Aerial dynamic intelligent reflector auxiliary network parameter configuration method

Abstract

The invention provides an auxiliary network parameter configuration method for an aerial dynamic intelligent reflecting surface, which is characterized in that an airborne reconfigurable intelligent surface (RIS-UAV) of an unmanned aerial vehicle is used for reflecting a signal of local first-type receiving and transmitting equipment (RFT1), and a data transmission service is provided for second-type receiving and transmitting equipment (RFT2). Specifically, the invention provides a communication process of an RIS-UAV, RFT1 and RFT2 with RFT1 as a computing center and a control center, establishes a three-dimensional air-ground channel model containing angle errors, and designs an RFT1 beam forming matrix, an RIS phase matrix and a UAV three-dimensional position optimization method based on iterative optimization and first-order Taylor expansion. And an RFT1 beamforming matrix, an RIS phase matrix, a UAV three-dimensional position updating mechanism and a decision signaling process based on the maximum percentage of user performance reduction are given.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to 3D dynamic deployment of aerial reconfigurable intelligent surfaces (ARISs) in wireless communication networks and anti-shake robustness transmission parameter configuration, including ARIS position optimization protocols, processes and methods. Background Art

[0002] In recent years, Unmanned Aerial Vehicle (UAV)-borne Reconfigurable Intelligent Surface (RIS)-assisted wireless communication has become a research hotspot in academia, because RIS has a wide range of high-gain business capabilities supported by the agility of UAV. If the ground RIS cannot be deployed in a location with a high Line of Sight (LoS) link to both the base station (BS) and the user, the interruption probability is very high. In addition, UAV can not only provide a high LoS link for RIS, but also reduce the reflection path loss through location optimization. Combining the above disadvantages and advantages, the combination of UAV and RIS, namely RIS-UAV, is a very promising new network element in the future 6G and has very important applications.

[0003] However, there are still many problems to be solved in the actual deployment and application of RIS-UAV-assisted networks. On the one hand, the communication performance of RIS-UAV is largely affected by the accuracy of air-to-ground (A2G) channel modeling. An accurate model is the basis for high-performance network deployment solutions. At present, most of the literature adopts the following three channel models: LoS channel model, probabilistic LoS channel model and Rician fading channel model. However, in order to facilitate the solution of optimization problems, most of the existing literature adopts fixed channel parameters, without considering the changes in channel parameters caused by the three-dimensional motion of RIS-UAV, so it is impossible to accurately characterize the real channel conditions of RIS-UAV-assisted networks.

[0004] On the other hand, most of the existing RIS-UAV communication works assume that the channel state information (CSI) of A2G is perfect. However, unlike ground-fixed cellular base stations, airflow and motor vibration can cause certain random deviations in the flight trajectory of the UAV platform in the horizontal or vertical direction, resulting in channel estimation errors, communication beam mismatch, and reduced communication rate. Therefore, it is still an open and important issue to establish an accurate three-dimensional A2G channel model with UAV jitter and further study the RIS-UAV-assisted network deployment scheme. Summary of the invention

[0005] The present invention proposes an aerial dynamic intelligent reflector-assisted network parameter configuration method, which utilizes RIS-UAV to reflect the signal of the local first-kind radio frequency transceiver (RFT1) to provide data transmission services for the second-kind radio frequency transceiver (RFT2). Specifically, the present invention proposes the communication process of RIS-UAV, RFT1 and RFT2, establishes a 3D air-to-ground channel model including angle errors, designs the RFT1 beamforming matrix, RIS phase matrix and UAV three-dimensional position optimization method based on iterative optimization and first-order Taylor expansion, and provides the RFT1 beamforming matrix, RIS phase matrix and UAV three-dimensional position update mechanism and decision signaling process based on "maximum percentage of user performance degradation".

[0006] The method for configuring parameters of an aerial dynamic intelligent reflector auxiliary network proposed by the present invention is as shown in the attached figure. Figure 1 As shown, it mainly includes steps 200 to 240.

[0007] Step 200, the proposed communication process of RIS-UAV, RFT1 and RFT2 is as shown in the attached Figure 2 shown.

