Dynamic omnidirectional metasurface-assisted indoor transmission optimization method

Abstract

The invention provides a dynamic omni-directional metasurface-assisted indoor transmission optimization method, which specifically comprises the following steps: firstly, constructing a dynamic IOS deployment scheme, deploying an IOS on a movable track embedded into a wall body, and enabling a base station to change the position of the IOS by controlling the track so as to provide a better reflection / transmission link for a user; then, a practical transmission protocol with both staged updating and instant updating is designed, the IOS position is updated once every N time slots, and the IOS coefficient and the base station beam forming are updated in each time slot. The green communication problem is solved through a two-stage optimization method: in the first stage, based on instantaneous channel state information, an alternating optimization and penalty dual decomposition technology is adopted to obtain a semi-closed solution of an IOS coefficient and base station beam forming; and in the second stage, on the basis of statistical channel state information, the IOS position is updated by utilizing a continuous convex approximation technology, so that the minimization of the total transmitting power of the base station is realized while the signal-to-noise ratio requirement of the user is met.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless communication, specifically to the technical field of intelligent surface assisted indoor communication, and particularly to the application of intelligent omni-surface (IOS) based on dynamic deployment in improving indoor network coverage and reducing wall penetration loss, which is suitable for indoor signal transmission optimization scenarios in high frequency bands in 5G and next generation wireless communication systems. BACKGROUND

[0002] With the development of 5G and next generation communication technology, high frequency bands such as millimeter waves are widely used due to abundant spectrum resources, but face the problem of large wall penetration loss. The high insulation materials of modern buildings further exacerbate signal attenuation, making it difficult to cover indoors.

[0003] To this end, the emerging field of intelligent building design includes wireless communication considerations in the early stages of building planning, suggesting the use of materials and structures that are more conducive to signal propagation, such as intelligent walls. The development of reconfigurable intelligent surfaces (RIS) provides a promising approach to realizing these intelligent walls. Current research in the field of wireless communication mainly focuses on reflective RIS, also known as intelligent reflective surfaces, which only allow signals to be reflected to users on the same side as the transmitter, limiting communication coverage.

[0004] To solve this problem, a new technology, intelligent omni-surface (IOS), has emerged. Unlike RIS, which only provides reflection, IOS can simultaneously reflect and transmit incident signals to users on both sides of the surface. By utilizing the ability to reflect and transmit jointly, IOS is able to achieve full-dimensional wireless communication regardless of the user's position relative to the surface. However, existing research assumes that IOS is fixedly embedded in the wall, lacking flexibility and failing to adapt to dynamically changing signal propagation environments, such as signal obstacles caused by user movement, renovations, and other activities. Therefore, in order to adapt to the dynamics of the propagation environment, more flexible and controllable intelligent wall solutions need to be explored. SUMMARY

[0005] The application discloses a dynamic omnidirectional metasurface assisted indoor transmission optimization method.

[0006] The application discloses a dynamic omnidirectional metasurface assisted indoor transmission optimization method. Figure 1 The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method.

[0007] The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method. Figure 2 The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method.

[0008] Specifically, the transmission protocol adopts a mechanism combining phased and instant updating: one transmission period contains N time slots, wherein the position of the IOS is updated every N time slots, and the IOS coefficient and the base station beam forming are instantaneously updated every time slot.

[0009] The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method. Figure 3 The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method. The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method.

[0010] The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method. The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method. The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method. The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method. The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method. The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method. The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method. The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method. The application discloses the dynamic omnidirectional metasurface assisted indoor transmission optimization method. Line of sight (LoS) component and non-line of sight (NLoS) component, respectively. is modeled as an element-wise independent and identically distributed (i.i.d.) zero-mean, unit-variance complex Gaussian random matrix, i.e. The computational formula of is

[0011] where, is the arrival elevation and azimuth angle from the base station to the IOS, and

[0012]

[0013] where, is the Kronecker product, denotes the transpose conjugate operation. is the departure angle from the IOS to the user k.

