Energy efficiency-first intelligent reflective surface-assisted UAV communication anti-interference robust design method

Abstract

The present invention provides an energy-efficiency-prioritized intelligent reflective surface-assisted robust anti-interference design method for UAV communication, comprising the following specific steps: S1. Obtaining global channel state information: Due to the high altitude and high line-of-sight link characteristics of UAVs, UAV communication channels are primarily line-of-sight channels. The present invention provides an energy-efficiency-prioritized intelligent reflective surface-assisted UAV communication anti-interference design method to address the problems of high energy consumption and low efficiency in the UAV communication anti-interference process. This method utilizes an alternating optimization algorithm combining continuous convex approximation, fractional programming, and S-procedure to iteratively optimize its transmit power allocation, the reflection coefficient of the intelligent reflective surface, and the UAV trajectory to continuously improve the UAV's receiving energy efficiency until convergence. This optimization strategy can be widely applied to UAV communication anti-interference scenarios, but is not limited to the scope listed above.

Description

Technical Field

[0001] The present invention relates to the field of anti-interference in UAV communications, and in particular to a robust design method for anti-interference in UAV communications assisted by intelligent reflective surfaces with energy efficiency priority. Background Art

[0002] Over the past few decades, the rapid development of drone communications has greatly improved human life. Drone communication systems overcome the inherent shortcomings of traditional communication systems, are not restricted by geographic location, and can be flexibly and rapidly deployed, thus providing fast wireless communication services. However, the high line-of-sight channels and open nature of wireless communication environments make drone communication systems vulnerable to security threats such as malicious interference or illegal eavesdropping, which can facilitate illegal activities such as endangering public safety and invading the privacy of others. To address malicious interference from adversaries, traditional drone communication anti-interference solutions rely on "spatial retreat" to mitigate the impact of interference. This not only increases drone flight energy consumption but also fails to completely address the impact of malicious interference. To effectively address and solve these challenges, there is an urgent need for more effective drone anti-interference measures to achieve intelligence gathering, terrorism / crime prevention, and investigation purposes.

[0003] Smart reflective surfaces have recently emerged as a promising new technology for implementing intelligent wireless channel environments in wireless communication systems. Specifically, a smart reflective surface is a plane containing a large number of passive reflective elements, each of which can independently generate controllable amplitude and / or phase variations for incident signals. Deploying a smart reflective surface provides an additional reflection path for user signal reception, one that is controllable, unlike other scattering and reflection paths in the environment. By deploying smart reflective surfaces in wireless networks and cleverly coordinating their reflections, the wireless channel between transmitters and receivers can be flexibly reconfigured to achieve a desired distribution, thereby addressing wireless channel fading impairments and interference, potentially achieving a quantum leap in wireless communication capacity and reliability.

[0004] Inspired by smart reflective surfaces, using them to assist drone communications in combating security threats is a promising new technology. However, this new technology faces challenges, firstly, in the joint optimization of the drone and the smart reflective surface. The artificial addition of a reflective link complicates the drone's trajectory design. Secondly, in practice, the precise location of the jamming node or eavesdropper is often unavailable, only approximate information can be obtained. This poses new challenges for both the drone's trajectory and the smart reflective surface's beamforming.

[0005] Therefore, the present invention provides a robust design method for anti-interference of intelligent reflective surfaces assisted in UAV communication with energy efficiency as the priority, which is used to address the problem of UAVs being subject to ground interference with uncertain locations. This method considers the uplink of UAV communication. The UAV, as the receiver, receives the transmitted signal from the ground node, which is subject to malicious interference attacks with uncertain location information. At the same time, intelligent reflective surfaces are deployed near the ground nodes to enhance communication effects. By utilizing an alternating iterative algorithm of continuous convex approximation optimization and fractional programming, the beamforming of the intelligent reflective surface and the trajectory of the UAV are continuously iteratively optimized to continuously improve the energy efficiency of the UAV until convergence. This optimization strategy can be widely applied to scenarios targeting ground nodes or multiple interference attacks, but is not limited to the scope listed above. Summary of the Invention

[0006] (1) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a robust design method for anti-interference of intelligent reflective surface-assisted UAV communication with energy efficiency priority, comprising the following specific steps:

[0008] S1. Obtaining global channel state information: Due to the high altitude and high line-of-sight link characteristics of drones, drone communication channels are mainly line-of-sight channels. Due to the flexible deployment of smart reflective surfaces, channels passing through smart reflective surfaces can also be considered line-of-sight links.

[0009] S2. Determine the UAV energy consumption model and establish an optimization problem: Build an energy consumption model for rotary-wing UAVs and a prediction model for interference locations to obtain the approximate locations of interference nodes. Simultaneously, calculate the total throughput of the UAV during its flight cycle. Under the condition of inaccurate interference locations, combine the transmit power of ground nodes, intelligent reflector beamforming, and UAV trajectory constraints to establish an energy efficiency maximization problem.