[0008] The communication process includes two stages, namely the initialization stage and the normalization stage. In the initialization stage, RFT2 reports its own position information; RFT1 calculates network parameters including the beamforming matrix, RIS phase matrix and UAV three-dimensional position of RFT1 according to the 3D air-ground channel model including the angle error established in step 210; then, RFT1 updates its own beamforming matrix and sends control information to RIS-UAV to update the RIS phase matrix and the three-dimensional position of UAV; then, it enters the normalization stage. In the normalization stage, RFT2 reports its position information and current rate information; RFT1 evaluates whether it is necessary to update the network parameters including the beamforming matrix, RIS phase matrix and UAV three-dimensional position of RFT1 according to the rate information of all RFT2; if it needs to be updated, the network parameters including the beamforming matrix, RIS phase matrix and UAV three-dimensional position of RFT1 are calculated according to the 3D air-ground channel model including the angle error established in step 210; if it does not need to be updated, no processing is performed. Then, RFT2 reports its position information and current speed information, and repeats the above process. RFT1, RFT2 and RIS-UAV communicate with each other in an unlimited manner.

[0009] In step 210, the established 3D air-to-ground channel model including the angle error is shown below.

[0010] It should be noted that the present invention ignores the direct link between RFT1 and RFT2, and only considers the reflection channel between RFT1 and RIS-UAV, and between RIS-UAV and RFT2. In addition, the channel model of the present invention considers the influence of the angle error caused by position movement and UAV jitter on the channel parameters, so the calculation result is more in line with the actual situation and more robust.

[0011] At time t, the channel H between RFT1 and RIS-UAV BR [t] can be calculated as follows

[0012]

[0013]

[0014]

[0015]

[0016] Where, β0 is the path loss per unit distance, d BR [t] is the distance between RFT1 and RIS-UAV; is the path loss index between RFT1 and RIS-UAV, α max and α minare the maximum and minimum values ​​of the path loss index, a1 and b1 are constants determined by the environment; is the Ricean factor of the channel between RFT1 and RIS-UAV, K max and K min is the maximum and minimum value of the Rice factor; are independent and identically distributed matrices that follow a circularly symmetric complex Gaussian distribution; H BR The first ((n) in [t] x +1)×(n y +1),m+1) elements are n x =0, 1, ..., N x -1,n y =0, 1, ..., N y -1,m=0,2,…,M-1,N x is the number of units in the horizontal dimension of RIS, N y is the number of units in the vertical dimension of RIS, M is the number of antennas of RFT1, θ BR [t] is the pitch angle of RFT1 relative to RIS-UAV, is the azimuth of RFT1 relative to RIS-UAV; 1N ×M is an all-1 matrix of N×M dimensions; A BR [t]'s ((n x +1)×(n y +1),m+1) elements are Δθ BR [t] is the pitch angle jitter error of RFT1 relative to RIS-UAV; B BR [t]'s ((n x +1)×(n y +1),m+1) elements are is the azimuth jitter error of RFT1 relative to RIS-UAV; definition

[0017] At time t, the channel between RIS-UAV and RFT2 It can be calculated as follows

[0018]

[0019]

[0020]

[0021]

[0022] in, is the distance between RIS-UAV and RFT2; is the path loss index between RIS-UAV and RFT2, is the Ricean factor of the channel between RIS-UAV and RFT2; are independent and identically distributed vectors that follow a circularly symmetric complex Gaussian distribution; H BR The first ((n) in [t] x +1)×(n y +1)) elements are n x =0, 1, ..., N-1, n y =0, 1, ..., N y -1, is the pitch angle of RFT2 relative to RIS-UAV, is the azimuth of RFT2 relative to RIS-UAV; The first ((n x +1)×(n y +1)) elements are is the pitch angle jitter error of RFT2 relative to RIS-UAV; The first ((n x +1)×(n y +1)) elements are is the azimuth jitter error of RFT2 relative to RIS-UAV; definition

[0023] Step 220, optimize the RFT1 beamforming, RIS phase and UAV 3D position based on iterative optimization and first-order Taylor expansion design. In this step, we focus on the optimization at a certain moment, so the time variable t is omitted.