[0014] Similarly, the reflection / transmission channel gain from the IOS to the user k is denoted as

[0015] where, is the departure elevation and azimuth angle from the IOS to the user . In addition, the coefficient of the IOS is defined as . According to the law of conservation of energy, there is . In addition, the transmission and reflection phase shift of the actual IOS hardware needs to satisfy: . Therefore, the signal-to-noise ratio (SINR) of the user is

[0016] where, is the diagonalization operation, is the user set, is the beamforming vector of the base station to the user k, is the additive white Gaussian noise power of the user .

[0017] Step 220, based on the Nth time slot of the communication protocol, the IOS position is designed by successive convex approximation, relaxation variable introduction and first-order Taylor expansion optimization.

[0018] Specifically, based on the designed transmission protocol (containing a period of N time slots), the overall optimization problem P1 is constructed:

[0019]

[0020] where [t] denotes the value of a variable in time slot t, t = {1, 2, · · ·, t}. The constraints include user SINR requirements, IOS hardware constraints (energy conservation and phase coupling), and location space limitations. Since IOS location movement cannot be completed instantaneously (physical adjustment is required) and depends on statistical channel state information (CSI), and base station beamforming and IOS coefficients can be optimized based on instantaneous CSI, P1 is decomposed into two stages: the first N - 1 time slots are optimized and the Nth time slot optimizes IOS location Thus, the IOS location optimization subproblem for the Nth time slot is simplified to P2:

[0021]

[0022] Since the P2 objective function is independent of the optimization variable q (IOS location), it is essentially a feasibility check problem. Therefore, by variable normalization (letting where , be the power allocation for user k), it is converted to:

[0023]

[0024] However, since there is no explicit expression and and are coupled, it is still difficult to solve. To solve the above problem, the distribution characteristics of need to be analyzed first. We rewrite as:

[0025] where is the path loss exponent of the IOS-to-user link, is the distance from the IOS to user k.

[0026] ,

[0027] where is the Rice factor of the IOS-to-user link.

[0028] The approximation relationship holds because the product of the two non-line-of-sight components has a mean of zero, a variance of 1, and a very small value that can be ignored in expectation operations. Therefore, which can be approximated as:

[0029] To further simplify, we introduce the following proposition: Proposition 1: obeys a complex Gaussian distribution and is independent of , only related to . According to Proposition 1, does not affect the distribution of , but affects its value, and further affects . This coupling relationship makes constraints more difficult to handle. To solve the above difficulties, we try to give suboptimal and to decouple and in constraint . Note that since , and have complex relationships in , it is difficult to obtain the optimal adapted to any . In , we are more concerned about the impact of IOS position on , so based on angle information, we design the following low complexity suboptimal closed-form solution:

[0030]

[0031] where only the reflection (transmission) angle information of users is used when calculating , and is the number of users served by IOS reflection (transmission). Obviously, the above formula can always satisfy the energy conservation constraint. Although and may not satisfy the constraint P1, according to Proposition 1, they do not affect the distribution of . These closed-form solutions mean that the transmission of multiple data streams in data transmission is no longer dependent on the base station, but is equivalent to the phase modulation characteristics of IOS. To further obtain the explicit expression of P2, we introduce another proposition: Proposition 2: Substitute and into P2, and the constraint can be approximated as:

[0032] where According to Proposition 2, P3 can be transformed as:

[0033]

[0034] However, due to and exist There is coupling in it. This remains a non-convex problem. We introduce two types of slack variables: .at this time, It can be equivalently transformed into:

[0035]

[0036]

[0037]

[0038] in, Let k be the coordinates of user k, for processing The nonconvexity of , we use The upper bound replaces it, that is:

[0039] at this time, It can be converted into:

[0040]

[0041]

[0042]

[0043] By iteratively solving “P2→P3→P4→P5→P6”, until the increment of the objective function (total power) corresponding to the IOS position optimization is less than a set threshold, the IOS position at this point can be considered as the configurable optimal position parameter.

[0044] Step 230: In the first N-1 time slots of the communication protocol, the BS beamforming and IOS coefficients (reflection coefficients) are optimized based on alternating optimization, penalized dual decomposition, and first-order Taylor expansion design. and transmission coefficient ).