[0010] S3. Given the ground node transmit power, smart reflector beamforming, and drone trajectory, find the interfering node position with the minimum lower limit for maximizing energy efficiency.

[0011] S4. Optimize the transmit power of ground nodes: Given the reflection coefficient of the smart reflector and the trajectory of the drone, optimize the transmit power of the ground nodes under the condition of imprecise interference position;

[0012] S5. Optimize the reflection coefficient of the smart reflector: Given the transmission power of the ground node and the trajectory of the UAV, robustly design the reflection coefficient of the smart reflector under imprecise interference;

[0013] S6. Optimize UAV trajectory: Robustly design the UAV trajectory under imprecise interference given the ground node transmit power and the reflection coefficient of the smart reflector;

[0014] S7. Iterate and optimize until the energy efficiency of the UAV converges: By repeating steps 3, 4, 5, and 6, the energy efficiency of the UAV is continuously improved, and finally the optimal solution that maximizes the energy efficiency of the UAV is obtained.

[0015] Preferably, the flight trajectory of the UAV is processed using a time discretization method in S1, the periodic flight duration is generally discretized into multiple short time slots, and the mobile UAV is approximately assumed to be stationary in each time slot.

[0016] Furthermore, the path discretization method is adopted in S1 to discretize the continuous time variable into N equally divided quasi-static time slots T = NΔt, where Δt is the length of a time slot. At the same time, the smart reflective surface is equipped with M reflective units to form a uniform linear array and a phase shift controller for intelligently adjusting each element, which is set as the diagonal phase shift matrix of the smart reflective surface in the time slot, and the smart reflective surface is deployed in the xz plane.

[0017] Furthermore, the S1 is used with h gu , h mu , h ru , h gr and h mr Indicates the channels from ground node to UAV, interference node to UAV, smart reflective surface to UAV, ground node to smart reflective surface, and interference node to smart reflective surface.

[0018] Furthermore, in S2, an energy consumption model of the rotary-wing UAV and a prediction model of the interference location are established to obtain the approximate location of the interference node. At the same time, the total throughput of the UAV during the flight cycle is calculated. Under the condition of inaccurate interference location, the energy efficiency maximization problem is established by combining the transmission power of the ground node, the intelligent reflector beamforming and the UAV trajectory constraints.

[0019] The energy consumption of rotary-wing drones usually consists of two parts, namely the energy consumption related to communication and the energy consumption to maintain the drone's flight. It is worth noting that compared with the energy consumption for flight, the energy consumption used by drones for communication is much smaller.

[0020] Furthermore, the interference position estimation is established in S3 when the minimum lower limit of the energy efficiency performance of the UAV is maximized given the ground node transmission power and the reflection coefficient of the smart reflector;

[0021] According to the actual situation in reality, the channels of the interfering nodes belong to a range, namely

[0022]

[0023] where β mu , β mrThey represent the amplitude of fading, and min and max represent the minimum and maximum limits of the amplitude determined by the interference uncertainty.

[0024] In addition, the phase of the smart reflector is processed first, and here we use

[0025]

[0026] Then, define

[0027]

[0028] Here, τ is an arbitrary phase change, so according to a H Γb=v H diag{a H The equation transformation relationship of}b is:

[0029] g g [n]=|v H [n]diag{h0[n]}h g [n]| 2 ,

[0030] g m [n]=|v H [n]diag{h0[n]}h m [n]| 2 ,

[0031] in In order to overcome the influence of imperfect channel information due to position uncertainty, it is considered to establish a K The convex hull method composed of samples is used to solve it, that is

[0032]

[0033] where h t =(diag{h0}h m ) t , α t represents the weighted value of the t-th sample, then formula (3-26) can be expanded to

[0034]

[0035] definition When v is fixed, the optimal weighting coefficient is to make The value is the largest. By using the Cauchy-Schwarz inequality, the solution of the weighting coefficient can be obtained as

[0036]

[0037] Furthermore, in S4, the transmission power of the ground node under the condition of imprecise interference position is optimized under the given reflection coefficient of the smart reflective surface and the trajectory of the UAV;

[0038] For a given UAV trajectory Q and smart reflector phase shift matrix Θ, in order to deal with the uncertainty of the interference position, the worst-case scheduling optimization problem is considered, where the interference is located closest to the UAV within the uncertainty range. Therefore, the problem can be reformulated as:

[0039]

[0040] The reason why it is difficult to solve the problem directly is that the interference location cannot be determined. By analyzing and estimating the interference location, the above objective function can be written as:

[0041]

[0042] in

[0043]

[0044] After the interference position is estimated, the above problem is a standard convex optimization problem and can be solved by the existing optimization tool CVX.