[0024] Among K RFT2 devices, the goal of the present invention is to maximize the transmission rate of the smallest RFT2 device by designing the optimal RFT1 beamforming, RIS phase, and UAV 3D position.

[0025] The RFT1 beamforming matrix V is obtained by solving the following problem using the Convex Optimization Toolbox:

[0026]

[0027] Among them, λ k is the slack variable, v kis the beamforming vector of RFT1 to the kth RFT2, is the phase matrix of RIS, φ=[φ1,…,φ N ] T ,φ n is the phase of the nth RIS unit, N is the number of RIS units, and N=N x N y , Blkdiag{·} is a block diagonal operation, P max is the maximum transmit power of RFT1, and the superscript [r] indicates the rth iteration value of the variable. is the noise power, Represents the real part of x.

[0028] The RIS phase matrix φ is obtained by solving the following problem using the convex optimization toolbox:

[0029]

[0030] Among them, γ k is the slack variable, Γ n is φ n The amplitude, ρ n is φ n phase.

[0031] The 3D position of the UAV is obtained by solving the following problem using the Convex Optimization Toolbox:

[0032]

[0033]

[0034] in, and f k,l (H R ) are about and H R The lower bound of τ k is the slack variable, and are the actual deployable space for the UAV’s horizontal position and altitude, respectively.

[0035] By iteratively solving "P1→P2→P3→P4→P1→P2→..." until the increment of η is less than a set threshold, the RFT1 beamforming, RIS phase and UAV three-dimensional position at this time can be considered as the configurable optimal network parameters.

[0036] Step 230, based on the principle of "maximum percentage of user performance degradation", determine whether it is necessary to update the RFT1 beamforming, RIS phase and UAV position.

[0037] According to the current transmission rate reported by each RFT2, RFT1 calculates its rate drop relative to the previous Q time slots, and calculates the average rate drop percentage. If the average rate drop percentage of more than X% of RFT2 exceeds Y%, the RFT1 beamforming, RIS phase and UAV position are updated and step 220 is performed; otherwise, no update is performed and no operation is performed.

[0038] The values ​​of Q, X, and Y can be preset according to actual conditions.

[0039] Step 240, the decision signaling process of RFT1 beamforming, RIS phase and UAV position update is as shown in the attached figure. Figure 3 shown.

[0040] After RFT1 calculates the optimal RFT1 beamforming, RIS phase, and UAV position based on the position information reported by all RFT2s at the last moment, it sends the calculated RIS phase and UAV position information to RIS-UAV. After receiving the information, RIS-UAV sends an ACK message to RFT1 and updates the RIS phase and its own position information. If RIS-UAV does not receive the message, RFT1 will continue to retransmit the information to RIS-UAV. After RIS-UAV updates the RIS phase and its own position, it will send an ACK message to RFT1 to ensure information synchronization.

[0041] Beneficial Effects

[0042] The invention discloses a method for configuring network parameters of an aerial dynamic intelligent reflective surface-assisted network. The basic device composition is introduced, and a new network element is introduced, namely, an intelligent reflective surface mounted on an unmanned aerial vehicle (RIS-UAV), which is composed of a RIS and a UAV. Through the reporting of the current transmission rate of RFT2 and the calculation of the average rate drop by RFT1, it can be determined whether the network parameters need to be updated, including RFT1 beamforming, RIS phase, and UAV position.

[0043] Furthermore, the present invention designs the communication process between RFT1, RFT2 and UAV-RIS, and deploys the central control and computing unit in RFT1 to ensure the computing processing and real-time control capabilities of the network. Furthermore, the 3D channel model established by the present invention takes into account the channel angle error and can provide more accurate input data for network parameter updates.

[0044] In addition, the present invention designs the decision signaling process of RFT1 beamforming, RIS phase and UAV position update, and through the RIS-UAV double ACK feedback mechanism, the execution of signaling and the synchronization of information can be guaranteed.

