A Design Method for Full-Duplex UAV Relay User Scheduling, Trajectory Optimization, and Resource Allocation

Abstract

The present invention relates to a method for full-duplex UAV relay user scheduling, trajectory optimization and resource allocation, belonging to the field of wireless communication technology. The present invention is a method for maximizing the minimum rate of users by jointly optimizing user scheduling, the flight trajectory of the UAV relay, beamforming and the transmission power of each transmitting end. This method decomposes the multi-dimensional complex joint optimization problem into relatively simple sub-problems of low dimension, namely user scheduling, UAV trajectory, beamforming and transmission power sub-problems, by introducing auxiliary variables and the block coordinate descent method, and uses methods such as introducing slack variables, successive convex approximation, and alternating interference suppression to transform and solve the sub-problems to obtain the final result. Thus, the minimum user rate is maximized.

Description

Technical Field

[0001] The present invention relates to a design method for full-duplex UAV relay user scheduling, trajectory optimization, and resource allocation, belonging to the field of wireless communication technology. Background Art

[0002] With the deployment of the fifth-generation wireless system, there has been an explosive growth in mobile traffic data. As an auxiliary communication means, unmanned aerial vehicles (UAVs) have also begun to be widely used in communication networks. Due to the flexibility and variability of UAVs, they have a wide range of applications in fields such as emergency rescue and surveillance. Currently, there are mainly three typical scenarios for UAVs as aerial base stations to assist and enhance communication, namely UAV aerial base stations, UAV-assisted information dissemination and data collection, and UAVs as aerial relays. In particular, using UAVs as aerial relays can improve the existing wireless communication capabilities. Compared with traditional ground relays, UAV relays can operate at higher altitudes, significantly increasing the system coverage. Moreover, UAV relays have a high probability of establishing a visual communication link between the base station and ground user equipment, effectively suppressing the loss of wireless channels for information transmission and enhancing the reliability of wireless communication.

[0003] At the same time, with the rapid development of communication technology, the limited spectrum resources have been difficult to meet the communication service requirements. In recent years, it has been found through research that full-duplex (Non-Orthogonal Multiple Access, NOMA) relay is a promising technology. Compared with traditional half-duplex communication (Orthogonal Multiple Access, OMA), it can support simultaneous and co-frequency information transmission and achieve higher spectral efficiency. In a full-duplex relay network, the base station and users can transmit data on the same spectrum at the same time through the relay. In a UAV relay communication system, adding full-duplex technology to assist communication between the base station and users is one of the current research hotspots. With full-duplex UAV-assisted communication, by optimizing user scheduling and UAV trajectory according to the user location and base station location, and simultaneously optimizing beamforming and transmission power, the minimum transmission rate of users in the communication system can be maximized, thereby improving the performance of the entire communication system. In existing research, there have been more studies on half-duplex UAV relay communication systems, but relatively few studies on full-duplex UAVs as relay-assisted communication, especially fewer studies on jointly optimizing user scheduling, UAV trajectory, beamforming, and transmission power. Therefore, this paper develops a design method for jointly optimizing user scheduling, UAV trajectory, beamforming, and transmission power of full-duplex UAV relays. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a design method for full-duplex UAV relay user scheduling, trajectory optimization and resource allocation, so as to solve the problems of waste of spectrum resources in existing UAV relay communication and poor auxiliary user communication effect.

[0005] The technical solution of the present invention is: a design method for full-duplex UAV relay user scheduling, trajectory optimization and resource allocation, and the specific steps are as follows:

[0006] Step1: Establish a system model of a full-duplex UAV relay system, and the system forwards the data received by the UAV from the base station to a ground user in each time slot.

[0007] Step2: Establish the problem of maximizing the minimum user transmission rate of the above system, and decompose the problem into a user scheduling sub-problem, a UAV trajectory optimization sub-problem, a beamforming optimization sub-problem and a transmit power optimization sub-problem by the block coordinate descent method.

[0008] Step3: Convert the non-convex sub-problems into convex problems by introducing slack variables, successive convex approximation and alternating interference suppression.

[0009] Step4: Design a joint optimization algorithm to jointly optimize user scheduling, UAV trajectory and resource allocation to maximize the minimum user rate.