[0045] Furthermore, the reflection coefficient of the smart reflector is optimized in S5: given the transmission power of the ground node and the trajectory of the UAV, the reflection coefficient of the smart reflector is robustly designed under imprecise interference;

[0046] For a given unmanned trajectory Q and ground node transmission power P, the optimization problem can be equivalently expressed as

[0047]

[0048] Since the objective function is relatively complex and multiple variables are coupled together, the above problem is difficult to solve directly. Since the log(·) function is a monotonically increasing function, we can find a set of phase matrices so that the throughput of the objective function in each time slot is maximized, and finally the average rate of the system is maximized. Define Define V[n]=v[n]v H [n], and Where V l,l [n] represents the value of the (l,l)th element of V[n]. However, the rank 1 constraint is a non-convex constraint. Using the semidefinite relaxation method, we first ignore the rank 1 constraint and solve it. Then, we obtain a feasible solution by Gaussian randomization or eigenvalue decomposition. Define and This can be transformed into

[0049]

[0050] stV l,l [n]=k[n],l=1,2...,M+1,

[0051]

[0052] The above is a standard convex optimization problem that can be effectively solved by CVX. However, the rank 1 constraint cannot be guaranteed to be met at this time. Specifically, if V[n] is rank 1, v[n] can be directly obtained by eigenvalue decomposition. Otherwise, v[n] must be approximated by Gaussian randomization. Finally, the phase coefficient of the smart reflector can be obtained by the following formula

[0053]

[0054] Furthermore, in S6, the trajectory of the UAV under imprecise interference is robustly designed under the given ground node transmission power and the reflection coefficient of the smart reflector;

[0055] For a given smart reflector matrix Γ and ground node transmission power P, the optimization problem can be equivalently expressed as

[0056]

[0057] Due to the uncertainty of the interference node position and the non-convex objective function, the above problem. In order to overcome the influence of the uncertainty of the interference node position, the slack variable is introduced To approximate the distance from the jamming node to the UAV, and for the distance from the jamming node to the smart reflective surface, consider the worst case, that is, the jammer is relatively far away. This makes it easier to transform the following problem. At this time, there is

[0058] d[n]≤||q[n]-q m || 2 ,d[n]≥0,

[0059] Among them, due to the uncertainty of the interfering node, q m Contains countless variables. In this case, consider using the alternative theorem of the biquadratic inequality Note that there is Satisfy the constraints and introduce slack variables At this time, the inequality condition is equivalently written as:

[0060]

[0061] in

[0062]

[0063] In order to make the problem easier to solve, we use the first-order Taylor series to get the lower bound, that is, at a given feasible point and E[n] can be newly expressed as a convex form Substituting this into the convex constraint we can obtain the uncertainty of the interference position:

[0064]

[0065] This is a typical semidefinite programming constraint that can be solved using CVX. Next, we will discuss the non-convexity of the objective function in solving the subproblem. Note that the variable g gu [n],g mu [n] and g ru [n] Compared with the trajectory position of the UAV, it is more complex and nonlinear, which makes it difficult to optimize the trajectory of the UAV. In order to overcome this difficulty, it is considered to use the position of the UAV trajectory in the previous iteration to obtain an approximate and As the value of the i-th iteration process. Based on the above approximation, define

[0066]

[0067] in Therefore, it can be restated as

[0068]

[0069] in In addition, two slack variables are introduced and Get the lower bound of the objective function

[0070]

[0071] in

[0072]

[0073] Then the objective function degenerates into finding the maximum value of the lower bound, which is expressed as the added constraints contain non-convex parts, and the slack variables are introduced. and definition To further process non-convex components, it is expanded into:

[0074]

[0075] in

[0076]

[0077] F2[n]=(x[n]-x r ) 2 +(y[n]-y r ) 2 +(H u -z r ) 2 .

[0078] At this point, the Taylor series is used to expand the still non-convex part, and at the feasible point and The specific expansion is F2[n] and also needs to be used Substitution, next, consider the changes to the denominator in the function. The goal of this chapter is to transform the denominator into a convex function.

[0079] First, introduce the slack variable have

[0080]

[0081] Next, make an equivalent transformation on the above and get

[0082]

[0083] Note that the left side is in the form of a convex function, and the next goal is to transform the right side into a concave function by and The Taylor expansion of

[0084]

[0085] So far, all the terms in the problem are convex, and the final problem to be solved is

[0086]

[0087] ξ3 -2 [n]-d[n]≤0,

[0088]

[0089] e[n]>0,

[0090] The above problem is a standard convex optimization problem and can be solved efficiently using the existing solver CVX.

[0091] (2) Beneficial effects

[0092] The present invention provides a robust design method for anti-interference in UAV communications assisted by intelligent reflective surfaces with energy efficiency as the priority. This method addresses the issues of high energy consumption and low efficiency in anti-interference UAV communications. This method utilizes an alternating optimization algorithm combining continuous convex approximation, fractional programming, and S-procedure to iteratively optimize transmit power distribution, the reflection coefficient of the intelligent reflective surface, and the trajectory of the UAV to continuously improve the receiving energy efficiency of the UAV until convergence. This optimization strategy can be widely applied to UAV communication anti-interference scenarios, but is not limited to the scope listed above. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] Figure 1 It is a robust anti-interference model for communication of UAVs assisted by intelligent reflective surfaces with energy efficiency as the priority.

