CN-122028050-A - RIS-assisted honeycomb-free large-scale MIMO network security energy efficiency optimization method

CN122028050ACN 122028050 ACN122028050 ACN 122028050ACN-122028050-A

Abstract

The invention provides a RIS-assisted honeycomb-free large-scale MIMO network security energy efficiency optimization method which comprises the steps of establishing a RIS-assisted honeycomb-free large-scale MIMO network model under a rice fading channel, acquiring estimated channel state information of an aggregate channel between a base station and user equipment and between the base station and an eavesdropper by adopting a linear minimum mean square error estimation method, carrying out downlink channel data transmission based on the estimated channel state information, deducing a user reachable rate lower bound and an eavesdropper leakage rate upper bound according to a downlink channel data transmission result, deducing a security reachable rate expression and a security energy efficiency expression according to the user reachable rate lower bound and the eavesdropper leakage rate upper bound, establishing a non-convex optimization problem containing RIS phase shift and power distribution based on the security energy efficiency expression, solving the non-convex optimization problem by adopting an iterative optimization algorithm, and decoupling the original problem into a RIS phase shift optimization sub-problem and a power distribution optimization sub-problem by adopting the iterative optimization algorithm until the original problem is converged to a local optimal solution meeting all constraints.

Inventors

Jin Sinian
ZHAO SHIHANG
Ju Moran
CHEN YILING
YUE DIANWU

Assignees

大连海事大学

Dates

Publication Date: 20260512
Application Date: 20260324

Claims (7)