[0010] The specific content of Step1 is as follows:

[0011] Step1.1: Construct a full-duplex UAV relay system, including 1 ground base station, M randomly distributed ground users and 1 UAV relay. The base station and users perform full-duplex communication through the UAV relay.

[0012] Assume that the UAV flies horizontally at a fixed height H, the flight period is T, the time period T is divided into K equal time slots, and the horizontal coordinates of the UAV at the k-th moment are defined as q[k] = [x(k), y(k)] T , k ∈ {1,...K}, the horizontal coordinates of the base station are q B = [x B , y B T , and the horizontal coordinates of the m-th ground user are q m = [x m , y m T .

[0013] Considering the limitation of flight ability, the movement constraint of the UAV is:

[0014] q[1] = q I

[0015] q[K] = q F

[0016] ||q[k] - q[k - 1]|| ≤ D max , k = 2, ..., K

[0017] Step 1.2: Assume that the channels between the UAV, the base station, and the users are modeled as Rice fading channels, and the channel between the base station and the ground users is modeled as a Rayleigh channel:

[0018] The channel gains from the base station to the UAV, from the UAV to the users, and from the base station to the users are respectively expressed as:

[0019]

[0020]

[0021]

[0022] Among them, β0 represents the channel gain at the reference distance d0 = 1m, represents the distance from the UAV to the base station at time slot k, represents the distance from the UAV to the ground user m at time slot k, represents the distance from the base station to the user m at time slot k, κ is the Rice factor, and respectively represent the scattering components of the channels from the base station to the UAV, from the UAV to the users, and from the base station to the users. Each element is a Gaussian random variable with a mean of 0 and a variance of 1. Among them,

[0023]

[0024] where i = {B, r, t, m}, 0 ≤ l ≤ M i , 0 ≤ j ≤ N i , represents the line-of-sight component from the base station to the UAV, represents the line-of-sight component from the UAV to the user m, θ B , φ B , θ r , φ r respectively represent the elevation departure angle, azimuth departure angle, elevation arrival angle, azimuth arrival angle between the base station and the UAV link, θ m , φ m , θ t , φ t respectively represent the elevation departure angle, azimuth departure angle, elevation arrival angle, azimuth arrival angle in the link from the UAV to the user m. d is the carrier wavelength, and λ is the distance between the antennas.

[0025] Step1.3: Use the binary variable α m [k] to represent user scheduling. The UAV serves 1 user per time slot. When α m [k]=1, it means that communication occurs between the base station and user m at time slot k. Otherwise, α m [k]=0.

[0026] Therefore, the user scheduling constraint is:

[0027]

[0028]

[0029] Step1.4: Let P max1 represent the maximum transmit power of the base station, P max2 represent the maximum transmit power of the UAV, P B [k] represent the transmission power from the base station to the UAV at time slot k, P m [k] represent the transmission power from the UAV to user m at time slot k, P all represent the total power at time slot k.

[0030] The power constraint is:

[0031]

[0032] 0≤P m [k]≤P max2

[0033] P B [k]+P m [k]=P all

[0034] The beamforming constraint is:

[0035]

[0036] Step1.5: The transmission rate from the base station to the UAV at the k-th time slot is:

[0037]

[0038] The transmission rate from the UAV to the m-th user at the k-th time slot is:

[0039]

[0040] Among them, h SI represents the self-interference channel between the full-duplex UAV receive antenna array and transmit antenna array, w B ,wr , w t , w m respectively represent the beamforming vectors at the base station, the beamforming vectors at the UAV receiving antennas, the beamforming vectors at the UAV transmitting antennas, and the beamforming vectors at the user receivers.

[0041] The achievable rate of this model in the k-th time slot is as follows:

[0042] R m [k] = min(R BU [k], R Um [k]) (7)

[0043] Therefore, the total transmission rate of user m over the entire communication period is expressed as:

[0044]

[0045] Specifically, Step 2 is as follows:

[0046] Step 2.1: Jointly optimize user scheduling, UAV flight trajectory, beamforming, and the transmit power of each transmitter to maximize the minimum user rate. The problem is modeled as follows:

[0047]

[0048]

[0049] In Equation (9), the C1 and C2 constraints represent user scheduling constraints for the communication between the base station and a user in each time slot. The C3 - C5 constraints represent the constraints on the UAV trajectory. The C6 - C8 constraints represent the constraints on the transmission power between the UAV and the m-th user. The C9 constraint represents the constant modulus constraints on the base station transmit beamforming, UAV receive beamforming, UAV transmit beamforming, and the m-th user receive beamforming vectors.