[0045] Specifically, the optimization problem corresponding to this stage is P7:

[0046]

[0047]

[0048]

[0049] Subsequently, given P7 is simplified to P8:

[0050]

[0051] Note: Then you can The constraints are rewritten as: ,in, , , , This indicates block diagonal operations. (or The The diagonal matrix is , for A matrix consisting entirely of zeros. Similarly, by... to Perform a first-order Taylor expansion to address its non-convexity:

[0052] in express The real part. Therefore. It can be converted into:

[0053]

[0054] in, Represents trace operation. and It is a block diagonal matrix. and These are the beamforming vector and signal-to-noise ratio threshold from the previous iteration, respectively. To take the real part of a complex number, For users The noise power. Clearly, It is a convex problem, which can be solved using Continuous Convex Approximation (SCA) and CVX tools.

[0055] Then, given Time optimization and at this time, It can be simplified to:

[0056]

[0057]

[0058]

[0059] because Objective function and and Irrelevant This is a feasibility study issue. There are:

[0060] By using the given Time optimization A similar process can be used to... The first constraint in the equation is transformed into:

[0061] in, , , The next step is to process... The non-convexity of the second and third constraints.

[0062] Define auxiliary variables and order .but It can be rewritten as:

[0063]

[0064]

[0065]

[0066]

[0067] Then, the augmented Lagrangian problem is constructed using the Penalty Dual Decomposition (PDD) framework, and the problem is solved by adding a penalty term to the objective function to remove [the negative variables]. The equality constraints yield:

[0068]

[0069]

[0070]

[0071] in, As a penalty factor for violating constraints, Let be the Lagrange dual variable. When At that time, the penalty term approaches zero, and the constraint... The approximate answer is true. Furthermore, under the Mangasarian-Fromovitz constraint specification, the optimal KTT solution can be obtained by alternately optimizing the original variables, dual variables, and penalty factors. Further optimization using block coordinate descent (BCD) can then be performed. The process continues until the increase in the total transmit power of the base station is less than a set threshold. At this point, the BS beamforming and IOS reflection coefficient... and transmission coefficient This can be considered as the configurable optimal network parameters.

[0072] Step 240: After the two-stage optimization is completed, this cycle is repeated. This repetitive process continues to adapt to the dynamically changing propagation environment.

[0073] Beneficial effects According to specific embodiments provided by the present invention, the present invention achieves the following technical effects: This invention addresses the need for IOS position updates during long-duration communication by proposing a novel concept of transparent transmission. This method deploys the IOS on a movable track within a wall to mitigate signal attenuation caused by wall penetration, and jointly optimizes IOS position, BS beamforming, and IOS coefficients. Furthermore, we propose a dynamic IOS transmission protocol where the IOS position is updated every N time slots, and employ a carefully designed innovative solution. Simulation and comparative results demonstrate that this scheme significantly reduces BS transmit power with a minimal number of IOS units, providing guidance for future applications of dynamic IOS-assisted transparent transmission. Attached Figure Description

[0074] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0075] Appendix Figure 1 This is the implementation process of the present invention.

[0076] Appendix Figure 2 This is a model diagram of the embedded dynamic IOS auxiliary network system designed in this invention.

[0077] Appendix Figure 3This is a conceptual diagram of the transmission protocol in the dynamic IOS-assisted network designed in this invention.

[0078] Appendix Figure 4 To the attached Figure 9 This is a comparison chart of simulation results between this scheme and the BCD&SDP (Semi-Definite Program) scheme, the RIS&ITS (Intelligent Transmissive Surface) scheme, and the fixed-position scheme.

[0079] Appendix Figure 4 These are simulation graphs comparing the convergence of various schemes.

[0080] Appendix Figure 5 This relates the minimum transmit power of each scheme to the number of IOS components.

[0081] Appendix Figure 6 This relates the minimum transmit power of each scheme to the number of users.

[0082] Appendix Figure 7 Comparison of minimum transmit power for each scheme under different ratios of transmitting users to all users.

[0083] Appendix Figure 8 This is a comparison of the relationship between the minimum transmit power and the number of slots for each scheme within a transmission cycle.

[0084] Appendix Figure 9 This is a comparison of the minimum transmit power and minimum SINR requirements for each scheme. Detailed Implementation

[0085] The steps and processes of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the examples described in this application are merely one application scenario of the present invention, and other results based on the content of the present invention without making substantial changes are within the protection scope of the present invention.