1. The RIS-assisted cellular-free large-scale MIMO network security energy efficiency optimization method is characterized by comprising the following steps of: s1, establishing an RIS-assisted honeycomb-free large-scale MIMO network model under a rice fading channel; s2, carrying out uplink channel estimation based on the constructed honeycomb-free large-scale MIMO network model, and acquiring estimated channel state information of an aggregate channel between a base station and single-antenna legal user equipment and between the base station and a single-antenna eavesdropper by adopting a linear minimum mean square error estimation method; S3, carrying out downlink channel data transmission based on the estimated channel state information, transmitting signals to the single-antenna legal user equipment by the base station through conjugate beam forming, and injecting artificial noise, wherein the artificial noise is located in a null space of the channel state of the single-antenna legal user equipment; S4, deducing a lower limit of the user reachable rate and an upper limit of the eavesdropper leakage rate according to the downlink channel data transmission result; S5, according to the lower limit of the user reachable rate and the upper limit of the eavesdropper leakage rate, a safe reachable and rate expression and a safe energy efficiency expression are deduced; S6, based on the safe energy efficiency expression, establishing a non-convex optimization problem containing RIS phase shift and power distribution, solving by adopting an iterative optimization algorithm, and decoupling the original problem into a RIS phase shift optimization sub-problem and a power distribution optimization sub-problem by adopting the iterative optimization algorithm, and solving alternately until the solution is converged to a local optimal solution meeting all constraints.
2. The RIS-assisted cellular-free massive MIMO network security energy efficient optimization method according to claim 1, wherein step S1 comprises: S11, is equipped with Base station BS of uniform planar array equipped with An intelligent reflection surface RIS of a uniform plane array, Single antenna legal user equipment UE A single antenna eavesdropper Eve, wherein each base station BS is equipped with Root antenna, each intelligent reflection surface RIS is composed of A plurality of reflection units; S12, connecting all base stations BS to a CPU (central processing unit) through a high-speed backhaul link, and communicating an intelligent reflection surface RIS with the CPU or an adjacent base station through a wired or wireless link; S13, in the honeycomb-free massive MIMO network, all base stations BS and the intelligent reflection surface RIS cooperate to transmit secret data to a single-antenna legal user, so an eavesdropper tries to eavesdrop all Assuming that all eavesdroppers are wirelessly connected to one eavesdropping convergence center, the center is responsible for collecting and aggregating all eavesdropping signals; S14, defining an index set as follows: a set of base stations BS is represented, Representing a set of intelligent reflective surfaces RIS, Representing a set of reflective elements, Representing a set of legitimate users of a single antenna, Representing a set of eavesdroppers; S15, in order to accurately describe the channel propagation characteristics between transceivers, consider the direct link BS-UE/Eve and the reflective link BS-RIS-UE/Eve, and assume that each link corresponding to the transmission path of the direct link BS-UE/Eve and the reflective link BS-RIS-UE/Eve follows the Laes fading model, thereby, from the first step Personal base station To the first Individual user From the first Personal base station To the first Eavesdroppers and eavesdroppers The channels of (a) are respectively expressed as: Wherein, the And Representing large scale path fading factors related to distance, which remain constant over a plurality of coherence times; And Is a rice factor representing the intensity ratio of the line-of-sight component to the non-line-of-sight component; in addition, in the case of the optical fiber, Representation of Is used for the distance-of-vision component of (a), Representation of Is a line of sight component of (a); Representation of Is used for the non-line-of-sight component of (c), Representation of The line-of-sight components are assumed to be deterministic and known a priori, whereas the non-line-of-sight components are modeled as rayleigh fading, i.e And (3) with By definition And (3) with For effective channel fading coefficients and define And (3) with Is the mean value vector to obtain And (3) with 。 Will be the first Personal base station And the first Intelligent reflecting surface Channels between (a) Intelligent reflecting surface And the first Individual user First, the Intelligent reflecting surface And the first Eavesdroppers and eavesdroppers The channels between are denoted as: Wherein, the 、 And Representing the large-scale fading coefficient, 、 And Representing the corresponding Lees factor, non-line-of-sight components 、 And Respectively correspond to 、 And Each element of the random Rayleigh distribution part is subjected to complex Gaussian random distribution with independent same distribution, namely Line of sight component 、 And Respectively correspond to 、 And Definition of certainty part 、 And To respectively correspond to 、 And Simultaneously defining the effective channel coefficients of (a) 、 And Channel vectors are the mean thereof; s16, for representing the line-of-sight component in the formula of the step S15, constructing a corresponding channel by adopting a uniform rectangular plane array model, wherein the size is as follows The array response vector of (a) is expressed as: Wherein, the Representing the Kronecker product of the matrix, The azimuth angle representing the departure angle or the corresponding arrival angle, A pitch angle representing a departure angle or a corresponding arrival angle; S17, assume that the total number of antennas is The array response vector along each coordinate axis is expressed as: Wherein, the Representing edges Shafts or The number of antenna elements in the axial direction, As the carrier wavelength is used, Is the cell pitch; S18, based on the array response vector defined in the step S16, the line-of-sight channel along the reflection link BS-RIS-UE/Eve is expressed as: Wherein, the Representing slave To the point of Is a separation angle from the azimuth angle of (c), Representing slave To the point of Is a separation angle from the azimuth angle of (c), Representing slave To the point of Is separated from the pitch angle of the roll, Representing slave To the point of Is separated from the pitch angle of the frame; Representing slave To the point of Is set to be a range of the azimuth departure angle, Representing slave To the point of Is a pitch angle departure angle; Representation of From the slave The azimuth angle of arrival of the received signal, Representation of From the slave Receiving a pitch angle arrival angle of a signal; Representing slave To the point of Is set to be a range of the azimuth departure angle, Representing slave To the point of Is set to be a range of the azimuth departure angle, Representing slave To the point of Is used for controlling the pitch angle and the departure angle of the vehicle, From the slave To the point of Is a pitch angle departure angle; S19, will And (3) with Aggregate channel sum between And (3) with The aggregate channel between is expressed as: Wherein, the Representation of Is used to determine the phase shift vector of (c), Representation of Is the first of (2) The reflection coefficient of each reflection unit, wherein Define the phase shift set of all RIS as 。