[0045] The aerial dynamic intelligent reflector auxiliary network parameter configuration method of the present invention can be applied to various outdoor communication systems, including UAV-RIS auxiliary network, UAV auxiliary network and other aerial communication systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to clearly explain the technical steps of the present invention, all the drawings used in the description of the present invention are briefly described below. It should be noted that the drawings described below are only some examples of the implementation of the present invention, and other ordinary technicians in this field can still obtain other drawings in other different scenarios based on these drawings.

[0047] Attached Figure 1 It is the implementation process of the present invention;

[0048] Attached Figure 2 It is the communication process of RIS-UAV, RFT1 and RFT2 proposed by the present invention;

[0049] Attached Figure 3 It is the decision signaling process of RFT1 beamforming, RIS phase and UAV position update designed by the present invention;

[0050] Attached Figure 4 This is an example scenario provided by the present invention;

[0051] Attached Figure 5 It is the relationship between the minimum transmission rate and the maximum value and the number of iterations in the strength scenario of the present invention;

[0052] Attached Figure 6 It is the optimization result of the proposed scheme and the baseline scheme under the strength scenario of the present invention;

[0053] Attached Figure 7 It is the relationship between the minimum transmission rate maximum value and the angle jitter amount under the strength scenario of the present invention;

[0054] Attached Figure 8 It is the relationship between the minimum transmission rate maximum value and the number of RIS units under the strength scenario of the present invention. DETAILED DESCRIPTION

[0055] The steps and processes of the present invention are described completely and clearly below in conjunction with the drawings in the present application. It is obvious that the example described in the present application is only an example application scenario of the present invention, and other results based on the content of the present invention without making substantial changes all fall within the protection scope of the present invention.

[0056] Attached Figure 4 This is an example scenario provided by the present invention, which shows all the devices included in the present invention and some parameters that will be used later. In this example scenario, the aerial reconfigurable smart surface consists of a smart reflective surface and a drone, which are deployed above an area that has no direct transmission link with the ground base station. Figure 4 The base station (BS) is the RFT1 device in the present invention, and there is a wireless transmission link between it and the RIS-UAV to control the phase of RIS and the position of UAV. The user equipment (UE) is the RFT2 device in the present invention.

[0057] Attached Figure 4 An example scenario provided is a single-base-station-multiple-user scenario. The communication between the user and the base station adopts the RIS-UAV, RFT1 and RFT2 communication processes proposed in the present invention, including the initialization phase and the normalization phase, RFT1 beamforming, RIS phase and UAV three-dimensional position optimization based on iterative optimization and first-order Taylor, as well as the decision signaling process of RFT1 beamforming, RIS phase and UAV position update.

[0058] According to the attached Figure 4 The relevant parameters for the instance scenario are: BS coordinate w BS =(2400,0,20) T , the number of UEs K = 5, and their coordinates are u1 = (-300, 150, 2) T ,u2=(-150,-200,2) T ,u3=(0,0,2) T ,u4=(250,-250,2) T ,u5=(300,200,2) T ,RIS-UAV initial and final positions q I =q F =(450,0,80), the number of BS antennas M = 8, the maximum BS transmission power P max =20dBm, process duration T D =160s, number of RIS units N = 100, unit path loss β0 = -60dBm, noise power spectral density σ 2 =-174dBm / Hz, the minimum value of Rice factor K min =0dB and maximum value K max=30dB, environmental constants a1 = 11.95, b1 = 0.14, the flight area of ​​the drone is a rectangular space, the x distance is [-450, 450], the y distance is [-300, 300], and the z distance is [80, 200]. The above distances are in meters.

[0059] Step 300: Establish a communication system model and target optimization problem in an example scenario.

[0060] In time slot t, the achievable rate R of UE k is k [t] is calculated as

[0061]

[0062] Among them, the 3D channel considering the UAV jitter error is expressed as (the time variable t is omitted for simplicity)

[0063]

[0064]

[0065] The definitions of the relevant parameters in (13)-(15) can be found in (1)-(12). The optimization problem of maximizing the minimum user rate in each time slot can be expressed as

[0066]

[0067] Among them, the first constraint is the minimum rate requirement of UE, the second constraint is the BS power budget, the third constraint is the optimization space of RIS reflection coefficient, and the fourth constraint is the feasible deployment space of UAVs.