[0050] Since the objective function of the problem of jointly optimizing user scheduling, UAV trajectory, beamforming, and transmit power is non-convex, in order to transform the objective function into a convex function for further solution, an auxiliary variable R req is introduced to represent the minimum user rate. Let R req be a function of q, {α m}, P, w. Then the above problem can be transformed into:

[0051]

[0052]

[0053] Step 2.2: Use the block coordinate descent method to decompose the transformed problem into four sub-problems, namely:

[0054] User scheduling sub-problem:

[0055]

[0056]

[0057] UAV trajectory sub-problem:

[0058]

[0059]

[0060] Beamforming sub-problem:

[0061]

[0062]

[0063] Transmit power sub-problem:

[0064]

[0065]

[0066] The specific content of Step 3 is as follows:

[0067] Step 3.1: With the fixed trajectory q, transmit power P, and ideal beamforming vector w, by introducing slack variables, the user scheduling sub-problem can be transformed into:

[0068]

[0069]

[0070] By introducing slack variables, this problem has been transformed into a standard linear programming problem, which can be easily solved by the existing convex optimization toolbox CVX.

[0071] Step 3.2: With the optimized user scheduling decision {α m}, fixed transmit power P, and fixed beamforming w, by introducing the successive convex approximation method and the first-order Taylor expansion, approximate the upper and lower bounds of the uplink rate and the downlink rate, so as to transform the UAV trajectory sub-problem into a convex problem for solution:

[0072]

[0073]

[0074] Step3.3: Under the condition of optimizing the user scheduling decision {α m}, optimizing the UAV trajectory q and the fixed transmission power P, by introducing slack variables, the user beamforming optimization sub-problem is transformed into:

[0075]

[0076]

[0077] Then, the beamforming problem is decomposed into four sub-problems: the beamforming vector at the base station, the beamforming vector at the user, the transmit and receive beamforming vectors at the full-duplex UAV relay. By using the alternating interference cancellation method to alternately optimize these four beam sub-problems, the final beamforming vector is obtained.

[0078] Step3.4: Under the condition of optimizing the user scheduling decision {α m}, optimizing the UAV trajectory q and the beamforming vector {w}, in order to maximize the end-to-end transmission rate, in the full-duplex UAV relay, the optimal power allocation should ensure that the UAV receive rate is equal to the UAV transmit rate.

[0079] The power allocation is represented by the intersection of R BU [k]=R Bm [k] and the line P m [k]+P B [k]=P all [k], that is:

[0080]

[0081] Where

[0082] The specific content of Step4 is as follows:

[0083] Step4.1: Initialize α (0) , q (0) , w (0) , P (0) , set the convergence accuracy ε, set the maximum number of iterations Z, and the iteration number z = 0.

[0084] Step4.2: According to the given q (z-1) , w (z-1) , P (z-1) α (0) , q (0) , w (0) , P (0) Optimize to obtain α (z) .

[0085] Step4.3: According to the given α(z-1) , w (z-1) , P (z-1) Optimize to obtain q (z) .

[0086] Step4.4: According to the given α (z-1) , q (z-1) , P (z-1) Optimize to obtain w (z) .

[0087] Step4.5: According to the given α (z-1) , q (z-1) , w (z-1) Optimize to obtain P (z) .

[0088] Step4.6: Until or l ≥ L max , end the iteration, otherwise continue to execute Step4.2.

[0089] The beneficial effects of the present invention are as follows: Most of the existing UAV relays are in half-duplex working mode and are single-antenna UAVs. The present invention uses a full-duplex multi-antenna UAV as a relay, and jointly optimizes user scheduling, UAV trajectory, and resource allocation according to the distribution of users, so as to better communicate with each user and improve the transmission rate of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Figure 1 is the system model diagram in the embodiment of the present invention;

[0091] Figure 2 is the flow chart of the present invention;

[0092] Figure 3 is the optimized trajectory diagram of the UAV in the embodiment of the present invention;

[0093] Figure 4 is the relationship diagram between the total transmission power and the minimum user rate in the embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0094] The present invention will be further described below in conjunction with the drawings and specific embodiments.