[0094] Figure 2 This is a schematic diagram of the trajectory changes of the UAV with energy efficiency priority under different strategies.

[0095] Figure 3 This is a schematic diagram of the change in drone energy efficiency with the transmission power of the interference node under different strategies.

[0096] Figure 4 This is a schematic diagram of the change in drone energy efficiency with the number of smart reflective surface units under different strategies. DETAILED DESCRIPTION

[0097] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0098] Example 1:

[0099] like Figure 1-4 As shown, the embodiment of the present invention provides an energy-efficiency-prioritized intelligent reflective surface-assisted UAV communication anti-interference robust design method. The implementation steps of the monitoring solution proposed in the present invention mainly include:

[0100] Step 1: Obtain global channel state information: First, in order to facilitate the design of UAV trajectory, the time discretization method is used to process the flight process of the UAV. The periodic flight duration is generally discretized into multiple short time slots, and the mobile UAV is approximately assumed to be stationary in each time slot. It is worth noting that the length of each time slot should be carefully selected, because short time slots will lead to high approximation accuracy of static UAVs and high complexity of design problems. We use the path discretization method to discretize the continuous time variable into N equally divided quasi-static time slots T = NΔt, where Δt is the length of a time slot. Therefore, the flight trajectory representation of the UAV can be obtained.

[0101]

[0102] The constraints of the UAV trajectory are:

[0103] q u [0] = q s ,

[0104] q u [N] = q F ,

[0105] ||q u [n]-q u [n-1]||≤V max △t,

[0106] Considering the power limitation of ground nodes, it can be expressed as The constraints are

[0107]

[0108] p[n]≤p max

[0109] At the same time, the intelligent reflective surface is equipped with M reflective units to form a uniform rectangular array, with M = M x ×M z , set as the diagonal phase shift matrix of the smart reflector at the nth time slot. The smart reflector is deployed in the xz plane and its position is represented by q r =[x r ,y r , z r ], let the phase matrix of the smart reflector be The constraints are:

[0110] θ i [n]∈[0, 2π), i∈{1, 2,...M}.

[0111] Next, consider the channel modeling problem in the wireless environment. Due to the high altitude and high line-of-sight link characteristics of drones, the communication channels of drones are mainly line-of-sight channels. In addition, due to the flexible deployment characteristics of smart reflective surfaces, the channels passing through smart reflective surfaces can also be regarded as line-of-sight links. Assuming that all channels are line-of-sight links, use h gu , h mu , h ru , h gt and h mr The channels from the ground node to the UAV, the interference node to the UAV, the smart reflector to the UAV, the ground node to the smart reflector, and the interference node to the smart reflector are represented. The direct channel power gain can be expressed using the free space path loss model as:

[0112]

[0113] where d gu [n]=||q u [n]-q g || represents the distance from the drone to the ground node, ρ represents the channel fading power gain at 1 meter, and λ is the wavelength of the carrier. Using the same method, when the position of the interfering node can be estimated, its channel h can be obtained. mu [n], for the reflection link, it is mainly composed of three parts cascade, first from the ground node to the smart reflective surface, through the phase adjustment of the smart reflective surface, and then from the smart reflective surface to the UAV. Specifically, the channel from the smart reflective surface to the UAV can be expressed as The specific form is:

[0114]

[0115] It is expressed as the phase response of the channel, which is specifically expressed as

[0116]

[0117] in

[0118]

[0119] Respectively represent the vertical departure angle and horizontal departure angle at the smart reflective surface. ru [n]=‖q[n]-q r || is the distance from the smart reflective surface to the drone, and d represents the distance between two adjacent units of the smart reflective surface. Using the same method, when the interference node position is estimated, the channel from the ground node to the smart reflective surface and the interference node to the smart reflective surface can be obtained. The expression of the channel gain is thus obtained as

[0120]

[0121] Similarly, we can get h gr [n] and h mr [n];

[0122] Step 2: Determine the UAV energy consumption model and establish an optimization problem: Build an energy consumption model for the rotary-wing UAV and a prediction model for the interference location to obtain the approximate location of the interfering node. Simultaneously, calculate the total throughput of the UAV during its flight cycle. Under the condition of inaccurate interference locations, combine the ground node transmit power, intelligent reflector beamforming, and UAV trajectory constraints to establish an energy efficiency maximization problem.