[0068] Step 310: Based on P5, solve the beamforming matrix {v k}.

[0069] Given the RIS phase {φ n} and the UAV 3D position w R , the solution of P5 can be simplified to the following problem

[0070]

[0071] The definitions of relevant parameters can be found in (9). This problem is a convex problem and can be quickly solved by convex optimization toolboxes, such as CVX.

[0072] Step 320, based on P5, solve the RIS phase {φ n}.

[0073] Given the beamforming matrix of BS {v k} and the UAV 3D position wR , the solution of P5 can be simplified to the following problem

[0074]

[0075] The definitions of relevant parameters can be found in (10). This problem is a convex problem and can be quickly solved by convex optimization toolboxes, such as CVX.

[0076] Step 330, based on P5, solve the three-dimensional position of the UAV

[0077] Given the beamforming matrix of BS {v k} and RIS phase {φ n}, the solution of P5 can be simplified to the alternating solution of the following two sub-problems:

[0078]

[0079]

[0080] The definitions of the relevant parameters can be found in (11) and (12). However, (11) and (12) do not give and f k,l (H R ) is a specific expression. Here, Take the example to derive its specific expression. This method is also applicable to obtaining f k,l (H R ) is a specific expression.

[0081] Extensions Focusing on solving the non-convexity problem caused by channel angle dependence, we can get the following expression

[0082]

[0083] Among them, H B is the BS height, H U is the UE height, is the 2D position of BS, is the two-dimensional position of user k,

[0084] Z 1,k,l ≤(X 1,k ΦX2+X 1,k ΦY2+Y 1,k ΦX2+Y 1,k ΦY2)v l (20)

[0085]

[0086]

[0087]

[0088]

[0089] Then, we can find that Relative to Z2 and Z 3,k are not jointly convex, so at the rth iteration we can use the following approximation

[0090]

[0091] Applying the same convex approximation operation to the right sides of (21) and (23), we can obtain the following inequality:

[0092]

[0093]

[0094] Although (22) and (24) are respectively relative to δ BR and convex function, but δ BR and Neither Therefore, we use the following approximation

[0095]

[0096] in,

[0097]

[0098] Similarly, we have

[0099]

[0100] in,

[0101]

[0102] The detailed derivations of (29) and (31) are given below.

[0103] So far, only the right side of (20) is relative to is non-convex. Focusing on the right side of (20), we can find that its value is X 1,k ΦX2,X 1,k φY2,Y 1,k ΦX2,Y 1,k The affine transformation of all elements in ΦY2 is due to vl The dot product of . The radial transformation does not change the convexity. Therefore, as long as X 1,k ΦX2,X 1,k φY2,Y 1,k ΦX2,Y 1,k All elements in ΦY2 are Here, we use X 1,k The elements in ΦX2 are used as an example to illustrate how to convert them into This method also applies to X 1,k ΦY2,Y 1,k ΦX2,Y 1,k Elements in ΦY2.

[0104] It's easy to know, The mth element can be written as (its phase can be expressed by {v k} and {φ n}Alignment, omitted here)

[0105]

[0106] Among them, K r =K max / K min To deal with the non-convexity of (32), we introduce a series of slack variables,

[0107]

[0108] Among them, μ 2,k and μ3 are defined in (31) and (29), respectively.

[0109] Therefore, we can obtain

[0110]

[0111] in, is μ 1,k The non-convexity of the right side of (33)(31)(29) is due to the angular dependence of the channel. We use the first-order Taylor expansion for convex approximation.

[0112]

[0113]

[0114]

[0115] therefore, All relative to The constraints of f become convex constraints, and each element of them can be obtained using the above method. k,l (H R ) calculation method can also be obtained.

[0116] Step 340, by iteratively solving "P1→P2→P3→P4→P1→P2→...", until the increment of η is less than a set threshold, at which time the RFT1 beamforming, RIS phase and UAV three-dimensional position can be considered as configurable optimal network parameters.