[0095] Embodiment 1: As Figure 1-2 shown, a design method for user scheduling, trajectory optimization, and resource allocation based on a full-duplex UAV relay, the specific steps are as follows:

[0096] Step1: Establish a system model of a full-duplex UAV relay system. This system forwards the data received by the UAV from the base station to a ground user in each time slot.

[0097] Step 2: Establish the problem of maximizing the minimum user transmission rate of the above system, and decompose the problem into a user scheduling sub-problem, a UAV trajectory optimization sub-problem, a beamforming optimization sub-problem, and a transmit power optimization sub-problem by using the block coordinate descent method.

[0098] Step 3: Convert the non-convex sub-problems into convex problems by introducing slack variables, successive convex approximation, and alternating interference suppression methods.

[0099] Step 4: Design a joint optimization algorithm to solve the user scheduling sub-problem, the UAV trajectory optimization sub-problem, the beamforming optimization sub-problem, and the transmit power optimization sub-problem after convex transformation.

[0100] Specifically, in Step 1, in this scenario, the UAV relay operates in full-duplex mode, that is, it can send and receive data simultaneously. The base station is equipped with a uniform planar array \(E = M\times N\), each user is equipped with a uniform planar array \(E = M\times N\), and the UAV is equipped with a uniform planar array \(E = M\times N\) for transmission and a uniform planar array \(E = M\times N\) for reception. The UAV takes off from the specified initial position during the mission interval \(T\), flies along the optimized relay communication trajectory, and then lands at the final position. The channels between the UAV and the base station and users can be modeled as Rice fading channels, and the channel between the base station and the ground users can be modeled as Rayleigh channels. In the \(k\)-th time slot, the channel gains from the base station to the UAV, from the UAV to the user, and from the base station to the user can be expressed as follows: B = M B × N B , each user is equipped with a uniform planar array \(E = M\times N\), m = M m × N m , the UAV is equipped with a uniform planar array \(E = M\times N\) for transmission, a uniform planar array \(E = M\times N\) for reception. The UAV takes off from the specified initial position during the mission interval \(T\), flies along the optimized relay communication trajectory, and then lands at the final position. The channels between the UAV and the base station and users can be modeled as Rice fading channels, and the channel between the base station and the ground users can be modeled as Rayleigh channels. In the \(k\)-th time slot, the channel gains from the base station to the UAV, from the UAV to the user, and from the base station to the user can be expressed as follows: t = M t × N t , a uniform planar array \(E = M\times N\) for transmission, r = M r × N r for reception. The UAV takes off from the specified initial position during the mission interval \(T\), flies along the optimized relay communication trajectory, and then lands at the final position. The channels between the UAV and the base station and users can be modeled as Rice fading channels, and the channel between the base station and the ground users can be modeled as Rayleigh channels. In the \(k\)-th time slot, the channel gains from the base station to the UAV, from the UAV to the user, and from the base station to the user can be expressed as:

[0101]

[0102]

[0103]

[0104] where, \(\beta_0\) represents the channel gain at the reference distance \(d_0 = 1m\). represents the distance from the UAV to the base station at time slot \(k\), represents the distance from the UAV to the ground user \(m\) at time slot \(k\), represents the distance from the base station to the user \(m\) at time slot \(k\). \(\kappa\) is the Rice factor. and Denoted as the scattering components of the base station to UAV, UAV to user, and base station to user channels respectively, where each element is a Gaussian random variable with a mean of 0 and a variance of 1. Among them

[0105]

[0106] where i = {B, r, t, m}, 0 ≤ l ≤ M i , 0 ≤ j ≤ N i , represents the line-of-sight component from the base station to the UAV, represents the line-of-sight component from the UAV to user m. θ B , φ B , θ r , φ r represent the elevation departure angle, azimuth departure angle, elevation arrival angle, and azimuth arrival angle between the base station and UAV links respectively. θ m , φ m , θ t , φ t represent the elevation departure angle, azimuth departure angle, elevation arrival angle, and azimuth arrival angle in the UAV to user m link respectively. d is the carrier wavelength, and λ is the distance between antennas.