[0123] The energy consumption of rotary wing UAVs usually consists of two parts, namely the energy consumption related to communication and the energy consumption for maintaining the UAV flight. It is worth noting that the energy consumption used by UAVs for communication is much smaller than the energy consumption for flight. Therefore, only the energy consumption used to maintain the UAV flight or hovering is considered, and E is used here. p [n] represents the unit of Joule, so the speed is v u The energy consumption model of the UAV of [n] is

[0124]

[0125] where v u represents the horizontal flight speed of the UAV, d' and s' represent the fuselage drag ratio and rotor strength respectively. A and ρ' represent the rotor disk area and air density respectively. P0 and P1 represent the blade profile power and induced power when the UAV is hovering, respectively, which are constants related to the UAV. u It can be obtained by the following formula

[0126]

[0127] and and It is given by the following formula

[0128]

[0129] Among them, U tip is the tip speed of the UAV rotor blade, and v0 represents the average rotor induced speed of the UAV.

[0130] For drones, the location of the ground equipment is known, while the location of the interference is partially known, which is consistent with most practical scenarios. Assuming that only the estimated location of the interference, that is, the center of the hemisphere, is known, it is Therefore, we can get the following formula

[0131]

[0132] Where (Δx m , Δy m , Δz m )∈ε m is the error between the estimated position and the actual position, which is subject to the following constraints:

[0133]

[0134] Among them D m is the radius of the hemisphere.

[0135] Assuming that the air-ground channel is allocated unit bandwidth, in order to ensure efficient transmission rate, this chapter aims to maximize energy efficiency, which ultimately leads to the following problem:

[0136]

[0137] stC1:q[0]=q S ,

[0138] C2:q[N]=q F ,

[0139] C3:||q u [n]-q u [n-1]≤V max △t,n=1,2,...,N,

[0140]

[0141] C6: θ i [n]∈[0,2π),i∈{1,...,M},

[0142] Among them, σ 2 represents the power of additive white Gaussian noise at the receiver, p m Represents the transmission power of interference, which can be obtained by the free space loss formula;

[0143] The interference location estimation is established to maximize the minimum lower bound of the UAV energy efficiency performance when given the ground node transmission power and the reflection coefficient of the smart reflector.

[0144] According to the actual situation in reality, the channels of the interfering nodes belong to a range, namely

[0145]

[0146] in

[0147]

[0148] They represent the amplitude of fading, and min and max represent the minimum and maximum limits of the amplitude determined by the interference uncertainty.

[0149] In addition, the phase of the smart reflector is processed first, and here we use

[0150]

[0151] Represents the phase change vector of the smart reflector, and then defines

[0152]

[0153] Here, τ is an arbitrary phase change, so according to a H Γb=v H diag{a H}b's equation transformation relationship is

[0154] g g [n]=|v H [n]diag{h0[n]}h g [n]| 2 ,

[0155] g m [n]=|v H [n]diag{h0[n]}h m [n]| 2 ,

[0156] where g g [n],g m [n] represents the channel gain of the useful signal and the interference signal In order to overcome the influence of imperfect channel information due to position uncertainty, it is considered to establish a K The convex hull method composed of samples is used to solve it, that is

[0157]

[0158] Ω represents the set of weighted values ​​in the convex hull, where h t =(diag{h0}h m ) t , α t represents the weighted value of the t-th sample, then g m [n] can be rewritten as

[0159]

[0160] definition When v is fixed, the optimal weighting coefficient is to make The value is the largest. By using the Cauchy-Schwarz inequality, the solution of the weighting coefficient can be obtained as:

[0161]

[0162] Step 4: Optimize the transmit power of the ground node: Given the reflection coefficient of the smart reflector and the trajectory of the UAV, optimize the transmit power of the ground node under the condition of imprecise interference position;

[0163] For a given UAV trajectory Q and smart reflector phase shift matrix Θ, in order to deal with the uncertainty of the interference position, the worst-case scheduling optimization problem is considered, where the interference is located closest to the UAV within the uncertainty range. Therefore, the problem can be reformulated as:

[0164]

[0165] The reason why it is difficult to solve the problem directly is that the interference location cannot be determined. By analyzing and estimating the interference location, the above objective function can be written as:

[0166]

[0167] in

[0168]

[0169] After the interference position is estimated, the above problem is a standard convex optimization problem, which can be solved by the existing optimization tool CVX;

[0170] Optimizing the reflection coefficient of the smart reflector: Given the ground node's transmit power and the drone's trajectory, the reflection coefficient of the smart reflector is robustly designed under imprecise interference.

[0171] For a given unmanned trajectory Q and ground node transmission power P, the optimization problem can be equivalently expressed as

[0172]

[0173] Since the objective function is relatively complex and multiple variables are coupled together, the above problem is difficult to solve directly. Note that since the log(·) function is a monotonically increasing function, we can find a set of phase matrices so that the throughput of the objective function in each time slot is maximized, and finally maximize the average rate of the system. Define Define V[n]=v[n]v H [n], and Where V l,l[n] represents the value of the (l,l)th element of V[n]. However, the rank-1 constraint is non-convex. Using the semidefinite relaxation method, we first ignore the rank-1 constraint and then obtain a feasible solution through Gaussian randomization or eigenvalue decomposition.