[0117] Step 350: Based on the principle of "maximum percentage of user performance degradation", determine whether it is necessary to update the RFT1 beamforming, RIS phase and UAV position. At this time, the values ​​of Q, X and Y are 10, 20 and 30 respectively.

[0118] According to the current transmission rate reported by each RFT2, RFT1 calculates its rate drop relative to the previous Q time slots, and calculates the average rate drop percentage. If the average rate drop percentage of more than X% of RFT2 exceeds Y%, the RFT1 beamforming, RIS phase and UAV position are updated and step 220 is performed; otherwise, no update is performed and no operation is performed.

[0119] Step 360, repeat the above process according to the time sequence.

[0120] The following is an explanation of some of the result graphs obtained in this example scenario.

[0121] In the attached Figure 5-8 In this paper, we compared 6 different schemes: 1) 3D mobile with jitter: 3D trajectory optimization considering drone jitter and 3D reflection, i.e., our proposed scheme; 2) 3D mobile without jitter: 3D trajectory optimization with 3D reflection but without considering drone jitter; 3) 2D mobile with jitter: 2D trajectory optimization considering drone jitter when drone altitude is fixed at minimum altitude; 4) 3D static with jitter: static deployment considering drone jitter and 3D reflection; 5) 3D static without jitter: static deployment with 3D reflection but without considering drone jitter; 6) 2D static with jitter: static deployment considering drone jitter when drone altitude is fixed at minimum altitude.

[0122] Attached Figure 5The convergence of the proposed RFT1 beamforming, RIS phase, and UAV three-dimensional position iterative optimization algorithm is demonstrated. The convergence of all schemes is similar. In addition, it is obvious that the proposed scheme achieves a larger minimum rate than other schemes. This is due to the use of RFT1 beamforming, RIS phase, and UAV three-dimensional position update process, which can fully utilize the high maneuverability of the UAV. In addition, it can be seen that with the increase in the number of iterations, the impact of UAV jitter and height optimization on performance gradually becomes prominent. The final performance is that compared with the other two mobile solutions, the performance of the mobile solution proposed in the present invention is improved by 17.65% and 29.87%, respectively. In the case of static RIS-UAV, these two values ​​reach 28.89% and 45.00%, respectively.

[0123] Attached Figure 6 (a) Intuitively shows the 3D position and trajectory of RIS-UAV after optimization of the six schemes. Figure 6 (b) shows the time-varying UE scheduling results for the mobile scenario and Figure 6 (c) shows the fixed power allocation result of the static scheme. Figure 6 (b) and Figure 6 (c) It can be seen that the transmission time allocation of different users in all mobile schemes and the power allocation of different users in the static scheme are proportional to the distance between the user and the base station, which is in line with our expectations. Figure 6 All the results in the above calculations are then used to control the flight trajectory of the RIS-UAV, the RIS phase, and the BS beamforming.

[0124] Attached Figure 7 The influence of the angle error caused by the UAV jitter on the maximum and minimum rates under different schemes is given. As the azimuth and elevation errors between the BS and the UAV increase, the performance of all schemes decreases to varying degrees, but our proposed scheme shows great advantages for both mobile and static UAVs. This is because we consider the height optimization under the UAV jitter and further compensate for the performance loss caused by the elevation error, which is not studied in other benchmark schemes. In addition, we also observed that when θ BR When the maximum change of θ is small, the performance gain obtained by optimizing the drone height is lower than the performance gain obtained by considering the drone jitter. However, as θ BR As the maximum change in

[0125] Attached Figure 8The relationship between the maximum and minimum rates and the number of RIS units N is shown. It can be seen that deploying more units in the passive RIS can improve the performance of the system. This performance gain gradually saturates with the increase in the number of RIS elements, especially for the static UAV solution, because the interference caused by the passive RIS reflection also increases with the increase of N. In addition, UAV jitter can greatly reduce the gain provided by RIS, while UAV 3D trajectory planning and considering 3D angular errors can minimize the impact of UAV jitter, which is also the reason for the poor performance of other baseline solutions. Increasing the number of antennas can improve system performance, but it also makes the system more sensitive to angular errors. This is because the more antennas there are, the narrower the beam formed, and the greater the performance loss caused by UAV jitter.