[0107] Considering a full-duplex UAV relay-assisted wireless communication network, which consists of a base station, a UAV, and M cellular users. The rate from the base station to the UAV in the k-th time slot is:

[0108]

[0109] where, h BU represents the channel matrix between the base station and the UAV, h SI represents the self-interference channel between the full-duplex UAV receive antenna array and transmit antenna array. P B , P m represent the transmit power from the base station to the UAV and the transmit power from the UAV to the m-th user respectively. w B , w r , w t represent the beamforming vector at the base station, the beamforming vector at the UAV receive antenna, and the beamforming vector at the user receive respectively.

[0110] The transmission rate from the UAV to the m-th user in the k-th time slot is:

[0111]

[0112] where h Um represents the channel matrix between the UAV and the user, h Bmrepresents the channel matrix between the base station and the user, w m represents the beamforming vector at the UAV transmit antenna.

[0113] The achievable rate of this model in the k-th time slot is as follows:

[0114] R m [k]=min(R BU [k],R Um [k])

[0115] Therefore, the total transmission rate of user m during the entire communication period is expressed as:

[0116]

[0117] Furthermore, the problem of jointly optimizing user scheduling, UAV flight trajectory, beamforming, and the transmit power of each transmitter to maximize the minimum user rate is modeled as follows:

[0118]

[0119]

[0120] Among them, the C1 and C2 constraints represent user scheduling constraints, and the communication between the base station and a user in each time slot. The C3 - C5 constraints represent the constraints of the UAV trajectory, the C6 - C8 constraints represent the constraints of the transmission power between the UAV and the m-th user communication, and the C9 constraint represents the constant modulus constraints of the base station transmit beamforming, UAV receive beamforming, UAV transmit beamforming, and the m-th user receive beamforming vector.

[0121] Since the objective function of the problem of jointly optimizing user scheduling, UAV trajectory, beamforming, and transmit power is non-convex, in order to transform the objective function into a convex function for further solution. An auxiliary variable R req is introduced to represent the minimum user rate, and let R req be a function of q, {α m}, P, w. The above problem can be transformed into:

[0122]

[0123]

[0124] The original problem is decomposed into four sub-problems using the block coordinate descent method.

[0125] Among them, the user scheduling sub-problem is expressed as:

[0126]

[0127]

[0128] Given a fixed trajectory \(q\), transmit power \(P\), and ideal beamforming vector \(w\), by introducing slack variables, the user scheduling sub - problem can be transformed into:

[0129]

[0130]

[0131] At this time, this problem has become a standard linear programming problem and can be easily solved by the existing convex optimization toolbox CVX.

[0132] When optimizing the user scheduling decision \(\{\alpha\) m \}\), with a fixed transmit power \(P\) and a fixed beamforming \(w\),

[0133] the UAV trajectory sub - problem can be expressed as:

[0134]

[0135]

[0136] By introducing the successive convex approximation method and the first - order Taylor expansion, the uplink rate and downlink rate are approximated to a lower bound. Therefore, the successive convex approximation method is used here to solve the optimization problem in (31). Although problem (31) is non - convex with respect to \(q[k]\), it is convex with respect to \(\left\lVert q[k]-q\right\rVert\) B \left\lVert\right\rVert 2 . According to the property of the first - order Taylor expansion, the first - order Taylor approximation is the global lower bound of a convex function at any local point. Therefore, we use the first - order Taylor expansion of \(R\) BU [k]\) at \(\left\lVert q[k]-q\right\rVert\) B \left\lVert\right\rVert 2 as the lower bound, denoted as

[0137]

[0138] where:

[0139]

[0140]

[0141] Here is a concave function with respect to \(q[k]\). Similarly, we can also obtain the lower bound of \(R\) Um [k]\), denoted as

[0142]

[0143] Among them:

[0144]

[0145]

[0146] Similarly, here is also a concave function with respect to q[k]. By replacing R m [k] in the constraint C1 in problem (31) with the lower bounds obtained in (32) and (35).