[0086] Appendix Figure 2 This is an example scenario provided by the present invention, illustrating all the devices included in the present invention. In this example scenario, it includes a device equipped with A base station with one antenna, K single-antenna user equipment, and a system containing... Intelligent omnidirectional metasurfaces with individual units These represent the number of elements in the horizontal and vertical directions, respectively. This intelligent omnidirectional metasurface is deployed on a low-cost track embedded in a wall. The base station can periodically control the track to change the position of the intelligent omnidirectional metasurface, thereby providing better channel conditions to adapt to dynamic environments. Both the base station antenna spacing and the intelligent omnidirectional metasurface element spacing are half a wavelength.

[0087] AppendixFigure 2 One example scenario provided is a single base station multi-user scenario. The communication between the user and the base station adopts a phased and instantaneous update transmission process proposed in this invention, which is based on IOS position optimization based on successive convex approximation, relaxation variable introduction and first-order Taylor expansion, as well as BS beamforming and IOS coefficient optimization based on alternating optimization, penalized dual decomposition and first-order Taylor expansion.

[0088] This invention uses Ricean fading to model the channel. The channel gain from the base station to the intelligent omnidirectional metasurface (IOS) is expressed as: in, For single-layer wall penetration loss, Free space path loss per unit distance. This is the path loss index. Rice's fading factor, , and These are the line-of-sight (LoS) component and the non-line-of-sight (NLoS) component, respectively. The model is a zero-mean, unit-variance complex Gaussian random matrix with independent and identically distributed elements, i.e. The calculation formula is:

[0089] in, The elevation and azimuth angles from the base station to the IOS are given.

[0090]

[0091] For Kronecker product, This indicates the transpose conjugate operation. This is the departure angle from the base station to the IOS.

[0092] Similarly, the reflection / transmission channel gain from IOS to user k is expressed as:

[0093] in, From iOS to users The elevation and azimuth of the departure angle.

[0094] In addition, the coefficient of IOS is defined as According to the law of conservation of energy, we have Furthermore, the transmission and reflection phase shifts of actual iOS hardware must meet the following requirements: .user The signal-to-noise ratio (SINR) is:

[0095] in, For diagonalization operations, For user sets, For users The power of additive white Gaussian noise.

[0096] This invention considers green communication and aims to minimize the BS transmit power while meeting the SINR requirements of all users by jointly optimizing BS beamforming, IOS coefficients, and IOS location. To this end, we have designed the following... Figure 2 The communication protocol shown has N time slots in one transmission cycle.

[0097] Specifically, the IOS has an initial position, or a position optimized in the previous transmission cycle. Next, in the first N-1 time slots, the BS beamforming and IOS coefficients are optimized based on the known IOS position. Then, in the nth time slot, the IOS position is optimized again. Finally, the above process is repeated to achieve long-term high-speed communication.

[0098] The logic of this design is that BS beamforming and IOS coefficients can be optimized based on instantaneous CSI, while the IOS position can only be optimized through statistical CSI. Furthermore, the IOS position cannot be moved instantaneously; moving the IOS in every time slot is impractical. Therefore, we move it once every N time slots.

[0099] Based on the above analysis, we consider the following optimization problem for time slots T (T≥2N):

[0100]

[0101] Where [t] represents the value of a variable in time slot t, t = {1, 2, ..., t}. These formulas respectively consider the user's minimum SINR requirement, the hardware constraints of the IOS, and the optimization space of the IOS location. The above optimization problem remains unsolved. According to the designed transmission protocol, problem P1 can be divided into two independent sub-problems: In the nth time slot, according to the pre-designed appropriate { }and{ }, statistical CSI optimization q for i∈{t, r}; In the first N-1 time slots, based on the q optimized in the previous time slot, optimize { respectively }and{ }, i∈{t, r}. Therefore, we adopt a two-stage method to solve P1: the first stage is to obtain the IOS position, and the second stage is to solve the BS beamforming and IOS coefficients.