3. The RIS-assisted cellular-free massive MIMO network security energy efficient optimization method according to claim 1, wherein step S2 comprises: S21, in a RIS auxiliary non-cellular large-scale MIMO network downlink adopting a time division duplex protocol, the implementation of downlink precoding depends on the acquisition of Channel State Information (CSI), and the quasi-Channel State Information (CSI) is efficiently acquired through uplink training by utilizing reciprocity between uplink and downlink channels; s22, in the pilot-assisted channel estimation process, all user equipments will simultaneously transmit respective allocated pilot sequences to the base station, each sequence comprising Since the pilot signals used in standardized communication networks are generally public and known, it is assumed that an eavesdropper has full knowledge of the pilot sequence information, and the eavesdropper transmits the pilot sequence exactly the same as any legitimate user device, thereby initiating a pilot spoofing attack. S23, due to the satisfaction of So the common pilot pollution phenomenon in the dense user scene is built in the network model, and the method is that Representation allocation to user equipment And the orthogonality characteristic of the pilot sequence is expressed as: And Wherein Representative and user equipment A set of user equipments sharing the same pilot; S24, considering that the user equipment and the eavesdropper can adopt different pilot frequency transmitting power, the first step Personal base station The pilot signal matrix received at that location is expressed as: Wherein, the Representing the transmit power used by the user equipment during the training phase, Representing the transmit power used by an eavesdropper during the training phase, Item representation The element of the additive Gaussian white noise matrix is modeled as independent random variables distributed in the same way, and the element is compliant with the distribution In the pilot signal matrix formula, each eavesdropper deliberately transmits a random pilot sequence to enhance the information leakage effect, and receives the pilot signal matrix After that, the processing unit is configured to, Despreading operations to estimate aggregate channels The process is expressed as: Wherein, the , ; S25, estimating an aggregate channel by adopting a linear minimum mean square error estimation method Channel state information CSI: S26, for deriving Calculating the statistical expectation and covariance matrix of the correlation, i.e. the closed LMMSE estimate of (2) 、、 And By utilizing the first-order and second-order statistical properties of the complex Gaussian random variables and performing proper matrix operation, the analytical expression is deduced as follows: Wherein, the , , , , , , , , , , , , , , , , , , , , ; S27, substituting the two analysis expressions in the step S26 into the formula in the step S25 to obtain the first analysis expression Personal base station And the first Individual user The closed LMMSE estimate for the inter-aggregate channel is: Wherein, the ; S28, defining the estimation error as Estimating a channel And estimation error Is two mutually uncorrelated random vectors and satisfies: Wherein, the , The expected value of the channel power is Wherein 。
4. The RIS-assisted cellular-free massive MIMO network security energy efficient optimization method according to claim 1, wherein step S3 comprises: S31, in the downlink transmission stage, all base stations simultaneously transmit data symbols to user equipment, and artificial noise is embedded and overlapped on a downlink secret signal, so that eavesdropping of an eavesdropper is prevented on the premise of not affecting normal communication of the user equipment; S32, each base station adopts conjugate beam forming technology and controls the coefficient according to long-term power And Allocating transmit power of each base station to data symbols Artificial noise symbol In which And (3) with All are normalized to meet Correspondingly, the first Personal base station Is expressed as: Wherein, the In order to allocate transmit power to the useful data, And The transmitting power and the beam forming vector of the artificial noise are respectively corresponding; S33, according to the first Personal base station Is the transmission signal of (1) Personal base station Is regulated by a power control factor to ensure that it does not exceed the maximum allowed transmit power Namely, the following conditions are satisfied: Wherein, the Is the matrix Hadamard product and, Is a matrix Is the first of (2) Row, and satisfies: S34 based on the first Personal base station Is used for transmitting signals to user equipment And eavesdroppers The signals received at the receiver are respectively expressed as: Wherein, the And Respectively represent user equipments Eavesdroppers Additive white gaussian noise at the point, the remaining symbols are defined as follows: 。
5. The method for optimizing security energy efficiency of an RIS-assisted non-cellular massive MIMO network according to claim 1, wherein in step S4, for an RIS-assisted non-cellular massive MIMO network under a rice fading channel, and under a condition that there are a plurality of colluded eavesdroppers and channel state information are imperfect, analyzing total reachable security rate and security energy efficiency includes: S41, establishing a lower bound of user accessibility and speed, which specifically comprises the following steps: In RIS-assisted non-cellular massive MIMO networks, uatF definition techniques are employed to obtain a closed lower bound expression of user equipment and rate, which is particularly effective when assuming that the base station only grasps massive channel statistics rather than instantaneous channel realizations, under which framework the deterministic equivalent terms Effective channel gain, which is considered to be completely known to the receiver, and therefore the user equipment A received signal The equivalent expression is as follows: Wherein, the Representing the desired signal component(s), In response to the beam forming uncertainty, Representing inter-user interference from other user devices, Referring to the interference introduced by the artificial noise, Is noise interference; the reachability and rate of the user equipment are expressed as: Wherein, the Is a preprocessing factor used for representing the distribution After pilot training of each symbol, coherence interval Ratio of the user equipment to the data transmission The corresponding effective signal-to-interference-and-noise ratio is expressed as: Through user equipment Deducing all expected values in the corresponding effective signal-to-interference-and-noise ratio formula, and when all legal user