[0174] definition and This can be transformed into

[0175]

[0176] stV l，l [n]=k[n],l=1,2.,M+1,

[0177]

[0178] The above is a standard convex optimization problem that can be effectively solved using CVX. However, the rank 1 constraint cannot be guaranteed to be met. Specifically, if V[n] is rank 1, v[n] can be directly obtained through eigenvalue decomposition. Otherwise, v[n] must be approximated through Gaussian randomization. Finally, the phase coefficient of the smart reflector can be obtained using the following formula:

[0179]

[0180] Optimizing UAV trajectories: Robustly designing UAV trajectories under imprecise interference, given the ground node transmit power and the reflection coefficient of the smart reflector.

[0181] For a given smart reflector matrix Γ and ground node transmission power P, the optimization problem can be equivalently expressed as

[0182]

[0183] Due to the uncertainty of the interference node position and the non-convex objective function, the above problem, in order to overcome the influence of the uncertainty of the interference node position, introduce the slack variable To approximate the distance from the jamming node to the UAV, and for the distance from the jamming node to the intelligent reflective surface, consider the worst case, that is, the jammer is relatively far away, so that it is convenient to transform the subsequent problems. At this time, there is

[0184] d[n]≤||q[n]-q m || 2 ,d[n]≥0,

[0185] Among them, due to the uncertainty of the interfering node, q m Contains countless variables. In this case, consider using the alternative theorem of the biquadratic inequality Note that there is Satisfy the constraints and introduce slack variables At this time, the inequality condition is equivalently written as

[0186]

[0187] in

[0188]

[0189] In order to make the problem easier to solve, we use the first-order Taylor series to get the lower bound, that is, at a given feasible point and E[n] can be newly expressed as a convex form Substituting into the convex constraint with uncertain interference position can be obtained

[0190]

[0191] This is a typical semidefinite programming constraint, which can be solved using CVX. Next, we will discuss the non-convexity of the objective function in solving the subproblem. Note that the variable g gu [n],g mu [n] and g ru [n] Compared with the trajectory position of the UAV, it is more complex and nonlinear, which makes it difficult to optimize the trajectory of the UAV. In order to overcome this difficulty, it is considered to use the position of the UAV trajectory in the previous iteration to obtain an approximate and As the value in the i-th iteration process, based on the above approximation, define

[0192]

[0193] in Therefore, it can be restated as

[0194]

[0195] in In addition, two slack variables are introduced and Get the lower bound of the objective function

[0196]

[0197] in

[0198]

[0199] Then the objective function degenerates into finding the maximum value of the lower bound, which is expressed as the added constraints contain non-convex parts, and the slack variables are introduced. and definition To further process non-convex components, it is expanded into:

[0200]

[0201] in

[0202]

[0203] F2[n]=(x[n]-x r ) 2 +(y[n]-y r ) 2 +(H u -z r ) 2 .

[0204] At this point, the Taylor series is used to expand the still non-convex part, and at the feasible point and The specific expansion is F2[n] and also needs to be used Next, consider the changes to the denominator in the function. The goal of this chapter is to transform the denominator into a convex function. First, introduce the slack variable have

[0205]

[0206] Next, make an equivalent transformation on the above and get

[0207]

[0208] Note that the left side is in the form of a convex function, and the next goal is to transform the right side into a concave function by and The Taylor expansion of

[0209]

[0210] So far, all the terms in the problem are convex, and the final problem to be solved is

[0211]

[0212] ξ3 -2 [n]-d[n]≤0,

[0213]

[0214] e[n]>0,

[0215] The above problem is a standard convex optimization problem and can be solved efficiently using the existing solver CVX;

[0216] Step 7: Iterate and optimize until the energy efficiency of the drone converges: By repeating steps 3, 4, 5, and 6, the energy efficiency of the drone is continuously improved, and the optimal solution that maximizes the energy efficiency of the drone is finally obtained;

[0217] The present invention uses intelligent reflecting surfaces to assist UAV communications, providing a new method for anti-interference of UAV communications. For the communicating parties, the main steps of this method are as follows: 1. Acquiring global channel state information; 2. Predicting the approximate location of interference; 3. Optimizing the power of ground nodes; 4. Designing the passive beamforming coefficients of the intelligent reflecting surface; 5. Designing the flight trajectory of the UAV. This method continuously iterates and predicts the interference location, optimizes ground power distribution, the reflection coefficient of the intelligent reflecting surface, and the UAV trajectory until the UAV energy efficiency converges by utilizing optimization algorithms such as continuous convex approximation, S-programming, and time-sharing planning. This optimization method can be widely used in the field of anti-interference of UAV communications.