[0102] First, IOS location optimization is performed. In this case, IOS location optimization is only performed in time slot N, so for simplicity, [t] is ignored. P1 is simplified to

[0103]

[0104] Due to the objective function and optimization variables Irrelevant This is a feasibility verification problem. However, feasibility verification rarely yields an optimal solution. To address this problem, we... Perform normalization processing, so that ,in , This is the power allocated to user k. At this time, It can be equivalently transformed into:

[0105]

[0106] However, due to the lack of explicit expressions, and and There is a coupling relationship. The solution remains elusive. To resolve this issue, an analysis is necessary. The distribution characteristics. We will Rewritten as:

[0107] in, ,

[0108] The approximation relationship holds because of the product of the two non-line-of-sight components. With a mean of zero and a variance of 1, the values ​​are extremely small and can be ignored in expectation calculations. Therefore, It can be approximated as:

[0109] To further simplify, we introduce the following proposition: Proposition 1: Follows a complex Gaussian distribution And with Irrelevant, only related to Relevant. According to Proposition 1, No impact The distribution of , but it will affect its value, and thus affect This coupling relationship makes Constraints are even more difficult to handle. To address these difficulties, we attempt to propose a suboptimal solution. and to decouple constraints In and It should be noted that, due to , and exist The relationship between China and the United States is complex and difficult to adapt to any situation. optimal .exist In the middle, we pay more attention to the iOS location. right Due to the influence of angle information, a low-complexity suboptimal closed-form solution is designed as follows:

[0110]

[0111] Among them, calculation Only the angle information of the user is used for reflection (transmission). The number of users serving IOS reflection (transmission). Clearly, the above equation always satisfies the energy conservation constraint. Although and The phase shifts may not satisfy constraint P1, but according to proposition 1, they do not affect The distribution of these closed-form solutions means that the transmission of multiple data streams in data transmission no longer depends on the base station, but is equivalently dependent on the phase modulation characteristics of the IOS. To further obtain an explicit expression for P2, we introduce another proposition: Proposition 2: and Substituting P2, the constraint can be approximately expressed as:

[0112] in, According to Proposition 2, P3 can be transformed into:

[0113]

[0114] However, due to and exist There is coupling in it. This remains a non-convex problem. We introduce two types of slack variables: .at this time, It can be equivalently transformed into:

[0115]

[0116]

[0117]

[0118] Among them, for processing The nonconvexity of , we use The upper bound replaces it, that is:

[0119] at this time, It can be converted into:

[0120]

[0121]

[0122]

[0123] At this point, For convex problems, the Continuous Convex Approximation (SCA) technique and CVX tool can be used for solving. The specific process is as follows: Step 300: Perform initialization settings and set the initial location of iOS. Iteration stopping precision Initialize the number of iterations. ; Step 310: Calculate the initial objective value of the convex optimization problem P6. .

[0124] Step 320: Perform iterative optimization, setting the number of iterations... Based on the position of the previous iteration Solve the convex optimization problem P6 in the current iteration to obtain the new IOS position. and the corresponding target value ; Step 330: Perform a convergence check. If the target values ​​before and after the iteration satisfy... If the iteration stops, output the current position. If the solution is the optimal one, then return to the iterative optimization step and continue iterating until the convergence condition is met.

[0125] Next, BS beamforming and IOS coefficient optimization are performed. Unlike most existing studies, this invention does not use the semidefinite programming (SDP) method with Gaussian randomization to solve the IOS coefficients, but instead develops a more efficient semi-closed solution.

[0126] According to the designed transmission protocol, in the initial Within each time slot, the location of the IOS is known and fixed. Therefore, in this... Minimizing the total transmit power of the base station in a given time slot is equivalent to minimizing the transmit power of the base station in each time slot. For any time slot, we can derive the following equivalent optimization problem from problem P1 (symbols omitted here for brevity). ):

[0127]

[0128]

[0129]

[0130] In this context, it is assumed that the instantaneous channel state information (CSI) is perfect, therefore [the following is omitted]. The expectation. Obviously, The objective function is about The function is convex, but all constraints are non-convex. We use alternating optimization (AO) technique to solve it. .