equipment adopts a conjugate beam forming scheme, the closed expression of the downlink reachable sum rate is as follows: Wherein, the Representing defined matrices Is the first of (2) Column, to simplify the expression of symbols in the subsequent derivation, an auxiliary function is defined as The remaining parameters are defined as follows: Wherein, the The function is defined as if Then Otherwise, it is Its complement function is ; Based on the closed expression of the downlink reachability and the rate, the closed expression of the downlink reachability and the rate when all legal user equipment adopts the conjugate beam forming scheme is equivalently expressed as: Wherein, the , , , , , , , , , , , ; S42, establishing an eavesdropper leakage rate upper bound, which specifically comprises the following steps: Under the assumption that all eavesdroppers completely know the instantaneous channel information and eavesdrop all legal user equipment simultaneously in a coexistence mode, in the scene, all eavesdroppers are connected to a centralized eavesdrop fusion center, the center performs coherent combination on received signals by adopting a maximum ratio combining technology, and the UatF method is applied to the eavesdroppers In the received signal, the concurrent leakage rate and SINR of all eavesdroppers are expressed as: by calculating all expected values in the above equation, a leak rate expression for multiple eavesdroppers using the maximum ratio combining technique is obtained: Wherein: based on the leak rate expression, the closed-form expression of all eavesdropper leak rates when the maximum ratio combining technique is adopted is equivalently written as: Wherein, the , , , , , , , , , , , , , , , 。
6. The RIS-assisted cellular-free massive MIMO network security energy efficient optimization method according to claim 1, wherein step S5 comprises: s51, expressing the lower bound of SASR as a subtraction form: Wherein the ramp function is defined as ; S52, modeling the total power consumption of the RIS-assisted non-cellular massive MIMO network as: Wherein, the , , The first term in (2) represents the total transmit power consumption of all base stations, where Representation of Power amplifier efficiency of (2) Corresponding to The power consumption of the back-haul link required to transfer data to the central processor, wherein Is the first The fixed power consumption of the postamble link, Is the bandwidth of the system and, Representing dynamic power consumption related to traffic, third item Representing static circuit power consumption, including Is the first of (2) Power consumption of individual antennas 、 Is the first of (2) Power consumption of each reflection unit And Power consumption of (2) ; S53, equivalently rewriting the total power consumption of the RIS-assisted non-cellular massive MIMO network into the following steps of simplifying the expression in the following deduction: Wherein, the , , ; S54, defining the spectrum energy efficiency as follows according to the formula in the step S51 and the rewritten total power consumption formula: S55, substituting the reachable rate of the legal user, the leakage rate of the eavesdropping user and the total rewritten power consumption into the formula and the frequency spectrum energy efficiency formula in the step S51 to obtain a closed expression of SASR and the SEE lower bound.
7. The RIS-assisted cellular-free massive MIMO network security energy efficient optimization method according to claim 1, wherein step S6 comprises: S61, developing a high-efficiency resource allocation framework aiming at a safety energy-efficiency optimization method, aiming at maximizing SEE and optimizing and jointly adjusting RIS phase shift And power control coefficients for the data signal and the artificial noise signal The overall optimization problem is expressed as follows: Wherein the formula is Unit mode constraint is carried out on all RIS, and a formula is adopted Ensuring that the transmit power of each base station does not exceed a maximum limit Formula (VI) And Requiring the achievable rate of each legitimate user to be no less than a minimum threshold And the leak rate of each eavesdropper does not exceed an upper limit ; S62, establishing a non-convex optimization problem of RIS phase shift, which specifically comprises the following steps: By adopting the technique of the extended Lagrangian multiplier method Conversion to unconstrained optimization problem on Riemann manifold, where inequality constraints are incorporated as penalty terms into the objective function, therefore Restated as follows: wherein penalty parameters Vector quantity Representing the lagrangian multiplier vector, the remaining parameters are defined as follows: Problem(s) Is an unconstrained optimization problem defined on a manifold, efficiently solves using the steepest descent method in Manopt toolbox, calculates the Riemann gradient of the objective function as the corresponding concept of Euclidean gradient on the manifold, and derives the Riemann gradient by projecting the Euclidean gradient onto the tangent space of the manifold The euclidean gradient of (c): Wherein: wherein the indication function At the position of Time equal to Otherwise, it is ; Mapping the derived Euclidean gradient onto the tangent space of the manifold by utilizing projection operation to obtain the Riemann gradient, wherein the expression is as follows: After obtaining the Riemann gradient, using the Riemann gradient as a descent direction to iteratively optimize the control variable, the updating is achieved by a back-off operation that maps the candidate points from the cut space back onto the manifold, thereby maintaining geometric feasibility throughout the optimization process, namely: Wherein, the Representing a step size; S63, establishing a non-convex optimization problem of power distribution, which specifically comprises the following steps: For a given set By introducing auxiliary variables 、 And Wherein Will optimize the problem Equivalently reconstruct as: Wherein, the , , , , , ; At the time of inspection When a path tracking (PF) algorithm is introduced, an operator is added Relaxed into And applying a multidimensional quadratic transformation to obtain an objective function in the form of a convex approximation: Wherein, the ; Using the convex inequality, the formula can be formulated Is approximated as its convex counterpart: Wherein, the ; Constraint formula is given by means of convex inequality The approximation is expressed as: Wherein, the ; The formula is developed by combining the chain law with a first-order taylor expansion The non-convex constraint approximation of (a) translates into the convex form: Wherein, the ; Further under PF frame Is the first of (2) The number of iterations is expressed as: finally, adopting PF optimization algorithm to solve the problems Solving; S64, obtaining local optimal phase shift based on IALMO algorithm Local optimal power control coefficient calculated by PF algorithm An iterative optimization framework is constructed to solve the problem Iterative updating by alternately solving RIS phase shift optimization sub-problems and power allocation optimization sub-problems And Until the network energy efficiency converges to a local maximum.