[0218] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An energy-efficiency-first intelligent reflective surface-assisted UAV communication anti-interference robust design method, characterized by: The specific steps include: S1. Obtaining global channel state information: Due to the high altitude and high line-of-sight link characteristics of drones, drone communication channels are mainly line-of-sight channels. Due to the flexible deployment of smart reflective surfaces, channels passing through smart reflective surfaces can also be considered line-of-sight links. S2. Determine the UAV energy consumption model and establish an optimization problem: Build an energy consumption model for rotary-wing UAVs and a prediction model for interference locations to obtain the approximate locations of interference nodes. Simultaneously, calculate the total throughput of the UAV during its flight cycle. Under the condition of inaccurate interference locations, combine the transmit power of ground nodes, intelligent reflector beamforming, and UAV trajectory constraints to establish an energy efficiency maximization problem. S3. Given the ground node transmit power, smart reflector beamforming, and drone trajectory, find the interfering node position with the minimum lower limit for maximizing energy efficiency. S4. Optimize the transmit power of ground nodes: Given the reflection coefficient of the smart reflector and the trajectory of the drone, optimize the transmit power of the ground nodes under the condition of imprecise interference position; S5. Optimize the reflection coefficient of the smart reflector: Given the transmission power of the ground node and the trajectory of the UAV, robustly design the reflection coefficient of the smart reflector under imprecise interference; S6. Optimize UAV trajectory: Robustly design the trajectory of the UAV under imprecise interference given the ground node transmit power and the reflection coefficient of the smart reflector; S7. Iterate and optimize until the energy efficiency of the UAV converges: By repeating steps 3, 4, 5, and 6, the energy efficiency of the UAV is continuously improved, and finally the optimal solution that maximizes the energy efficiency of the UAV is obtained.

2. The energy efficiency-first intelligent reflective surface-assisted UAV communication anti-interference robust design method according to claim 1 is characterized by: In S1, the flight trajectory of the UAV is processed using a time discretization method, the periodic flight duration is discretized into multiple short time slots, and the mobile UAV is assumed to be stationary in each time slot.

3. The energy efficiency-first intelligent reflective surface-assisted UAV communication anti-interference robust design method according to claim 1 is characterized by: In the S1, a path discretization method is used to discretize the continuous time variable into N equally divided quasi-static time slots T = NΔt, where Δt is the length of a time slot. At the same time, the smart reflective surface is equipped with M reflective units to form a uniform linear array and a phase shift controller for intelligently adjusting each element, which is set as the diagonal phase shift matrix of the smart reflective surface in the nth time slot. The smart reflective surface is deployed in the xz plane.

4. The energy efficiency-first intelligent reflective surface-assisted UAV communication anti-interference robust design method according to claim 1 is characterized by: In the S1, h gu , h mu , h ru , h gr and h mr Indicates the channels from ground node to UAV, interference node to UAV, smart reflective surface to UAV, ground node to smart reflective surface, and interference node to smart reflective surface.

5. The energy efficiency-first intelligent reflective surface-assisted UAV communication anti-interference robust design method according to claim 1 is characterized by: In S2, an energy consumption model of the rotary-wing UAV and a prediction model of the interference location are established to obtain the approximate location of the interference node. At the same time, the total throughput of the UAV during the flight cycle is calculated. Under the condition of inaccurate interference location, the energy efficiency maximization problem is established by combining the transmission power of the ground node, the intelligent reflector beamforming and the UAV trajectory constraints. The energy consumption of rotary-wing drones consists of two parts: energy consumption related to communication and energy consumption for maintaining drone flight. It is worth noting that compared with flight energy consumption, the energy consumption used by drones for communication is much smaller.

6. The energy efficiency-first intelligent reflective surface-assisted UAV communication anti-interference robust design method according to claim 4 is characterized by: In said S3, an interference position estimation is established when the minimum lower limit of the energy efficiency performance of the UAV is maximized when the ground node transmission power and the reflection coefficient of the smart reflective surface are given; According to the actual situation in reality, the channels of the interfering nodes belong to a range, namely where β mu , β mr Represent the amplitude of fading, φ mr represents the arrival angle of interference to the smart reflector, min and max represent the minimum and maximum amplitude limits determined by the interference uncertainty; In addition, the phase of the smart reflector is processed first, and here we use Then, define Here, τ is an arbitrary phase change, so according to a H Γb=v H diag{a H The equation transformation relationship of}b is: g g [n]=|v H [n]diag{h0[n]}h g [n]| 2 , g m [n]=|v H [n]diag{h0[n]}h m [n]| 2 , in In order to overcome the influence of imperfect channel information due to position uncertainty, it is considered to establish a K The convex hull method composed of samples is used to solve it, that is where h t =(diag{h0}h m ) t , α t represents the weighted value of the t-th sample; then formula (3-26) can be expanded to definition When v is fixed, the optimal weighting coefficient is to make The value is the largest. By using the Cauchy-Schwarz inequality, the solution of the weighting coefficient can be obtained as 7. The energy efficiency-first intelligent reflective surface-assisted UAV communication anti-interference robust design method according to claim 1 is characterized by: In said S4, under the given reflection coefficient of the smart reflective surface and the trajectory of the UAV, the transmission power of the ground node under the condition of imprecise interference position is optimized; For a given UAV trajectory Q and smart reflector phase shift matrix Θ, in order to deal with the uncertainty of the interference position, the worst-case scheduling optimization problem is considered, where the interference is located closest to the UAV within the uncertainty range. Therefore, the problem can be reformulated as The reason why it is difficult to solve directly is that the interference position cannot be determined. By analyzing and estimating the interference position, the objective function can be written as in After the interference position is estimated, the above problem is a standard convex optimization problem and can be solved by the existing optimization tool CVX.