[0131] First, given Time optimization at this time, It can be simplified to:

[0132]

[0133] Note: Then you can The constraints are rewritten as: ,in, , , , This indicates block diagonal operations. (or The The diagonal matrix is , for A matrix consisting entirely of zeros. Similarly, by... to Perform a first-order Taylor expansion to address its non-convexity:

[0134] in express The real part. Therefore. It can be converted into:

[0135]

[0136] in, Represents trace operation. and It is a block diagonal matrix. and These are the beamforming vector and signal-to-noise ratio threshold from the previous iteration, respectively. To take the real part of a complex number, For users The noise power. Clearly, It is a convex problem, which can be solved using Continuous Convex Approximation (SCA) and CVX tools.

[0137] Then, given Time optimization and at this time, It can be simplified to:

[0138]

[0139]

[0140]

[0141] because Objective function and and Irrelevant This is a feasibility study issue. There are:

[0142] By using the given Time optimization A similar process can be used to... The first constraint in the equation is transformed into:

[0143] in, , , The next step is to process... The non-convexity of the second and third constraints.

[0144] Define auxiliary variables and order .but It can be rewritten as:

[0145]

[0146]

[0147]

[0148]

[0149] Furthermore, a penalized dual decomposition (PDD) framework is adopted to construct... The augmented Lagrange problem is solved by adding a penalty term to the objective function to remove... Equality constraints:

[0150]

[0151]

[0152]

[0153] in, As a penalty factor for violating constraints, Let be the Lagrange dual variable. When At that time, the penalty term approaches zero, and the constraint... The approximate answer is yes. Furthermore, under the Mangasarian-Fromovitz constraint, the optimal KTT solution can be obtained by alternately optimizing the primal variables, dual variables, and penalty factor. Therefore, we employ the block coordinate descent (BCD) method for alternating optimization.

[0154] about The subproblems are:

[0155]

[0156] This subproblem only contains There are convex constraints, and the objective function is a convex function, which can be solved using the CVX tool.

[0157] about The subproblems are:

[0158]

[0159]

[0160] in The objective function can be rewritten as:

[0161] It is evident that optimizing variables Only exists in the third item China. (General) Rewritten as ,in , Then to and Perform alternating optimization.

[0162] about The subproblems are:

[0163]

[0164]

[0165] in , * indicates conjugate operation. Based on the expression of the objective function, It can be divided into M independent subproblems. The first constraint can be transformed into:

[0166] Substituting the above equation into... We can obtain:

[0167]

[0168] The closed-form solution is:

[0169] in express The phase. From this we can obtain The closed-form solution.

[0170] about The subproblems are:

[0171]

[0172]

[0173] in, .like It only requires superimposing their phases. This can be converted to the range of 0 to 1, therefore The second constraint does not affect the optimality of the problem. Note that... ,in . based on Given the constraints, let: in .therefore, The objective function can be rewritten as:

[0174] in, Minimizing the above expression, we get:

[0175] in, It is possible Calculations show that This is a sign function. Therefore, we can obtain... The closed-form solution can be obtained. Therefore, by alternately solving P13, P15, and P17, the solution can be obtained given... In the case of obtaining The optimal solution.

[0176] Combining the two sub-parts mentioned above and utilizing the alternating optimization technique, the overall optimization algorithm for P7 can be obtained, as shown in Algorithm 2. Dual variables and penalty factor The algorithm is updated according to the PDD framework, and Algorithm 2 converges to the Karush-Kuhn-Tucker (KTT) solution. Therefore, Algorithm 2 can obtain at least one stable local optimum.

[0177] The specific process of Algorithm 2 is as follows: Step 400, Initialize variables Set update coefficient ; Step 410, update by solving P9 ; Step 420, update by solving P13

[0178] Step 430, based on the transmission amplitude coefficient of equation IOS and reflection amplitude coefficient Using angle variables Relationships and angle variables value expression update ; Step 440, based on the relationship between the IOS reflection phase and the IOS transmission phase coefficient. Closed-form solution update ; Step 450, Calculate ;like Then update the dual variable: Otherwise, update the parameters: Until Less than the preset threshold.

[0179] The following explains some of the resulting images obtained in this example scenario.

[0180] Simulation results are attached. Figure 4 To the attached Figure 9 As shown. (Attached) Figure 4 The convergence of the proposed method is demonstrated. Clearly, the convergence value of the proposed scheme outperforms other benchmark schemes. Furthermore, a noteworthy finding is that the proposed scheme exhibits a faster convergence speed compared to the BCD&SDP scheme. This is because the optimal IOS coefficients provided in the method of this invention are derived in a closed-form solution, thereby eliminating the need for CVX-based solutions.