Description

RIS-assisted honeycomb-free large-scale MIMO network security energy efficiency optimization method Technical Field The invention relates to the technical field of wireless communication, in particular to a RIS-assisted honeycomb-free large-scale MIMO network security energy efficiency optimization method. Background As wireless networks evolve towards 6G, data-intensive traffic proliferates pushing the urgent need for high-speed traffic. The honeycomb-free large-scale MIMO network is used as an innovative architecture, and the distributed access points are used for cooperatively serving the users, so that the inter-cell interference is obviously reduced, and the link robustness is improved. However, densely deployed active access points also present high power consumption, which poses serious challenges for long-term sustainable operation of the network. RIS is an emerging green communication technology that can enhance signal coverage with very low energy overhead by passively regulating the electromagnetic wave propagation environment. RIS is introduced into a honeycomb-free large-scale MIMO network, so that the defect of high energy consumption caused by dense deployment can be effectively overcome, dual promotion of network spectrum efficiency and energy efficiency is realized, and a key path is provided for efficient and energy-saving operation of a future network. The combination of RIS and non-cellular massive MIMO not only can further enhance the received signals of legal users through intelligent reflection, but also can effectively inhibit the signal receiving capability of eavesdroppers, thereby obviously enhancing the physical layer security of the network. This synergistic architecture provides a very potential solution to address the dual security and energy efficiency challenges in future wireless networks. Although research has explored physical layer security in RIS-assisted non-cellular MIMO networks, systematic analysis and optimization research on SEE is still very limited for imperfect channel state information scenarios with multiple eavesdroppers in mind. An accurate analysis framework and efficient joint resource allocation scheme is lacking to quantify and maximize the safe energy efficiency of the architecture. Therefore, in the RIS-assisted non-cellular MIMO network under the Rice fading channel, the problems of multi-eavesdropper cooperation and imperfect CSI confidentiality and energy efficiency maximization are mainly studied, a set of complete performance analysis, theoretical derivation and optimization algorithm is provided, and the method has important theoretical value and practical application significance. Disclosure of Invention According to the problem of SEE maximization in the downlink RIS-assisted non-cellular massive MIMO network with coexistence of multiple eavesdroppers, the RIS-assisted non-cellular massive MIMO network security energy efficiency optimization method is provided. The invention firstly carries out uplink channel estimation by an LMMSE method, and all user equipment and eavesdroppers in the scene share the same pilot frequency set. Based on imperfect channel state information, on the basis of introducing artificial noise and coordinated beamforming precoding, a closed lower bound of SASR and SEE is deduced, and an analysis performance reference is established. In order to solve the non-convex SEE maximization problem, a novel alternative optimization framework integrating Riemann manifold optimization and path tracking algorithm is provided. The method successfully converts the original coupling problem into a series of processable sub-problem sequences. The simulation result verifies the superiority of the submitted replacement optimization algorithm in terms of performance and convergence robustness, and compared with a reference scheme, the simulation result realizes remarkable SEE gain. The invention adopts the following technical means: A RIS-assisted cellular-free large-scale MIMO network security energy efficiency optimization method comprises the following steps: s1, establishing an RIS-assisted honeycomb-free large-scale MIMO network model under a rice fading channel; s2, carrying out uplink channel estimation based on the constructed honeycomb-free large-scale MIMO network model, and acquiring estimated channel state information of an aggregate channel between a base station and single-antenna legal user equipment and between the base station and a single-antenna eavesdropper by adopting a linear minimum mean square error estimation method; S3, carrying out downlink channel data transmission based on the estimated channel state information, transmitting signals to the single-antenna legal user equipment by the base station through conjugate beam forming, and injecting artificial noise, wherein the artificial noise is located in a null space of the channel state of the single-antenna legal user equipment; S4, deducing a lower limit of the