[0147] Next, solve the beamforming optimization sub-problem. When optimizing the user scheduling decision {α m}, optimizing the UAV trajectory q, and fixing the transmit power P, introducing slack variables, the user beamforming optimization sub-problem is transformed into:

[0148]

[0149]

[0150] Then, decompose the beamforming problem into four sub-problems: the beamforming vector at the base station, the beamforming vector at the user, the transmit and receive beamforming vectors at the full-duplex UAV relay. Alternately optimize these four beam sub-problems by the alternating interference cancellation method. In the nth iteration, solving the UAV receive beamforming vector sub-problem can be expressed as:

[0151]

[0152]

[0153] Where and are the fixed base station beamforming vector and the UAV transmission beamforming vector obtained in the (n - 1)th iteration respectively, and δ1 (n) is the interference cancellation factor. Similarly, the other three beamforming vector problems can also be directly obtained, and finally the beamforming vector is obtained through alternating solution.

[0154] Finally, when optimizing the user scheduling decision {α m}, optimizing the UAV trajectory q, and the beamforming vector {w}, in order to maximize the end-to-end transmission rate, in the full-duplex UAV relay, the optimal power allocation should ensure that the transmission rate in (23) is equal to the transmission rate in (24). So the power allocation can be obtained from R BU [k] = R Bm [k] and the line P m [k] + P B [k] = P allThe intersection point of [k] indicates, that is:

[0155]

[0156] Wherein:

[0157] By solving the above equations:

[0158] When AD - BC ≠ 0:

[0159]

[0160] Among them,

[0161] When AD - BC = 0

[0162]

[0163] The optimal solution conclusion of the power distribution problem is as follows:

[0164] 1. If P B [k] ∈ [P all - P max2 , P max1 , and AD - BC ≠ 0, the power optimal solution is (40), when AD - BC = 0, then the power optimal solution is (41);

[0165] 2. If P B [k] ∈ [0, P all - P max2 , the power optimal solution is (P all - P max2 , P max2 );

[0166] 3. If P B [k] ∈ (P max1 , P all , the power optimal solution is (P max1 , P all - P max1 ).

[0167] The solution process of the final problem:

[0168] Step4.1: Initialization; Set the convergence accuracy ε; Set the maximum number of iterations Z; The number of iterations z = 0;

[0169] Step4.2: According to the given q (z-1) , w (z-1),P (z-1) α (0) ,q (0) ,w (0) ,P (0) Seek optimization to obtain α (z)

[0170] Step4.3: According to the given α (z-1) ,w (z-1) ,P (z-1) Optimize to obtain q (z)

[0171] Step4.4: According to the given α (z-1) ,q (z-1) ,P (z-1) Optimize to obtain w (z)

[0172] Step4.5: According to the given α (z-1) ,q (z-1) ,w (z-1) Optimize to obtain P (z)

[0173] Step4.6: Until or l≥L max , end the iteration; otherwise continue to execute Step Five.

[0174] In this implementation case, 1 base station and 6 users are randomly distributed in a 1200*1200 scenario, and a full-duplex UAV relay provides services for the base station and users. The flight time of the UAV T = 100s, the maximum flight height of the UAV H = 100m, and the maximum flight speed total power v max = 30m / s. The total power P all = 30dbm, the maximum transmit power limit of the base station P max1 = 17dbm, the maximum transmit power limit of the UAV P max2 = 18dbm, the carrier frequency f = 38GHz, the noise power σ 2 = -110dbm, the reference channel gain β0 = -50db. According to the settings of these parameters, MATLAB is used to simulate the system.

[0175] As Figure 3 shown, the optimized UAV trajectory trend conforms to the geographical distribution of the base station and users, indicating that the UAV optimizes the trajectory for better communication between the base station and users.

[0176] As Figure 4 shown, the minimum user rate increases with the increase of the total transmission power, and when it increases to a certain value, the minimum user rate remains basically unchanged.

[0177] The specific embodiments of the present invention have been described in detail with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments, and various changes can be made without departing from the spirit of the present invention within the scope of knowledge possessed by those of ordinary skill in the art.

Claims

1. A design method for full-duplex UAV relay user scheduling, trajectory optimization, and resource allocation, characterized in that: Step 1: Establish a system model of a full-duplex UAV relay system, where the UAV forwards the data received from the base station to a ground user in each time slot; Step 2: Establish the problem of maximizing the minimum user transmission rate of the above system, and decompose the problem into a user scheduling sub-problem, a UAV trajectory optimization sub-problem, a beamforming optimization sub-problem, and a transmit power optimization sub-problem by the block coordinate descent method; Step 3: Convert the non-convex sub-problems into convex problems by introducing slack variables, successive convex approximation, and alternating interference suppression; Step 4: Design a joint optimization algorithm to jointly optimize user scheduling, UAV trajectory, and resource allocation to maximize the minimum user transmission rate.