[0181] Appendix Figure 5 The performance of the four schemes was compared from the perspective of the number M of intelligent omnidirectional metasurface units. It can be seen that when... At that time, the proposed solution achieved a minimum transmit power of 21.2 dBm for the base station. A horizontal comparison shows that to achieve this transmit power, the number of IOS units required by the BCD&SDP and RIS&ITS solutions would increase by 40.8% and 65.4%, respectively. This indicates that, comparatively, the proposed solution offers both cost-effectiveness and efficiency advantages. Furthermore, a vertical comparison shows that when… At that time, the minimum transmit power required by the base station in the fixed-location scheme is 25.5 dBm, which is 4.3 dB higher than the scheme proposed in this invention. This gap will widen if the number of IOS units is further increased. For example, when... At that time, the power difference reached 5.9 dB.

[0182] Appendix Figure 6 This demonstrates the impact of environmental changes, such as the number of users K, on ​​the base station's transmit power requirements. Two scenarios are considered: one is environmental parameters... That is, the number of users transmitting data via the intelligent omnidirectional metasurface is 0; secondly... 0.5, meaning the distribution of reflecting and transmitting users in the IOS is uniform. From a macroscopic perspective, the proposed solution performs best. Especially when At this time, because the reflecting user is closer to the base station, the base station requires less transmission power. From a microscopic perspective, the performance differences between different schemes are significant. For example, when the number of users is 10, as environmental parameters change from... arrive The base station transmit power of the fixed location scheme increased by 1.4 dB, while the increase in transmit power of the scheme proposed in this invention is relatively small.

[0183] Appendix Figure 7 Using the user distribution ratio as the independent variable, the impact of user distribution on base station transmit power was compared in more detail in a scenario with 8 users. The results show that the performance fluctuation of the proposed solution is relatively small as the proportion of transmitting users changes, which is consistent with the solution's ability to perform optimal configuration for different user scenarios. Under all user distribution ratios, the proposed solution requires the minimum transmit power.

[0184] Appendix Figure 8 The advantages of the proposed scheme are highlighted from the perspective of protocol parameter configuration. As the number of time slots N in a transmission cycle increases, the minimum transmit power required by the base station to maintain the same user transmission rate also increases, eventually stabilizing. This is because a larger number of time slots in a transmission cycle means a lower update frequency for the IOS location, resulting in a lower match between the IOS location and the current communication environment. Therefore, the base station needs higher transmit power to maintain the user's transmission rate. A comparison of different schemes shows that the proposed scheme has the lowest requirement for the IOS location update frequency. Specifically, with a base station power budget of 21.2 dBm, the proposed scheme only needs to update the IOS location once every 1000 time slots, while the RIS&ITS scheme and the fixed-location scheme require an update frequency increase of more than 5 times and more than 3 times, respectively. This demonstrates the stronger robustness, lower latency sensitivity, and higher practicality of the proposed scheme. Furthermore, when N increases from 100 to 1000, the transmit power of the proposed scheme increases by only 2.5 dB, while the transmit power of the fixed-location scheme increases by 3.1 dB.

[0185] Appendix Figure 9 This demonstrates the different signal-to-noise ratio requirements of users. The minimum transmit power required for the base station is determined. Unsurprisingly, the required transmit power increases with user SINR demands. Notably, compared to other benchmark schemes, the proposed scheme achieves transmit power that is less sensitive to user SINR demands, making it more energy efficient. For example, when... When the transmit power is 1 dB, the fixed-position scheme has a transmit power 1.4 dB higher than the proposed scheme; when When the value is 10 dB, this gap widens to 11.2 dB. Furthermore, the BCD&SDP and RIS&ITS schemes also require higher transmit power compared to the proposed scheme, albeit by different amounts.