Claims

1. A method for configuring network parameters of an aerial dynamic intelligent reflector, characterized in that: include: The reconfigurable intelligent surface onboard an unmanned aerial vehicle (RIS-UAV) is used to reflect the signal of the local first-type transceiver device (RFT1) to provide data transmission services for the remote second-type transceiver device (RFT2). The communication process of RIS-UAV, RFT1 and RFT2 with RFT1 as the calculation and control center is proposed, and a 3D air-ground channel model including angle error is established. The RFT1 beamforming matrix, RIS phase matrix and UAV three-dimensional position optimization method based on iterative optimization and first-order Taylor expansion are designed, and the RFT1 beamforming matrix, RIS phase matrix and UAV three-dimensional position update mechanism and decision signaling process based on "maximum percentage of user performance degradation" are given.

2. The method according to claim 1, characterized in that The proposed communication process of RIS-UAV, RFT1 and RFT2 includes two stages, namely the initialization stage and the normalization stage. In the initialization stage, RFT2 reports its own position information, and RFT1 calculates the network parameters, including the beamforming matrix of RFT1, the RIS phase matrix and the three-dimensional position of UAV, according to the 3D air-ground channel model containing the angle error. Then, RFT1 updates its own beamforming matrix and sends control information to RIS-UAV to update the RIS phase matrix and the three-dimensional position of UAV, and then enters the normalization stage. In the normalization stage, RFT2 reports its position information and current rate information. RFT1 evaluates whether the network parameters need to be updated according to the rate information of all RFT2, including the beamforming matrix of RFT1, the RIS phase matrix and the three-dimensional position of UAV. If update is required, the network parameters, including the beamforming matrix of RFT1, the RIS phase matrix and the three-dimensional position of UAV, are recalculated. If update is not required, no processing is performed. Then, RFT2 reports its position information and current rate information, and the above process is repeated.

3. The method according to claim 1, characterized in that The channel model of the present invention takes into account the effect of the angle error caused by position movement and UAV jitter on the path loss index. Rice Factor The influence of channel parameters is expressed as follows: Among them, α max and α min are the maximum and minimum values ​​of the path loss index, a1 and b1 are constants determined by the environment, and K max and K min are the maximum and minimum values ​​of the Rice factor, θ is the pitch angle of RFT1 or RFT2 relative to RIS-UAV, and Δθ is the pitch angle jitter error of RFT1 or RFT2 relative to RIS-UAV.

4. The method according to claim 1, characterized in that Based on iterative optimization and first-order Taylor expansion design, RFT1 beamforming (corresponding to question P1 in the manual), RIS phase (corresponding to question P2 in the manual) and UAV three-dimensional position are optimized. Among them, the UAV three-dimensional position optimization is divided into two serial sub-problems: 2D position optimization (corresponding to question P3 in the manual) and height optimization (corresponding to question P4 in the manual). By iteratively solving "P1→P2→P3→P4→P1→P2→…", until the increment of the minimum user transmission rate η is less than a set threshold, the RFT1 beamforming, RIS phase and UAV three-dimensional position at this time can be considered as configurable optimal network parameters.

5. The method according to claim 1, characterized in that According to the current transmission rate reported by each RFT2, RFT1 calculates its rate drop relative to the previous Q time slots and calculates the average rate drop percentage; if more than X% of the average rate drop percentages of RFT2 exceed Y%, the RFT1 beamforming, RIS phase and UAV position are updated; otherwise, no update is performed and no operation is performed; the values ​​of Q, X, and Y can be pre-set according to actual conditions.

6. The method according to claim 1, characterized in that In the decision signaling process of RFT1 beamforming, RIS phase and UAV position update, RFT1 sends the calculated RIS phase and UAV position information to RIS-UAV. After receiving the information, RIS-UAV sends an ACK message to RFT1 and updates the RIS phase and its own position information at the same time. If RIS-UAV does not receive the message, RFT1 will continue to retransmit the information to RIS-UAV; after RIS-UAV updates the RIS phase and its own position, it will send an ACK message to RFT1 to ensure information synchronization.

Images