna power constraint, wherein the objective function of the non-convex optimization problem is to maximize the total illumination power of the radar on the target of interest; The solution module is used to solve the non-convex optimization problem by using the SCA method to obtain the optimal transmission signal matrix.

Citation Information

Cited By

  • Single-bit radar communication integrated transmitted waveform design method based on symbol-level precoding

    CN119814094A

  • Interference utilization and sensing integrated precoding method and system based on low-resolution DAC / PS

    CN120710546A