2. The design method for full-duplex UAV relay user scheduling, trajectory optimization, and resource allocation according to claim 1, characterized in that Specifically, Step 1 is as follows: Step 1.1: Construct a full-duplex UAV relay system, which includes 1 ground base station, M randomly distributed ground users, and 1 UAV relay. The base station and users perform full-duplex communication through the UAV relay; Suppose the UAV flies horizontally at a fixed altitude H with a flight period of T. The time period T is divided into K equal time slots. Define the horizontal coordinates of the UAV at the k-th moment as q[k] = [x(k), y(k)] T , where k ∈ {1,...K}, and the horizontal coordinates of the base station are q B = [x B , y B T , and the horizontal coordinates of the m-th ground user are q m = [x m , y m T ;​​ The movement constraint of the UAV is: q[1] = q I q[K] = q F ||q[k] - q[k - 1]|| ≤ D max , k = 2, ..., K In the formula, q[1] represents the horizontal coordinate of the UAV at the first moment, q[K] represents the horizontal coordinate of the UAV at the Kth moment, q[k - 1] represents the horizontal coordinate of the UAV at the (k - 1)th moment, and Dmax represents the maximum value of the horizontal distance of the UAV between the kth moment and the (k - 1)th moment; Step 1.2: Assume that the channels between the UAV and the base station and users are modeled as Rice fading channels, and the channel between the base station and ground users is modeled as Rayleigh channel: The channel gains from the base station to the UAV, from the UAV to the user, and from the base station to the user are respectively expressed as: where β0 represents the channel gain at the reference distance d0 = 1m, represents the distance from the UAV to the base station at time slot k, represents the distance from the UAV to the ground user m at time slot k, represents the distance from the base station to the user m at time slot k, κ is the Rice factor, and represent the scattered components of the channels from the base station to the UAV, from the UAV to the user, and from the base station to the user, respectively, where each element is a Gaussian random variable with a mean of 0 and a variance of 1. Among them, where \(i = \{B,r,t,m\}\), \(0\leq l\leq M\) i , \(0\leq j\leq N\) i , the base station is equipped with a uniform planar array \(E\) B = M B × N B , each user is equipped with a uniform planar array \(E\) m = M m × N m , the UAV is equipped with a uniform planar array \(E\) t = M t × N t for transmission, and a uniform planar array \(E\) r = M r × N r for reception, represents the line-of-sight component from the base station to the UAV, represents the line-of-sight component from the UAV to user \(m\), \(\theta\) B , \(\varphi\) B , \(\theta\) r , \(\varphi\) r respectively represent the elevation departure angle, azimuth departure angle, elevation arrival angle, azimuth arrival angle between the base station and the UAV link, \(\theta\) m , \(\varphi\) m , \(\theta\) t , \(\varphi\) t respectively represent the elevation departure angle, azimuth departure angle, elevation arrival angle, azimuth arrival angle in the link from the UAV to user \(m\), \(d\) is the carrier wavelength, and \(\lambda\) is the distance between antennas; Step 1.3: Use the binary variable α m [k] to represent user scheduling. The UAV serves one user per time slot. When α m [k]=1, it means that communication occurs between the base station and user m at time slot k. Otherwise, α m [k]=0; Therefore, the user scheduling constraint is: Step 1.4: Let P max1 represent the maximum transmit power of the base station, and P max2 represent the maximum transmit power of the UAV. Let P B [k] denote the transmission power sent from the base station to the UAV in time slot k, and P m [k] denote the transmission power sent from the UAV to user m in time slot k. Let P all denote the total power at time slot k; The power constraint is: 0 ≤ P m [k] ≤ P max2 P B [k] + P m [k] = P all The beamforming constraint is: Step 1.5: The transmission rate from the base station to the UAV in the kth time slot is: The transmission rate from the UAV to the mth user in the kth time slot is: Among them, h SI represents the self-interference channel between the full-duplex UAV receive antenna array and the transmit antenna array, w B , w r , w t , w m respectively represent the beamforming vectors at the base station, the beamforming vectors at the UAV receive antennas, the beamforming vectors at the UAV transmit antennas, and the beamforming vectors at the user receivers. σ represents the noise power, and σ m represents the noise power of the ground user m; The achievable rate of this model in the kth time slot is as follows: R m [k] = min(R BU [k], R Um [k]) (7) Therefore, the total transmission rate of user m during the entire communication period is expressed as:

3. The design method for full-duplex UAV relay user scheduling, trajectory optimization, and resource allocation according to claim 1, characterized in that Specifically, Step 2 is as follows: Step 2.1: The problem of jointly optimizing user scheduling, UAV flight trajectory, beamforming, and the transmit power of each transmitter to maximize the minimum user transmission rate is modeled as follows: In formula (9), the C1 and C2 constraints represent user scheduling constraints. The base station communicates with one user through the UAV in each time slot. The C3 - C5 constraints represent the constraints of the UAV trajectory. The C6 - C8 constraints represent the constraints of the transmission power between the UAV and the mth user. The C9 constraint represents the constant modulus constraints of the base station transmit beamforming, UAV receive beamforming, UAV transmit beamforming, and the mth user receive beamforming vectors; Introduce an auxiliary variable R req Denote the minimum user transmission rate and let R req be a function of q, {α m}, P, w, then the above problem can be transformed into: Step 2.2: Use the block coordinate descent method to decompose the transformed problem into four sub-problems, which are respectively: User scheduling sub-problem: UAV trajectory sub-problem: Beamforming sub-problem: Transmit power sub-problem:

4. The design method for full-duplex UAV relay user scheduling, trajectory optimization, and resource allocation according to claim 1, characterized in that Specifically, Step 3 is as follows: Step 3.1: With the fixed trajectory q, transmit power P, and ideal beamforming vector w, by introducing slack variables, the user scheduling sub-problem can be converted to: Step 3.2: Under the condition of optimizing the user scheduling decision {α m}, fixing the transmit power P and the fixed beamforming w, by introducing the successive convex approximation method and the first-order Taylor expansion, approximating the uplink rate and the downlink rate to a lower bound, so as to convert the UAV trajectory sub-problem into a convex problem for solution: Step 3.3: Under the condition of optimizing the user scheduling decision {α m}, optimizing the UAV trajectory q and the fixed transmission power P, by introducing slack variables, the user beamforming optimization sub-problem is transformed into: Then, the beamforming problem is decomposed into four sub-problems: the beamforming vector at the base station, the beamforming vector at the user, the transmit and receive beamforming vectors at the full-duplex UAV relay. These four beam sub-problems are alternately optimized by the alternating interference suppression method to obtain the final beamforming vector; Step 3.4: In the case of optimizing the user scheduling decision {α m}, optimizing the UAV trajectory q and the beamforming vector {w}, in order to maximize the end-to-end transmission rate, in full-duplex UAV relaying, the optimal power allocation should ensure that the UAV reception rate is equal to the UAV transmission rate; P in power distribution m [k] and P B [k] is determined by R BU [k] = R Bm [k] and line P m [k] + P B [k] = P all [k] is obtained from the intersection of the solution represented by, i.e.: Among them 5. The design method for full-duplex UAV relay user scheduling, trajectory optimization, and resource allocation according to claim 1, characterized in that, Specifically, step 4 is as follows: Step 4.1: Initialize α (0) , q (0) , w (0) , P (0) , set the convergence accuracy ε, set the maximum number of iterations Z, and the number of iterations z = 0; Step 4.2: According to the given q (z-l) , w (z-l) , P (z-l) , α (0) , q (0) , w (0) , P (0) Optimize to obtain α (z) ; Step 4.3: According to the given α (z-l) , w (z-l) , P (z-l) optimize to obtain q (z) ; Step 4.4: According to the given α (z-l) , q (z-l) , P (z-l) optimize to obtain w (z) ; Step 4.5: Optimize to obtain P according to the given α (z-l) , q (z-l) , w (z-l) ; (z) ; Step 4.6: Until either l≥Z, end the iteration, otherwise continue to execute Step 4.2.

Images