8. The energy efficiency-first intelligent reflective surface-assisted UAV communication anti-interference robust design method according to claim 1 is characterized by: Optimizing the reflection coefficient of the smart reflector in S5: Robustly designing the reflection coefficient of the smart reflector under imprecise interference under given ground node transmission power and drone trajectory conditions; For a given unmanned trajectory Q and ground node transmission power P, the optimization problem can be equivalently expressed as Since the objective function is relatively complex and multiple variables are coupled together, the above problem is difficult to solve directly. Since the log(·) function is a monotonically increasing function, we can find a set of phase matrices so that the throughput of the objective function in each time slot is maximized, and finally the average rate of the system is maximized. Define Define V[n]=v[n]v H [n], and Where V l,l [n] represents the value of the (l,l)th element of V[n]. However, the rank 1 constraint is a non-convex constraint. Using the semidefinite relaxation method, we first ignore the rank 1 constraint and solve it. Then, we obtain a feasible solution by Gaussian randomization or eigenvalue decomposition. Define and This can be transformed into: The above is a standard convex optimization problem that can be effectively solved by CVX. However, the rank 1 constraint cannot be guaranteed to be met at this time. Specifically, if V[n] is rank 1, v[n] can be directly obtained by eigenvalue decomposition. Otherwise, v[n] must be approximated by Gaussian randomization. Finally, the phase coefficient of the smart reflector can be obtained by the following formula 9. The energy efficiency-first intelligent reflective surface-assisted UAV communication anti-interference robust design method according to claim 1 is characterized by: In S6, under the given ground node transmission power and the reflection coefficient of the smart reflector, the trajectory of the UAV under imprecise interference is robustly designed; For a given smart reflector matrix Γ and ground node transmission power P, the optimization problem can be equivalently expressed as Due to the uncertainty of the interference node position and the non-convex objective function, the above problem, in order to overcome the influence of the uncertainty of the interference node position, introduce the slack variable To approximate the distance from the jamming node to the UAV, and for the distance from the jamming node to the intelligent reflective surface, consider the worst case, that is, the jammer is relatively far away, so that it is convenient to transform the subsequent problems. At this time, there is Among them, due to the uncertainty of the interfering node, q m Contains countless variables. In this case, consider using the alternative theorem of the biquadratic inequality Note that there is Satisfy the constraints and introduce slack variables At this time, the inequality condition is equivalently written as in In order to make the problem easier to solve, we use the first-order Taylor series to get the lower bound, that is, at a given feasible point and E[n] can be newly expressed as a convex form Substituting into the convex constraint with uncertain interference position can be obtained This is a typical semidefinite programming constraint, which can be solved using CVX. Next, we will discuss the non-convexity of the objective function in solving the subproblem. Note that the variable g gu [n],g mu [n] and g ru [n] Compared with the trajectory position of the UAV, it is more complex and nonlinear, which makes it difficult to optimize the trajectory of the UAV. In order to overcome this difficulty, it is considered to use the position of the UAV trajectory in the previous iteration to obtain an approximate and As the value in the i-th iteration process, based on the above approximation, define in Therefore, it can be restated as in In addition, two slack variables are introduced and Get the lower bound of the objective function in Then the objective function degenerates into finding the maximum value of the lower bound, which is expressed as the added constraints contain non-convex parts, and the slack variables are introduced. and definition In order to further process the non-convex components, it can be expanded into in F2[n]=(x[n]-x r ) 2 +(y[n]-y r ) 2 +(H u -z r ) 2 At this point, the Taylor series is used to expand the still non-convex part, and at the feasible point and The specific expansion is F2[n] and also needs to be used Substitution; Next, consider the changes in the denominator in the function. The goal of this chapter is to transform the denominator into a convex function. First, introduce the slack variable have Next, make an equivalent transformation on the above formula and get Note that the left side is in the form of a convex function, and the next goal is to transform the right side into a concave function by and The Taylor expansion of So far, all the terms in the problem are convex, and the final problem to be solved is: The above problem is a standard convex optimization problem and can be solved efficiently using the existing solver CVX.

Images