Claims

1. A method for dynamic omni-directional metasurface assisted indoor transmission optimization, characterized in that, Comprise: By deploying IOS on a movable track on the wall, the base station (BS) changes the IOS position by controlling the track to provide a better transmission link, achieving a transparent transmission with wall penetration loss approaching zero; A phased and instantaneous updating transmission protocol is proposed, in which the IOS position is updated every N time slots, and the IOS coefficient and BS beamforming are updated every time slot; An optimization problem is established to achieve long-term green communication, which jointly optimizes the IOS position, IOS coefficient and BS beamforming, minimizes the BS transmit power while meeting user demand, and uses a two-stage method to solve.

2. The method of claim 1, wherein, The transmission protocol adopts a mechanism combining phased and instantaneous updating: a transmission period contains N time slots, in which the IOS position is updated every N time slots, and the IOS coefficient and base station beamforming are updated every time slot.

3. The transmission process according to claim 2, specifically comprising two stages: the first stage is the first N-1 time slots, and the IOS starts working at the initial position (or the position optimized in the last period), and the semi-closed solution of the IOS coefficient and BS beamforming is solved by alternating optimization and penalty dual decomposition technology to ensure that the signal-to-interference plus noise ratio of the user in the current time slot meets the demand.

4. The first stage of claim 3, wherein Based on alternating optimization, penalized dual decomposition and first-order Taylor expansion, the BS beamforming, IOS coefficients (reflection coefficients and transmission coefficients ) are designed, where the joint optimization problem is split into two serial sub-problems, i.e., BS beamforming optimization with fixed IOS coefficients (corresponding to the problems P8→P9 in the specification) and IOS coefficients optimization with fixed BS beamforming (corresponding to the problems P10→P17 in the specification), which are solved iteratively by alternating "P8→P9 and P10→P17", until the increment of the base station total transmit power is less than a set threshold, at which time the BS beamforming, IOS reflection coefficients and transmission coefficients can be considered as the configurable optimal network parameters.

5. The method of claim 4, wherein, The optimization process uses the block coordinate descent (BCD) method to jointly solve the IOS coefficient optimization subproblem and the BS beamforming subproblem, where each subproblem has a convex objective function and a convex constraint, and can be iteratively converged to the optimal or suboptimal solution by convex optimization tools (such as CVX).

6. The transmission process according to claim 2, the second stage is the Nth time slot, and the IOS position is updated based on statistical CSI using successive convex approximation technology, thereby providing better channel conditions for communication in the next period, and the above process is executed in a loop to achieve long-term high-rate communication.

7. The second stage of claim 6, wherein, The second stage of the two-stage optimization method is to design an IOS position based on successive convex approximation, relaxation variable introduction and first-order Taylor expansion, wherein the IOS position optimization problem is split into multiple serial sub-problems, i.e., from the original problem P2 to the power normalization P3, and then to P4 and P5 by introducing a relaxation variable, and finally to the convex optimization problem P6, and the "P2→P3→P4→P5→P6" is solved iteratively until the IOS position optimization corresponding objective function increment is less than a set threshold, at which time the IOS position can be considered as the optimal configurable position parameter, and specifically comprising: converting the original feasibility check problem into a total power minimization problem, introducing a relaxation variable The distance power of the base station to the IOS and the distance power of the IOS to the user are respectively approximated (both jointly quantify the signal attenuation degree of the base station-IOS-user transmission link), the non-convex product term is approximated by the first-order Taylor expansion, the IOS position is iteratively updated by the successive convex approximation technique, and the iteration is stopped until convergence.

8. The method of claim 7, wherein, When optimizing the IOS position, the relaxation variable and the first-order Taylor expansion method are introduced to convert the original non-convex optimization problem into a series of subproblems that can be solved by successive convex approximation (SCA), thereby effectively reducing the solution complexity and improving the algorithm stability and convergence speed.

9. The method of claim 7, wherein, the relaxation variable satisfies: where is the base station position, is the user k position, q is the IOS position.

10. The method of claim 1, wherein, The channel model adopts Rician fading modeling, and the channel gain expression from the base station to the IOS is: The reflection / transmission channel gain expression from the IOS to user k is: wherein, is the single-layer wall penetration loss, is the free space path loss per unit distance, is the path loss exponent, is the Rician fading factor, is the distance from the base station to the IOS, is the distance from the IOS to user k, is the line-of-sight (LoS) component, is the non-line-of-sight (NLoS) component.

Images